The role of sensory systems in directional perception of the fiddler crab, Uca pugilator

Transcription

1

2 The role of sensory systems in directional perception of the fiddler crab, Uca pugilator A thesis submitted to the Graduate School of the University of Cincinnati in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE in the Department of Biological Sciences of the College of Arts and Sciences 2011 by Jessica D. Ebie B.A. Wittenberg University, 2004 Committee Chair: John E. Layne, PhD.

3 Abstract Fiddler crabs, like many other organisms, have independently evolved to have compensatory eye movements that help stabilize gaze [1-4]. Humans and non-human taxa have been shown to use these eye movements to help accurately perceive the location of objects in the world around them. This study examines the fiddler crab, Uca pugilator, to investigate whether the role of compensatory eye movements in directional perception has evolved in a similar way as in other taxa. Two experiments were performed with the fiddler crab, Uca pugilator, to study the role of sensory systems in its perception of and interactions with its environment. The first experiment tested the function of the oculomotor reflexes in the maintenance of subjective body axis and awareness of where objects are in space. The second experiment tests the efficiency of magnetic statoliths as replacements for normally non-magnetic statoliths in the statocysts and the role of the statoliths in vertical compensatory eye movements. Oculomotor reflexes (i.e., gaze-stabilizing eye reflexes) were tested to determine whether they contribute to directional perception. Fiddler crabs were exposed to two types of oculomotor stimuli an optokinetic stimulus alone or optokinetic and vestibulo-ocular reflex stimuli presented simultaneously. Following exposure to the oculomotor stimulus, fiddler crabs were presented with a looming stimulus (i.e., a simulated predator) to invoke an escape response. The direction of escape was recorded and compared to that of looming stimulus treatment crabs that only received the looming stimulus. The escape direction was recorded as an indicator of the role of the oculomotor reflexes in directional perception. There was a significant difference in escape direction between looming stimulus treatment crabs and those receiving only the optokinetic stimulus, which suggests that the optokinetic reflex does not update the mapping of object 2

4 images on the retina to perceived location of those objects in space. However, there was no significant difference between looming stimulus treatment crabs and those receiving the optokinetic and vestibulo-ocular stimuli simultaneously, which is a contradictory result. However, fiddler crabs may be using other information when perceiving direction. During the second experiment, fiddler crabs normally non-magnetic statoliths were replaced with metal filings and then exposed to an oscillating magnet to simulate movement in the pitch and roll planes. Simulating movement in these planes allows the role of the statolith in vertical compensatory eye movement to be examined in addition to determining the effectiveness of metal filings as a statolith replacement. The eye oscillation lagged behind the magnet oscillation, but the frequencies were consistent. Measured gain (i.e., eye oscillation amplitude divided by magnet oscillation amplitude) was slightly less than what would be expected if the fiddler crab eye perfectly compensated for the simulated body tilt created by the magnetic apparatus. This attenuation may be due to the lagging response of the fiddler crab to the dynamic stimulus. Knowing that metal filings are effective statolith replacements opens up many possibilities for behavioral manipulations to investigate how information provided to the fiddler crab by the statocyst is used in perception. 3

5 4

6 Acknowledgments I would like to thank Dr. John E. Layne for his support and guidance during this project. I would like to thank Dr. George Uetz and Dr. Elke Buschbeck for their valuable contributions. I would like to thank Theresa Hauser for her assistance with data collection. I would like to thank Roger Ruff for his technical support and Dr. Theodore P. Pavlic for his assistance with Matlab. I would like to thank my family, Dr. Theodore Pavlic, Bonnie Ebie, Flora Ebie, and Ray Ebie, for all their support while I was completing my master s work. I would also like to thank Dr. Dan Rittschoff, Dr. Kathy Reinsel, and Dr. Jim Welch for collecting fiddler crabs in Beaufort, NC for this research. 5

7 Table of Contents ABSTRACT... 2 ACKNOWLEDGMENTS... 5 I. CHANGE IN RETINAL IMAGE LOCATION DUE TO OCULOMOTOR REFLEX AND BODY ORIENTATION AFFECTS ESCAPE DIRECTION IN THE FIDDLER CRAB UCA PUGILATOR... 8 INTRODUCTION... 8 The Oculomotor System... 8 Eye Movements and Perception Eye Movements or their Equivalent in Various Taxa MATERIALS AND METHODS Experimental Procedures Optokinetic Stimulus Optokinetic and Vestibulo-ocular Stimuli Simultaneously Video Analysis DATA ANALYSIS RESULTS Optokinetic Stimulus Results Optokinetic and Vestibulo-ocular Stimuli DISCUSSION II. THE ROLE OF GRAVITY IN INDUCING COMPENSATORY EYE MOVEMENT IN THE VERTICAL PLANE IN THE FIDDLER CRAB UCA PUGILATOR INTRODUCTION

8 MATERIALS AND METHODS Experimental Procedures for Inducing Precocious Ecdysis Experimental Procedures for Vertical Compensatory Eye Movements Video Analysis Data Analysis RESULTS Inducing Precocious Ecdysis Results Vertical Compensatory Eye Movements Results DISCUSSION Inducing Precocious Ecdysis Vertical Compensatory Eye Movements CONCLUSIONS REFERENCES

9 I. Change in retinal image location due to oculomotor reflex and body orientation affects escape direction in the fiddler crab Uca pugilator Introduction The Oculomotor System The oculomotor system has evolved in humans and many non-human taxa and includes both directed and stabilizing eye movements [1-3, 5]. Directed eye movements occur when a decision is made to shift one s gaze to a different focal object. For example, while reading, an individual follows along the text on the page, and the eye moves from word to word. Likewise, if someone wanted to follow a ball sailing through the air, directed eye movements would allow the individual to direct their gaze toward the ball and track it. Unlike directed eye movements, stabilizing eye movements are the oculomotor reflexes that stabilize an image on the retina, allowing an organism to maintain a stable visual field, which assists in the organism s ability to keep track of objects around it and effectively interact with its surroundings [2, 6-9]. Stabilizing eye movements also help to prevent motion blur, which can result in loss of contrast and a smearing of images. Smearing of images occurs as a result of the physiological limitation of the photoreceptors in the eye. Each photoreceptor has its own field of view, its acceptance angle, which varies by eye type and organism [2, 10]. Photoreceptors also have a length of time that they need to be stimulated by an image in order to fire fully. If the image projected on the photoreceptor does not remain stationary for the full firing time before moving, the photoreceptor will produce a signal that is not proportional to the stimulus intensity. When 8

10 the retinal projection of an object moves faster than one acceptance angle per required response time, motion blur or smearing occurs [2, 10]. So the oculomotor reflexes reduce retinal slip of images and thus reduce motion blur. The oculomotor system also contributes to the awareness of what is and is not moving in space and of where things are relative to the organism s body [1, 3, 5]. Specifically, as an organism moves its eyes, retinal image location changes, and the oculomotor system updates the organism s subjective body axis coordinates to reflect the eye movments [1, 5]. This study investigates the role of stabilizing eye movements in the directional perception of the fiddler crab, Uca pugilator. There are various types of eye movements, which play different physiological roles, that compose the directed and stabilizing eye movements. Directed Eye Movements Directed eye movements are not examined during this study, but it is important to understand their function in order to distinguish them from stabilizing eye movements. There are three categories of directed eye movements: smooth pursuit, vergence, and saccades [1, 11]. They occur as a result of a decision to look at an object and track it as it moves through space by moving the eye so that it matches the relative angular velocity of the focal object while following the focal object s movement trajectory to maintain the object s position on the retina. Smooth pursuit movements are discussed in detail in this subsection as one example of these directed eye movements. Smooth pursuit movements enable the eye to keep the image of the focal object within the fovea (i.e., area of the retina with the highest resolution, which images of focal objects are projected on) of the retina [1, 3, 11]. If the object begins to move too fast for the eye to smoothly track it, the eye will use small saccades, another type of directed eye movement consisting of quick jerking movements, in an attempt to maintain the object s position on the fovea. Think of watching a ball during a tennis match, for instance. 9

11 Saccades are rapid jerky eye movements that that help to maintain an image on the fovea of the eye. These movements work in conjunction with smooth pursuit movements when we read, for example. Finally, vergence movements are eye movements that allow an individual to focus on an object at different distances. Vergence occurs when the eyes move closer to the median of the face of farther away depending on the distance of the object from the face (e.g., moving a newspaper closer and farther from the face while reading) [1, 2]. Stabilizing Eye Movements This study focuses on stabilizing eye movements, which perform a very different job for the oculomotor system than directed eye movements. The stabilizing types of eye movement or oculomotor behavior, the vestibulo-ocular and optokinetic responses (VOR and OKR, respectively), maintain the position of an image on the retina and are reflexive eye movements that do not result from a conscious decision to track or change the focal object being viewed. They work in conjunction with each other to compensate for movement of the head by moving the eyes in the opposite direction at an equal velocity [1, 3, 12]. However, the stimuli that initiate these two reflexes are different. The VOR is initiated by stimulation of the vestibular system in humans and other species. In particular, it is initiated by stimulation of the semicircular canals, which are sensitive to angular acceleration. For example, if the canals indicate clockwise angular acceleration in the yaw (i.e., horizontal) plane, the VOR produces eye yaw movement in a counter-clockwise direction. In humans, the VOR helps prevent the appearance of a bouncing environment while an individual is walking by mitigating changes in the visual field by compensating for natural movement of the head [1]. However, the oculomotor system receives no visual feedback from the retina and thus cannot guarantee a stable visual field. The VOR system provides an open-loop (i.e., no feedback) 10

12 control of eye movement that sends sensory signals from the vestibular system to the oculomotor system in response to angular acceleration of the head [1, 11, 13]. After the vestibular system is triggered by head movement, the information gained is used to produce a sudden VOR movement, which has pre-determined parameters, such as the velocity of the eye, that result from the information in the vestibular system. Because the VOR does not receive any direct visual feedback from retinal slip (i.e., movement of an object at an angular velocity that results in movement of the image on the retina when it is unintended) and is not practically capable of removing all such motion. So information about the remaining retinal slip of the visual surround is transmitted to the oculomotor system which generates an OKR to robustly stabilize the image on the retina [1, 13]. In contrast, the OKR compensates for movement of the head by moving the eyes in the same direction as the image motion. For example, when riding in a car and watching out the side windows, our eyes will track telephone poles and other environmental images as they pass by. However, instead of being driven by the vestibular system, it is driven by large-scale image motion on the retina. Such a pattern of motion is unlikely to arise from movement of objects in the environment but instead indicates that the head is moving, and the OKR results in eye movement that compensates for it. The OKR can compensate more precisely than the VOR because it is a feedback system and is better able to maintain the position of the large-scale image on the retina as long as head movements are not too abrupt. These two reflexes complement each other to maintain a stable visual field; the VOR compensates for gross, high frequency components of head movement, and the OKR compensates for the residual low frequency components of the head movements. 11

13 Eye Movements and Perception The eye movements of the oculomotor system play a role in perception by regulating information that the brain receives about the surrounding environment, which the brain adds to a perceived so-called retino-cortical map [1, 3]. This retino-cortical map provides an individual with information about the location of her surroundings relative to her own body by sending information about object locations from the receptors in the retina to the brain, which are then mapped by her conscious facilities onto perceived body-location information, creating a perception of where objects are in the environment relative to a subjective body axis. The retinocortical map allows an organism to navigate and interact with its environment (e.g., successfully pick up a fork from the table when preparing to eat). It is updated when the eyes make directed movements (e.g., smooth pursuit, vergence, or saccades) to reflect any changes of object location relative to the body. However, the oculomotor reflexes occur without conscious awareness, and there is no update to the map when they occur [1, 14]. Visual information is actually divided between two major maps: a visual-motor map, which receives information from reflexive movements of the eyes and peripheral vision, and the retino-cortical map (or cognitive map), which receives information from conscious movements of the eye and receptors in the fovea [14]. The involvement of VOR and OKR in spatial perception can be determined by inducing these oculomotor reflexes and measuring subsequent deviations in spatial judgment. The OKR is induced with a pattern of black and white vertical stripes displayed on a drum rotating around the subject or projected on a screen and made to appear to move [1, 15]. This results in large-scale image motion across the retina and causes the OKR to respond as though the head has moved. The stimulation of an erroneous OKR leads the individual to perceive objects with erroneous 12

14 positions because the visual surround has shifted, but the retino-cortical map has not [16]. For example, Tozzi et al. [15] illustrated the role OKR plays in perception of object location by stimulating OKR in subjects using black and white vertical bars projected onto a screen. The bars were then moved across the screen. At different points during the OKR stimulus, a vertical bar of light was projected onto the screen and flashed on and off so that it appeared and disappeared, and participants would be asked to verbally indicate the location of the bar along an ever-present number line. In all cases, the participants indicated that the flashed bar was in a position that was shifted in the direction of the black and white bar movement from where it was actually displayed, which is explained by the shift of the position on the retina that was stimulated by the flash between the time of flash and verbal indication of the location. The magnitude of the shift depended on whether the flashed bar was displayed during the fast or slow phase of the nystagmus (i.e., the smooth movement of the eye to track the moving environment, and the saccade that occurs when the eye quickly rotates back to a previous position) [15]. When the flashed bar was displayed during the slow phase of the nystagmus, participants reported a shift of 5. However, when the flashed bar was shown between 100 ms prior to the saccade and the beginning of the saccade, participants reported shift magnitudes that were much smaller. Tozzi et al. [15] suggest that the reduction in shift magnitude was due to competing movement because the saccade was moving the eyes in the opposite direction as the black and white vertical bars. Additional research has provided evidence of a visual-motor map in addition to the retinocortical map [14, 17]. In particular, participants asked to complete similar tasks as in the study by Tozzi et al. [15] showed two different results based on whether they were asked to verbally indicate the position of the image that was presented to them after OKR stimulation or to point to it. Those asked to verbally indicate its location indicated a shift, but those asked to point to it 13

15 were able to correctly identify the location. Thus, a second uncorrupted map known as the visual-motor map was inferred [14, 17]. The VOR can be induced by placing an organism (or human) on a platform (or chair) and rotating it in complete darkness. Alternatively, the organism can be rotated inside of a surround with black and white vertical stripes and the ability to rotate at the same velocity as the organism is being rotated [18-20]. Both techniques result in angular acceleration, which stimulates the vestibular system leading to a VOR. When an individual is rotated in complete darkness, the VOR results in a gain of 0.6 rather than a gain of 1.0, which illustrates that the reflex cannot completely compensate for head movement on its own. If an individual is rotated in ambient light, the OKR is able to assist the VOR and a gain of 1.0 results showing that the two reflexes together are able to compensate completely for head movement stimulating an equal and opposite movement of the eyes [1, 18]. Ivanenko et al. [18] showed that participants receiving visual stimuli that conflicted with the vestibular input would rely more heavily on the visual input than vestibular input to interpret their movement through space. Participants were placed on a chair and were given goggles containing a screen that showed them a virtual square room with vertical black and white stripes. The participants were rotated 180, but the virtual room was only rotated 90. Participants did not report a discrepancy. When asked to report how much they were rotated, they reported approximately 90. This research further supports the importance of visual feedback in the oculomotor system. Eye Movements or their Equivalent in Various Taxa Because some organisms have limited eye mobility or cannot move their eyes at all, they have evolved an analogous ability to move their head or body to carry out the oculomotor movements that mobile-eyed animals perform [6, 21, 22]. For example when they are hunting 14

16 prey or feeding, birds such as terns and pigeons have limited eye mobility but use their head and neck movement to move the eyes in a way that mimics oculomotor movement in organisms with movable eyes [6, 23-25]. Additionally, honeybees placed in a rotating black and white vertically striped drum compensate for the drum rotations by moving their head and thorax in place of the oculomotor reflexive eye movements seen in humans [26]. Eye Movements in Uca pugilator Crustaceans, like many animals including humans, have independently evolved an oculomotor system composed of directed and stabilizing eye movements [4, 27-29]. There is also evidence that some crustaceans make similar perception errors as humans when their oculomotor reflexes are artificially stimulated. Fiddler crabs Uca pugilator and Uca rapax have been shown to fail to update their retino-cortical map when exposed to an OKR or VOR stimulus [4, 30]. Fiddler crabs received OKR or VOR stimuli and were allowed to return to their burrow, which they excavate and live in, after receiving the stimulus. Both species of fiddler crabs made errors approximately equal to the angular rotation when returning to their burrows suggesting that they did not update their subjective body axis to reflect adjustments made by the oculomotor reflexes. This study focuses on a fiddler crab, Uca pugilator, with tall eyestalks placed close together as opposed to other species of crab with short eyestalks spaced far apart on the carapace of the crab [29]. Similar to humans and other mammals, crabs rotate their eyes to maintain a stable visual field in order to compensate for motion due to their own body movements [28, 31]. These highly moveable eyes atop eyestalks make fiddler crabs ideal for this type of research. This study attempts to understand the role of directed eye movements (optokinetic and vestibulo-ocular) in the spatial perception of the fiddler crab, Uca pugilator to see if the motor-perceptual phenomena have evolved in parallel with that of other animals like the 15

17 neuro-muscular machinery. Fiddler crabs were exposed to oculomotor stimuli followed by a looming stimulus (i.e., a simulated predator) to invoke an escape response. If the oculomotor reflexes fail to update their equivalent of the retino-cortical map, their subjective body axis should shift by an amount similar to the angular eye rotation. Consequently, the escape direction will have a similar erroneous shift compared to the escape direction of control crabs not exposed to the OKR stimulus. Materials and Methods Male fiddler crabs, Uca pugilator, were collected from Carrot Island, Beaufort, NC, USA. Crabs were housed in the lab in a 1.2 m diameter arena filled with a sand mud mixture collected from Carrot Island. A simulated diurnal tidal cycle was created using a pump to move 25 ppt salt water made with Instant Ocean in and out of the arena twice daily. Crabs were fed ground tropical fish flakes (Omega One Goldfish Flakes) ad libitum. While housed in the arena, fiddler crabs behaved as normal by excavating burrows, foraging, and courting mates. Experimental Procedures A flag (6 mm x 1 mm) made of flexible paper was attached to each crab with a drop of superglue to the distal median end of the carapace covering the eyestalk. The flag extended over the dorsal carapace. Crabs were numbered for identification using nail polish on the dorsal carapace. After receiving an eye flag and number, crabs were allowed to acclimate for 24 h prior to testing. All crabs were placed individually inside of a glass bowl (8 cm base diameter) with a thin (approximately 1 cm) layer of sand covering the bottom. The sand layer was thin enough to prevent burrow excavation, which would have provided a visual reference point inside of the bowl. Ground tropical fish flakes (Omega One Goldfish Flakes) and a few drops of 1 M dextrose 16

18 made with 25 ppt salt water were placed in the center of the bowl. Food was used to encourage the crabs to stay in the center of the bowl, and a feeding response was used to indicate that the crabs had relaxed prior to beginning the experiment. The bowl was centered inside of a drum (32 cm diameter) with black and white vertical stripes placed equidistantly around the drum. A pico projector (Optoma pico pocket projector PK301) was placed above the drum at an angle allowing a looming stimulus (i.e., a repeating sequence of progressively larger concentric black dots) to be projected on the side of the drum at a height slightly above the bowl. Any object projected on a fiddler crab s eyes above the horizon of the eye would be interpreted as potentially threatening, and so arranging the looming stimulus in this location ensured the fiddler crabs would respond as though it was a threat [32, 33]. A Plexiglas bar pointed from the location of the looming stimulus towards the drum s center. A piece of red tape (1.3 cm length) was wrapped around the end of the bar and used for a distance scale during data analysis. The bowl, crab, and bar were recorded from above with a camcorder (Sony Handycam HD) mounted above the apparatus. Crabs were oculomotor stimulus followed by a looming stimulus or just a looming stimulus to investigate whether the escape direction differed when an oculomotor stimulus was presented (Figure I-1). 17

19 A. B. Figure I-1. Explanation of the hypothesis and methods using OKR treatment as an example. The blue circles in A and B represent the eye of a fiddler crab, Uca pugilator. (A) The black bird positioned on the eye represents the focal retinal receptor prior to OKR stimulation. When the crab receives the OKR stimulus, a black and white vertically striped drum is rotated clockwise around the crab (black and white arrow), and the eye follows it (blue arrow). (B) After the OKR stimulus, the eye has rotated in response to the OKR stimulus resulting in the position of the retinal receptor represented by black bird on the eye being rotated to a new position. The yellow bird on the eye indicates the position of the receptor prior to OKR stimulation. Following the OKR stimulus, the looming stimulus is projected in front of the crab (large black bird), and the image is formed on the retina on the focal receptor. If the crab s retino-cortical map updated to reflect the change in retinal receptor location, it will escape in the direction of the black arrow. If the crab s retino-cortical map did not update, it will escape in the direction of the yellow arrow indicating that it is not aware of eye movement (i.e., it perceived the looming stimulus as being located where the yellow bird is in figure B). 18

20 Optokinetic Stimulus All trials for the optokinetic stimulus were conducted in ambient fluorescent light. Individuals either received a looming stimulus only (LS treatment) or received an optokinetic stimulus followed by a looming stimulus (OKR treatment). Each animal in the LS treatment (n = 25) was placed inside of the bowl and allowed to acclimate indicated by the initiation of feeding and demonstration a relaxed body position. Following acclimation, the looming stimulus video was played to stimulate an escape response in the crab. Each OKR treatment animal (n = 25) was placed in the bowl and allowed to acclimate. Following acclimation, the crab received an optokinetic stimulus created by rotating the drum clockwise (xˉ = 0.08 ± 0.04 revolutions/s) around the bowl containing the crab until the crab s eye with the eye flag noticeably rotated indicating OKR stimulation. Following the optokinetic stimulus, the looming stimulus video was played to stimulate an escape response in the crab. Optokinetic and Vestibulo-ocular Stimuli Simultaneously All trials exposing animals to both OKR and VOR stimuli were conducted in low fluorescent light. Individuals either received a looming stimulus only (LS treatment) or a simultaneous optokinetic and vestibulo-ocular stimuli followed by a looming stimulus (OKR + VOR treatment). As in the previous set of experiments, each LS treatment animal (n = 26) was placed inside of the bowl and allowed to acclimate. Following acclimation, the looming stimulus video was played to stimulate an escape response in the crab. Each OKR + VOR treatment animal (n = 24) was placed in the bowl and allowed to acclimate. Following acclimation, the crabs received a vestibulo-ocular stimulus created by rotating the bowl clockwise (xˉ = 0.01 ± revolutions/s) while the drum remained stationary until the crab s eye with the eye flag noticeably rotated indicating a reflex stimulation. Because the 19

21 experiments were conducted in visible light, the animals also received an optokinetic stimulus produced by rotating the bowl, which caused the drum to appear to be rotating counter clockwise. Following the oculomotor stimuli, the looming stimulus video was played to stimulate an escape response. Video Analysis All trials were recorded using a Sony Handycam HD video recorder and analyzed in Matlab R2007b. For LS treatment analysis, one video frame was chosen before the introduction of the looming stimulus, one video frame was chosen when the first step was taken during the escape run response, and every fifth frame following the first step until the crab reached the edge of the test arena was chosen. For OKR treatment and OKR + VOR treatment analysis, one frame was selected before the oculomotor reflex stimulus, one frame was selected immediately before drum or bowl rotation began, two frames were chosen during drum or bowl rotation, the last frame of drum or bowl rotation was selected, the frame during which the first escape step was taken was selected, and a final frame when the crab reached the edge of the test arena was selected. Every fifth frame between the first step frame and the final frame were analyzed in addition to all frames selected prior to the frame containing the crab s first escape response step. Custom software (Matlab, The Mathworks, Inc., Natick, MA.) was used to calculate drum rotation speed or bowl rotation speed (revolutions/s), drum rotation or bowl rotation angle, eye rotation angle, body orientation angle prior to the looming stimulus, and crab escape direction. This was done by constructing a vector from the location of the looming stimulus through the center of the crab, which represented an ideal 180 escape path shown in Figure 2 as a dashed line originating at the looming stimulus and continuing through crab A. The body orientation angle of the crab relative to the 180 escape path was calculated prior to the 20

22 presentation of the looming stimulus to the crab (angle a in Figure I-2). Escape direction angle was calculated by using a linear regression of the x and y coordinates from every fifth frame during the crab s escape run. The regression line was used as the escape path vector, and the angle of escape was calculated as the angle between the escape path vector and the 180 vector. Eye rotation was calculated by finding the angle between a vector representing the initial position of the eye flag prior to the oculomotor reflex stimulus and a vector representing the position of the eye flag following the oculomotor stimulus. Data Analysis Matlab CircStats [34] and R (CRAN packages used: car, pwr, compute.es, psych, sciplot, and circular) [35-42] were used to analyze data. Correlations examining the relationship between body orientation and escape angle as well as eye rotation and escape angle were run. A onetailed t-test was used to compare mean escape angle of the LS treatment and OKR treatment groups and the LS treatment and OKR + VOR treatment groups. An ANCOVA model was used to examine the effect of treatment and body orientation on escape direction in the LS and OKR treatments and the LS and OKR + VOR treatments. A two-tailed t-test was used to compare LS and OKR treatment crabs body orientation prior to looming stimulus exposure. The body orientation of LS and OKR + VOR treatment crabs were also compared using a two-tailed t-test. A v-test was used to examine the distribution of the escape directions of LS treatment crabs, OKR treatment crabs, and OKR + VOR treatment crabs around a circle to determine whether they were randomly distributed. The v-test is given a angular parameter to compare the mean escape direction to in order to determine whether the mean is significantly similar to the given parameter. Mean body orientation and the predicted 180º escape direction were used as reference 21

23 parameters. A v-test was also used to examine the distribution of body orientation around a circle for all treatments. The predicted 180º escape direction was used as the reference parameter. B A a b Figure I-2. Method used to calculate angular data. The dashed line connecting the light grey oval (which represents the looming stimulus) through the center of crab A represents the predicted 180 escape path. Crab A represents the position of the crab following exposure to the oculomotor stimulus and prior to the looming stimulus. Angle a illustrates the calculation of the crab body orientation angle. Crab B represents the position of the crab following the escape. The dashed line connecting crab A to crab B represents the escape path vector, and angle b represents the escape direction angle. Results Optokinetic Stimulus Results The LS treatment and OKR treatment crabs showed a significant difference in escape direction following exposure to a looming stimulus (one-tailed t-test, DF = 48, t = 1.71, p = 0.047; Figure I-3). There was no significant correlation between the OKR treatment crabs escape direction and eye rotation resulting from the OKR stimulus (Pearson correlation, DF = 23, R 2 = 0.054, p = 0.263; Figure I-4A). Body orientation prior to exposure to the looming 22

24 stimulus was significantly correlated with the crab escape direction in both the LS treatment and OKR treatment groups (Pearson correlation, LS treatment: DF = 23, R 2 = 0.25, p = 0.011; OKR treatment: DF = 23, R 2 = 0.17, p = 0.042; Figure I-4B and Figure I-4C). An ANCOVA model was used to account for the effects of treatment, body orientation, and the interaction on crab escape angle. The interaction between body orientation and treatment was removed from the model because it was not significant. The results indicate that body orientation was the only factor with a significant effect on crab escape direction (ANCOVA body orientation: F 1,47 = 11.87, p = ; Table I-1). However, the treatment was very close to having a significant effect on escape angle direction (ANCOVA treatment: F 1, 47 = 3.57, p = 0.065). Body orientation of the LS treatment and OKR treatment crabs prior to the looming stimulus exposure was examined and no significant difference was found between the two treatment groups (two-tailed t-test, DF = 48, t = 1.21, p = 0.23). Escape direction for both the LS treatment and OKR treatment groups were nonrandomly spread around a circle and were significantly related to the predicted 180 escape direction and the group s mean body orientation (v-test, escape referenced to predicted 180 OKR treatment: N = 25, v = 21.07, p < ; LS treatment: N = 25, v = 21.93, p < ; escape referenced to mean body orientation OKR treatment: N = 25, v = 19.66, p < ; LS treatment: N = 25, v = 20.30, p < ). A v-test exploring the distribution of both the LS treatment and OKR treatment groups body orientation showed non-random distribution and a 23

25 significant relationship between body orientation and the predicted 180 escape direction (v-test, OKR treatment: N = 25, v = 13.08, p < ; LS treatment: N = 25, v = 16.36, p< ). A. Escape Angle (deg) LS Treatment OKR B LS Treatment Mean LS Treatment OKR Treatment Mean OKR Treatment Figure I-3. Escape run directions of looming stimulus (LS) and optokinetic response (OKR) treatment crabs. Uca pugilator were exposed to a looming stimulus if they were in the LS treatment and an optokinetic stimulus followed by a looming stimulus if they were in the OKR treatment. The looming stimulus triggered an escape run response in the crabs. The escape response was significantly different between the two groups (one-tailed t-test, DF = 48, t = , p = 0.047). Mean escape directions are shown in A for each treatment. Individual crab escape directions as well as the mean for each treatment group are shown in B using a polar plot.

26 Escape Angle (deg) A. y=0.56x ; r^2= Eye Rotation (deg) Escape Angle (deg) B. y=0.19x ; r^2=0.17 y=0.29x ; r^2=0.25 Escape Angle (deg) C Body Orientation (deg) Body Orientation (deg) Figure I-4. Eye rotation and body orientation correlated with escape direction in the looming stimulus (LS) and optokinetic response (OKR) treatments. Fiddler crabs in the OKR treatment group received an OKR stimulus, which resulted in the rotation of their eyes. The OKR stimulus was followed by a looming stimulus, which evoked an escape response. There was no significant correlation between escape direction and eye rotation angle in the OKR treatment crabs (A.; Pearson correlation, DF = 23, R 2 = 0.054, p = 0.26). Fiddler crabs in the LS treatment were exposed to the looming stimulus, but they were not exposed to the OKR stimulus. There was a significant correlation between body orientation prior to the looming stimulus and escape direction in both the OKR (B) and LS (C) treatment groups (Pearson correlation, OKR treatment: DF = 23, R 2 = 0.17, p = 0.042; LS treatment: DF = 23, R 2 = 0.25, p = 0.011). The dashed grey lines in all figures show 95-percentile confidence intervals in all graphs. 25

27 Table I-1. Effect of treatment and body orientation on escape direction in the OKR treatment. An ANCOVA model was used to examine the effect of body orientation, treatment, and the interaction on escape direction in fiddler crabs, Uca pugilator, exposed to an optokinetic stimulus and a looming stimulus in the OKR treatment and exposed to a looming stimulus in the LS treatment. The interaction was not significant and was removed from the model. Body orientation was the only factor that had a significant effect on escape direction (p = ) DF SS MSS F-value p-value Treatment Body Orientation ** Residuals Totals *<0.05, **<0.01, ***<0.001 Optokinetic and Vestibulo-ocular Stimuli The LS treatment and OKR + VOR treatment crabs showed no significant difference in escape direction following exposure to a looming stimulus (one-tailed t-test, DF = 48, t = -1.43, p = ; Figure I-5). There was no significant correlation between the OKR + VOR treatment crabs escape direction and eye rotation resulting from the OKR and VOR stimuli (Pearson correlation, DF = 22, R 2 = , p = 0.74). Body orientation prior to looming stimulus exposure was significantly correlated with the crab escape direction in both the LS treatment and OKR + VOR treatment groups (Pearson correlation, LS treatment: DF = 24, R 2 = 0.50, p < ; OKR + VOR treatment: DF = 22, R 2 = 0.45, p = ; Figure I-5). An ANCOVA model was used to account for the effects of treatment, body orientation, and the interaction on crab escape angle. The interaction was not significant and was removed from the model. The results indicate that body orientation was the only factor that had a significant effect on crab escape direction (ANCOVA body orientation: F 1, 47 = 41.84, p = 7.2 x 10-8 ; Table I-2). However, treatment was close to having a significant effect on escape direction (ANCOVA treatment: F 1, 47 = 3.79, P = 0.058). Body orientation of the LS treatment and OKR + VOR treatment crabs prior to looming stimulus exposure was examined and no 26

28 significant difference was found between the two groups (two-tailed t-test, DF = 48, t = , p = 0.99). Escape directions for both the LS treatment and OKR + VOR treatment groups were nonrandomly spread around a circle and were significantly related to the predicted 180 escape direction and the treatment group s mean body orientation (v-test, escape referenced to predicted 180 OKR + VOR treatment: N = 24, v = 18.98, p < ; LS treatment: N = 26, v = 23.09, p < ; escape referenced to mean body orientation OKR + VOR treatment: N = 24, v = 19.85, p < ; LS treatment: N = 26, v = 23.13, p < ). A v-test exploring the distribution of both the LS treatment and OKR + VOR treatment groups body orientation showed non-random distribution and a significant relationship between body orientation and the predicted 180 escape direction (v-test, OKR + VOR treatment: N = 24, v = 14.37, p < ; LS treatment: N = 26, v = 17.71, p < ). 27

29 A. Escape Angle (deg) LS Treatment OKR + VOR B LS Treatment Mean LS Treatment OKR + VOR Treatment Mean OKR + VOR Treatment Figure I-5. Escape run directions of looming stimulus (LS) and optokinetic response + vestibulo-ocular response (OKR +VOR). Uca pugilator were exposed to a looming stimulus following simultaneous exposure to an optokinetic stimulus and a vestibulo-ocular reflex in the OKR + VOR treatment. Crabs in the LS treatment received the looming stimulus, but they did not receive the oculomotor reflex stimuli. The looming stimulus triggered an escape response in the crabs. The direction of the escape response was not significantly different between the two treatment groups (one-tailed t-test, DF = 48, t = -1.43, p = 0.92). Mean escape directions are shown in A. Individual escape directions as well as the mean for each group are plotted as a polar plot in B.

30 Escape Angle (deg) A. y = 0.28x ; r^2 = Eye Rotation (deg) Escape Angle (deg) B. y = 0.43x ; r^2 = 0.45 y = 0.40x ; r^2 = 0.50 Escape Angle (deg) C Body Orientation (deg) Body Orientation (deg) Figure I-6. Eye rotation and body orientation correlated with escape direction in the looming stimulus (LS) and optokinetic response + vestibule-ocular response (OKR + VOR) treatments.. Fiddler crabs in the OKR + VOR treatment group received OKR and VOR stimuli simultaneously, which resulted in the rotation of their eyes. The OKR + VOR stimuli were followed by a looming stimulus, which evoked an escape response. There was no significant correlation between escape direction and eye rotation angle in the OKR + VOR treatment group (A.; Pearson correlation, DF = 22, R 2 = 0.081, p = 0.18). Fiddler crabs in the LS treatment were exposed to the looming stimulus, but they were not exposed to the OKR + VOR stimuli. There was a significant correlation between body orientation and escape direction in both the OKR + VOR (B) and LS (C) treatments (Pearson correlation, experimental group: DF = 22, R 2 = 0.45, p= ; control group: DF = 24, R 2 = 0.50, p = ). The dashed grey lines in all figures show 95-percentile confidence intervals in all graphs. 29

31 Table I-2. Effect of treatment and body orientation on escape direction in looming stimulus (LS) and optokinetic response + vestibule-ocular response (OKR + VOR) treatments. An ANCOVA model was used to examine the effect of body orientation, treatment, and the interaction on escape direction in fiddler crabs, Uca pugilator, exposed to an optokinetic stimulus and a looming stimulus in the OKR treatment and exposed to a looming stimulus in the LS treatment. The interaction was not significant and was removed from the model. Body orientation was the only factor that had a significant effect on escape direction (p = 7.2 x 10-8 ). DF SS MSS F-value p-value Treatment Body Orientation x10-8 *** Residuals Totals *<0.05, **<0.01, ***<0.001 Discussion Uca pugilator exposed to an optokinetic stimulus responded to a looming stimulus with a significantly different escape direction than fiddler crabs that were exposed to the same looming stimulus but were not exposed to the optokinetic stimulus (Figure I-3). In both groups, the looming stimulus triggered an escape response that resulted in the fiddler crabs running away from the looming stimulus. The LS treatment group escape direction was similar to the 180 predicted escape direction. However, the shift in escape run direction in the experimental group deviated from the predicted 180 escape direction in the predicted direction based on the clockwise rotation of the black and white striped drum providing the optokinetic stimulus. This result is similar to those found by Layne et al. [4] and Layne et al. [30], which demonstrated that fiddler crabs made a homing error when trying to return to their covered burrow after receiving an OKR or VOR stimulus. Layne et al. suggested that the optokinetic stimulus resulted in a shifting of the fiddler crabs subjective body axis with no update to the retino-cortical map. This resulted in the homing error observed. However, in this study, fiddler crabs exposed to the vestibulo-ocular stimulus and the optokinetic stimulus simultaneously did not show a significant difference in escape run direction compared to the control crabs (Figure I-5). Nevertheless, 30

32 Figure I-4B differs from Figure I-4C in qualitatively similar way to how Figure I-6B differs from Figure I-6C. In both cases, LS treatment escape-direction body-orientation correlations show data distributed uniformly around the line of best fit. However, OKR treatment and OKR + VOR treatment escape-direction body-orientation correlations show the data have a conical spread with the smallest portion of the cone near 0 and spreading out from there. This suggests that there is some effect of the oculomotor stimuli on escape direction. This is further suggested by the ANCOVAs for both the OKR treatments and the simultaneous OKR and VOR treatments. Although treatment (control or experimental) did not have a significant effect in the ANCOVA models, it was close to significant ( Table I-1 and Table I-2). It might be expected that eye-rotation angle resulting from the oculomotor stimulus will be correlated with the angular shift in escape run direction. However, thought we did see a shift in escape direction there was no significant correlation between eye rotation angle and escape direction in the experimental group receiving the optokinetic stimulus (Figure I-4A). The fiddler crabs receiving both oculomotor stimuli did not show a shift in escape direction nor was there a significant correlation between escape direction and eye rotation in the experimental group (Figure I-6). The fiddler crabs in the control and experimental groups for both portions of this study did show a significant correlation between body orientation and escape direction suggesting that body orientation played a greater role in determining escape direction than any other factor in this study (Figure I-4 and Figure I-6). The v-test results showed that the escape direction of the fiddler crabs was not significantly different than the predicted 180 escape 31

33 direction nor the crabs body orientations. This may suggest that there was a large variation in body orientations and escape directions, which prevented the mean escape direction from being distinguished from either body orientation or the predicted 180º escape direction. It is also possible that the setup needs to be redesigned in a way that allows the body orientation of the fiddler crabs to be controlled. It is important to take into account the literature that suggests that there are two maps, visual-motor and retino-cortical (cognitive), that receive visual information [14, 16, 17]. It is possible that the escape response of the fiddler crabs, running away, is similar to the act of pointing at a target in humans, which has been shown to use the visual-motor map and fail to perceive a shift in perception of the environment resulting from stimulation of oculomotor reflexes [14, 16, 17]. The oculomotor reflexes do cause a shift in the perception map, which human participants indicate by verbally stating the location of a target [14-17]. The difference in escape direction between the control and experimental groups of crabs receiving only the OKR stimulus did show a significant difference in escape direction, but there was no correlation with eye rotation angle in the experimental group. These results indicate that in this case the stimulus did not update the retino-cortical map because there was a significant shift in escape direction, but it is not clear why there is no correlation between eye rotation angle and escape run direction. There was a strong correlation in the experimental and control groups between body orientation and escape direction. It may be that the fiddler crabs deferred to running in the direction their body was oriented because they did not have a home burrow that they were targeting. In the absence of an end goal for the run, they ran in the most convenient direction to escape looming stimulus. In work with blue crabs, Woodbury [43] suggested that there are four general strategies for avoiding a potential predator: running directly away from the predator, 32

34 running towards a target (e.g., a fiddler crab s burrow), running in a direction that is between the target and directly away from the predator, and running in a completely random direction. Blue crabs were shown to use physical characteristics (e.g., change in water color indicating direction of movement needed to move off of the beach and into deeper water) to determine which way to run in the presence of a predator and tended to run in a direction that was a compromise between directly away from the predator and towards a target [43, 44]. It is reasonable then that fiddler crabs would do the same and will run directly away from the looming stimulus in the absence of a target burrow. Because a shift in subjective body axis may affect target acquisition but not visual-motor map, we would thus not see a correlation between eye rotation angle and escape direction. The study results suggest that optokinetic reflexes do not update the retino-cortical map and results in a shift in subjective body axis, but the crabs default to running in a direction that is correlated to initial body orientation in the absence of a target to run towards. However, there was no significant difference between the escape directions of the control and experimental groups receiving both the VOR and OKR stimulus. It is possible that the stimuli these crabs received needed to be stronger in order to elicit a difference in escape direction. Ensuring that angular eye rotation resulting from the stimuli is sufficiently large would be an important change to make if this experiment were to be repeated. Further research administering only a VOR stimulus in complete darkness is needed to see if the VOR has a significant effect on escape direction, which would indicate whether VOR eye movements update the retino-cortical map of fiddler crabs. 33

35 II. The role of gravity in inducing compensatory eye movement in the vertical plane in the fiddler crab Uca pugilator Introduction Decapods use statocysts, gravireception (i.e., gravity sensing) organs, to sense their body position and movement in space. These organs are analogous to the human vestibular system and can sense movement relative to gravity in the pitch and roll planes as well as angular acceleration [45, 46]. Each statocyst is an invagination located in the base of an antennule in most decapods, and it is composed of vertical and horizontal fluid-filled canals, sensory hairs, and a statolith (i.e., a ball of sand granules) [47-50]. The vertical and horizontal canals form connected, orthogonal tori, and the statolith is positioned at the base of the vertical canal atop a sensory cushion composed of statolith hairs [49]. The information gained from the statocyst is used to right the body and for oculomotor control. The statolith hairs in brachyuran (crab) statocysts form two, crescent-shaped rings on the base of the vertical canal, and the tips of the hairs are cemented to the statolith, which rests on top of them, with a secretion from the tegumental glands (Figure II-1A) [47, 49, 50]. The statolith hairs are composed of a hollow shaft that is feathered on the tip in some crustaceans, and the tip bends toward the center of the crescent in a manner that resembles a hook (Figure II-1B) [47, 51-53]. The base of the shaft attaches to a bulbous structure referred to as the ampulla, which is composed of the fulcrum, tooth, and lingual. These structures serve as a hinge and enable the hair cell to bend when shear force (i.e., a force acting perpendicular to the hair) is applied to it. When decapods tilt their bodies in the pitch or roll plane, the force of gravity acting 34

36 on the statoliths causes them to move and imposes shear force [47]. That force is transduced to the receptor cell in the form of tension by way of the chorda, a structure that is attached to the inside of the hair shaft [47, 54]. A. B. Statolith Figure II-1. The Structure of a decapod statocysts and statolith hair. A. is a visual representation of a brachyuran (crab) statocyst as presented by Sandeman and Okajima [45]. The statocyst is composed of two circular canals, which indicate angular acceleration, and the statolith, which indicates movement in the pitch and roll planes (grey arrow). B. illustrates the structure of a single statolith hair and receptor cell of a brachyuran as presented by Cate and Roye [53]. The statolith hair (grey arrow) is bent by shear force applied to it by the statolith during movement in the pitch and roll planes. The hair pulls on the chorda (black arrow), which stimulates the receptor cell (black and white dashed arrow) to fire. The bipolar receptor cells of the statocysts are directionally sensitive to the force being applied to the statolith hairs [45]. The two rows of statolith hairs produce the highest firing rates in the receptor cells when they are deflected perpendicular to the rows of statolith hairs and do not always fire when deflected parallel to the rows [53, 55]. The receptors for the inner row of crab statolith hairs respond to shear force with a phasic firing pattern that is directionally sensitive (i.e. more firing occurs in perpendicular hair displacement). The rate of receptor cell adaptation is dependent on the intensity of the stimulus and corresponds with amplitude of the force. The receptors for the outer row of statolith hairs have a tonic firing pattern and respond strongly to vibration in addition to shear force caused by statolith movement [55]. The statolith 35

37 hair cells have been shown to have a preferred orientation with respect to gravity at which the receptor cells fire most rapidly, and when the cell moves to either side of the preferred orientation, the firing rate decreases [51]. As the statolith hairs in both rows return to their normal resting position after stimulation, the cell immediately returns to its resting potential [45, 51]. The statolith and statolith hairs, specifically the inner row of hairs, have been shown to serve as position indicators for crabs and other decapods [46, 55, 56]. Prawns and crayfish with normally non-magnetic statoliths replaced with ferrite filings have been shown to adjust their body position based on the position of a magnet moved around them to simulate gravity [56, 57]. Because the only part of the statocyst that could be responding to the magnet is the crescent of statolith hairs, these studies provided evidence for the role of statolith hairs as body position and equilibrium sensors. Additionally, Ozeki [58] was able to record from statolith hairs while moving a magnet about a crayfish statocyst in four planes. While doing so, he noted that some statolith hairs fired in multiple directions while other hairs were unidirectional, and some hairs demonstrated inhibitory responses in one or more directions to statolith hair deflection. These findings were in agreement with similar findings by Cohen [48] in the American lobster, Homarus americanus. While recording from the statolith hairs and the eye cups of the American lobster, Patton and Grove [59] were able to see evidence of the lobsters combining signals from multiple statolith hair neurons to determine eye movement. As they stimulated the statolith hairs by manipulating the statolith in the pitch and roll plane, they observed compensatory eye cup movements that corresponded to the way the lobster eyes would move if it was actually being rotated in the pitch or roll plane. This work provides additional evidence that decapods use statoliths and statolith hairs to orient themselves or at least their eyes in space. 36

38 When crustaceans molt, the statolith is shed and replaced with sand granules from the environment, which are cemented together by the tegumental gland to form a ball [60]. Crustaceans have been successfully induced to replace the statolith with magnetic filings during molt in place of the sand granules, and the sense of equilibrium and position can then be magnetically manipulated causing reorientation based on the summed gravitational and magnetic forces [56, 57]. In addition to reorienting their bodies, crustaceans have been observed to reorient their eyestalks to compensate for body tilt after statocyst stimulation [47, 61]. Ecdysis (i.e., molting) in crustaceans is inhibited by a molt-inhibiting hormone (MIH) produced by the X-organ, which is located in the eyestalk of the crustacean [62-64]. Moltinhibiting hormone not only limits production of molting hormone, but it also reduces the responsivity of the tissue to any molting hormone that is produced [65]. When the nervous system signals the X-organ to stop producing MIH, the Y-organ begins producing molting hormone (MH), which initiates proecdysis, the premolt stage. The molting hormone produced by the Y-organ is α-ecdysone, which is then converted to 20-hydroxyecdysone [64]. Removal of the eyestalks, which removes the X-organ, allows the initiation of proecdysis as if the nervous system blocked the production of MIH [62, 63]. Other approaches to inducing precocious ecdysis include leg removal, injections of molting hormones (α-ecdysone and 20-hydroxyecdsone), and exogenous exposure to molting hormones [66-69]. This study aims to induce fiddler crabs, Uca pugilator, to replace their normally nonmagnetic statoliths with ferrite filings using exogenous exposure to molting hormones (α-ecdysone and 20-hydroxyecdsone) while in the presence of only the magnetic filings for statolith replacements. Because α-ecdysone is converted to 20-hyroxyecdyone, it is 37

39 expected that fiddler crabs exposed to 20-hydroxyecdysone will molt sooner than those exposed to α-ecdysone. It is expected that both groups will molt sooner than the control group. Following ecdysis, fiddler crab statoliths will be manipulated using a magnet to illustrate the effectiveness of metal filings as statolith replacements. Manipulating the statoliths with metal filings will cause stimulation of statolith hairs and result in vertical compensatory eye movements. It is expected that stimulation of the statolith will result in compensatory eyestalk movement that correlates with the motion of an external magnet. Materials and Methods Male fiddler crabs, Uca pugilator, were collected from Carrot Island, Beaufort, NC, USA. Crabs were housed in the lab in a 1.2 m diameter arena filled with a sand mud mixture collected from Carrot Island. A simulated diurnal tidal cycle was created using a pump to move 25 ppt brackish water made with Instant Ocean in and out of the arena twice daily. Crabs were fed ground tropical fish flakes (Omega One Goldfish Flakes) ad libitum. While housed in the arena, fiddler crabs behaved as normal by excavating burrows, foraging, and courting mates. Experimental Procedures for Inducing Precocious Ecdysis Forty-five culture dishes (11.43 cm diameter, Carolina Biological Supply Company) were used to house fiddler crabs, Uca pugilator, individually (carapace width xˉ = mm) and were divided into three treatment groups of 15 bowls each. 30 g of steel filings (from Alfa Aesar, type 410-L) were placed inside of each bowl along with 225 ml of 25 ppt brackish water (Instant Ocean). The α-ecdysone treatment had the α-ecdysone molting hormone added to the bowl. The 20-hydroxyecdysone had the molting hormone 20-hydroxyecdysone added to the bowl. The no 38

40 hormone treatment did not have any molting hormones added to the bowl and received additional water in place of the hormones. Stock solutions for the three groups were made as follows. Each crab in the no hormone treatment received 225 µl of stock solution made composed of 3 drops of 100% ethanol (ETOH), and 3375 µl of 5 % ETOH. Each crab in the α-ecdysone treatment received 225 µl of the stock solution, which is made by adding three drops of 100% ETOH to 900 µg ± 1 µg of α- ecdysone to put it into solution µl of 5% ETOH was added to the solution to make the treatment stock solution. Each crab in the 20-hydroxyecdysone treatment received 225 µl of the stock solution, which is made by adding three drops of 100% ETOH to 900 µg ± 1 µg of 20- hydroxyecdysone to put it into solution µl of 5% ETOH was added to the solution to make the treatment stock solution. Crabs were fed four pellets of sinking tropical fish food (Omega One Goldfish Flakes) and aerated daily. Bacteria and excess food were removed from the bowls as needed. Water was changed every four days and new stock solutions were made at that time. Dates of molting and death events were recorded. Crabs that molted were left in the bowl for 24 h following completion of ecdysis and then were moved to an aquarium containing sand and 25 ppt brackish water (Instant Ocean). After a crab molted or died, it was replaced with a new crab. Nineteen fiddler crabs out of 98 molted during the course of this study. The remaining 79 fiddler crabs died during the course of the study. The number of days it took for each crab to molt or die following initial exposure to the hormone treatments was recorded. 39

41 Experimental Procedures for Vertical Compensatory Eye Movements The experimental apparatus is shown in Figure II-3. A round drum 21 cm in diameter and 22 cm in height with black and white stripes placed equidistantly along its height was laid on its side and attached to a hollow steel rod that was inserted through the drum s base and extended into its center (Figure II-3). A second hollow steel rod was inserted inside the first rod and attached to one end of a thin metal bar. The thin metal bar was bent in a 90 angle, and the end that was not attached to the steel rod was attached to a bar magnet with a previously characterized magnetic field 1. A third hollow steel rod was inserted inside the second rod with a platform made of acrylic and covered with sandpaper attached to its end. The third steel rod remained stationary while the first and second rods were able to be rotated independently using stepper motors controlled by a custom Matlab R2007b GUI. A fiddler crab (N = 1) was affixed by his carapace with superglue to a small strip of transparency plastic which was positioned above the platform and attached to a small wooden dowel rod. The dowel rod rested loosely inside of two thin pieces of wood in such a way that the crab s legs touched the platform in a normal standing posture and he was supporting his own weight. Not all crabs survived long enough after statolith replacement to be tested in the drum apparatus. Statocysts were stimulated by using the stepper motor (step size = 1.8 ) to oscillate the magnet in the roll plane beneath the crab in a sinusoidal profile with 30 amplitude from the vertical axis and a frequency of 0.25, 0.50, 0.75, or 1.0 Hz. The bar magnet used for all trials was calibrated so that the placement of the magnet (3.4 cm from statocyst) was such that the magnetic force attracting the statolith to it would be approximately equal in magnitude to the 1 In particular, a separate procedure was used to determine the distance at which the force on metal filings used in the experiment was exactly balanced by the force of gravity. The magnet was then spaced that distance below the platform in this apparatus. 40

42 force of gravity on the statolith. Because the two forces acting on the statolith have equal magnitudes, the resultant simulated gravity force shown alongside the tail-to-tip addition of the two vectors forms an isosceles triangle. Consequently, the simulated gravitational force on the crab acts at an angle away from the vertical axis that is half of the angle that the magnet is positioned from the vertical axis at any given time during oscillation (Figure II-4). Trials were recorded in low visible light with a Sony Handycam HD camcorder for later analysis. Figure II-2. Drum used for gravireception trials. A round drum 21 cm in diameter and 22 cm in height with black and white stripes placed equidistantly along its height was laid on its side. A fiddler crab, Uca pugilator, was placed on an acrylic platform covered with a piece of sandpaper. The fiddler crabs was secured on the platform using a drop of superglue on the dorsal side of the carapace, which attached it to a thin piece of transparency. The transparency was secured to a wooden dowel rod that rested loosely in between two thin pieces of wood. This method of securing the crab allowed it to support its own weight and stand on the platform. The magnet was oscillated beneath the crab at a set frequency and amplitude during each trial. 41