View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Scholar Commons | University of South Florida Research

University of South Florida Scholar Commons

Graduate Theses and Dissertations Graduate School

5-21-2004 The Organization of the Visual System in the Bonnethead Shark (Sphyrna tiburo) Amy L. Osmon University of South Florida

Follow this and additional works at: https://scholarcommons.usf.edu/etd Part of the American Studies Commons

Scholar Commons Citation Osmon, Amy L., "The Organization of the Visual System in the Bonnethead Shark (Sphyrna tiburo)" (2004). Graduate Theses and Dissertations. https://scholarcommons.usf.edu/etd/1188

This Thesis is brought to you for free and open access by the Graduate School at Scholar Commons. It has been accepted for inclusion in Graduate Theses and Dissertations by an authorized administrator of Scholar Commons. For more information, please contact [email protected].

The Organization of the Visual System in the Bonnethead Shark (Sphyrna tiburo)

by

Amy L. Osmon

A thesis submitted in partial fulfillment Of the requirements for the degree of Cognitive and Neural Sciences Department of Psychology College of Arts and Sciences University of South Florida

Major Professor: Toru Shimizu Sarah Partan, Ph.D. Robert Hueter, Ph.D. Cynthia Cimino, Ph.D.

May 21, 2004

Keywords: Bonnethead, shark, ganglion, vision, retina

© Copyright 2004, Amy L. Osmon

Table of Contents

List of Tables ii

List of Figures iii

Abstract iv

Overview 5

Organization of retinal ganglion cells in non-shark species 6

Illumination 9

Habitat 9

Behavior 10

Organization of retinal ganglion cells in sharks 12

Illumination 13

Habitat 14

Behavior 14

Possible retinal ganglion cell topography of the bonnethead shark 16

Methods 17

Topographic mapping of retinal ganglion cells 18

Retrograde labeling of retinal ganglion cells 19

Data Analysis 20

Results 22

Discussion 35

References 42

Appendix A 50

i List of Tables

Table 1 Comparison of shark species from varied habitats 8

Table 2 Retinal ganglion cell counts and statistics 25

ii List of Figures

Figure 1. Typical photographic image with retinal ganglion cells 22

Figure 2. Case 2627-1 26

Figure 3. Case 2626-4 27

Figure 4. Case 2626-5 28

Figure 5. Case 2626-8 29

Figure 6. Case IMF1 30

Figure 7. Case IMF2 31

Figure 8. Case 2625-2 32

Figure 9. Case 2626-6 33

Figure 10. Case 2627-1 50

Figure 11. Case 2626-4 51

Figure 12. Case 2626-5 51

Figure 13. Case 2626-8 52

Figure 14. Case IMF1 52

Figure 15. Case IMF2 53

Figure 16. Case 2626-5 53

Figure 17. Case 2626-6 54

iii

The Organization of the Visual System in the Bonnethead Shark (Sphyrna tiburo)

Amy L. Osmon

ABSTRACT

The goal of this project was to examine the visual system of the bonnethead shark (Sphyrna tiburo). The eyes of this shark are located at the extreme lateral ends of a broad, elongated cephalofoil. Better understanding of their visual system may aid in determining the adaptive benefits of their usual head shape. The proposed project examined one specific aspect of their visual system: the organization of retinal ganglion cells and identification of areas of increased resolution. Two experiments were conducted to realize these aims: (1) staining of retinal ganglion cells, to examine their distributional pattern, and (2) retrograde staining of retinal ganglion cells to determine morphology.

iv Chapter One

Introduction

Elasmobranchs had long been thought to possess poor vision and utilize other sensory systems for navigation and detection of both prey and predators (Gruber, 1977). However, during the 1960’s and 1970’s the scientific community began to publish both anatomical and physiological research indicating that the visual systems of sharks and rays were capable of higher visual resolution than previously believed (Ali and Anctil, 1974a; Gruber, 1977; Gruber,

Gulley, and Brandon, 1975; Gruber, Hamasaki, and Bridges, 1963; Hamasaki and Gruber, 1965;

Stell, 1972; Stell and Witkovsky, 1973). Concurrent studies also attempted to test the visual acuity and learning ability of sharks by investigating visually mediated behaviors (Graeber, 1978;

Tester and Kato, 1966; Wright and Jackson, 1964). These studies revealed complexity within sharks’ visual system and provided insight into the significance of the visual system in their daily existence.

Research regarding retinal anatomy in sharks has shown variability in rod-to-cone ratios, the distribution of ganglion cells within the retina, as well as the presence of visual streaks

(Bozzano and Collin, 2000; Gruber, 1977; Gruber et al., 1975; Gruber et al., 1963; Hamasaki and

Gruber, 1965; Hueter, 1988; Peterson and Rowe, 1980; Stell, 1972; Stell and Witkovsky, 1973).

The implications of these retinal variations have not yet been thoroughly explored. However, several authors (Bonazzo and Collin, 2000; Gruber et al., 1975; Hueter, 1989; Hueter and Gruber,

1982) have related the variability of sharks’ visual systems to their feeding behaviors and habitats. This study examined the retinal anatomy of one shark species, the bonnethead shark.

The bonnethead shark is one of nine species in the Sphyrnidae family, commonly referred to as “hammerhead” sharks, possessing a broad and elongated head shape. The bonnethead shark inhabits clear to turbid inshore waters of the Gulf of Mexico, the Atlantic Ocean, and along the coasts of Central and South America (Cortes and Parsons, 1996; Hoese and Moore, 1958). Their

5 diet consists of a variety of swift-moving crabs and cephalopods (Cortes, Manire, and Hueter,

1996; Motta and Wilga, 2000). As this shark species adjusts well to captivity (Cortes et al.,1996;

Cortes and Parsons, 1996) and their ecology is well documented (Cortes et al., 1996; Cortes and

Parsons, 1996; Hoese and Moore, 1958; Myreberg and Gruber, 1974), it is an excellent subject for an investigation into the relationship between retinal anatomy and ecological niche.

Theories regarding the function of Sphyrinidae sharks’ unique cephalofoil focus on their head shape providing increased hydrodynamic lift, an area useful for capture of large prey items, and/or an enlarged area for electroreception and olfaction (Antcil and Ali, 1976; Compagno,

1984; Johnsen and Teeter, 1985; Kajiura and Holland, 2002; Martin, 1993; Nakaya, 1995; Strong,

Gruber, and Snelson, 1990). Only two studies have examined the visual system of Sphyrinidae sharks (Anctil and Ali, 1974b; Gruber et al., 1963). Anctil and Ali (1974b) investigated the retinal morphology of the scalloped hammerhead shark (Sphyrna lewini) and designated some retinal ganglion cells as giant ganglion cells due to their large soma size. Gruber et al. (1963) reported similarities in the morphology of cones between the lemon shark (Negaprion brevirostris) and the great hammerhead shark (Sphyrna mokarran). However, no research has been conducted regarding the visual system of the bonnethead shark. Information pertaining to the bonnethead shark’s visual system may lead to a better understanding of the true function and significance of this shark’s unusual head shape compared to other shark species.

Organization of retinal ganglion cells in non-shark species

How does retinal cell topography relate to the diverse habitats of different vertebrate species? According to Hughes’ (1977) terrain theory, animals with a predominantly two- dimensional horizon in their visual environment (e.g., the ocean with a sand-water boundary for a benthic aquatic species) gain an advantage from possession of a visual streak. A visual streak is an elongated area of increased ganglion cell density relative to other areas within the retina

6 (Bozzano and Collin, 2000). The advantage of possessing a visual streak is higher resolving power in the visual fields corresponding to the area of increased cell density (Bozzano and Collin,

2000). A visual streak may also negate the necessity of utilizing distinctive eye movements while an animal is gazing across an expansive horizon (Collin and Pettigrew, 1988b).

Hughes’ theory (1977) has been tested and validated in many non-aquatic vertebrates including the fat-tailed dunnart (Sminthopsis crassicaudata) (Arrese, Dunlop, Harman,

Braekevelt, Ross, Shand, and Beazley, 1999), several ungulates (e.g. the pig, sheep, ox, dog, and horse) (Hebel, 1976), the tammar wallaby (Macropus eugenii) (Wong,Wye-Dvorak, and Henry,

1986), and the African elephant (Loxodonta africana) (Stone and Halasz, 1989). Visual streaks have also been found in teleosts occupying open areas with a distinct visual horizon such as the blue tuskfish (Choerodon albigena), red-throated emperor (Lethrinus chrysostomas), collared sea bream (Gymnocranius bitorquatus), clown triggerfish (Balistoides conspicillum), and painted flutemouth (Aulostoma chinensis) (Collin and Pettigrew, 1988b). In addition, Hughes’ theory has been applied to other marine animals, including the loggerhead (Caretta caretta), leatherback

(Dermochelys coriacea), and green (Chelonia mydas) turtles (Oliver, Salmon, Wyneken, Hueter, and Cronin, 2000), and marine mammals including the sea otter (Enhydra lutris) (Mass and

Supin, 2000). However, there is scant information regarding elasmobranch species (Bozzano and

Collin, 2000; Hueter, 1989; Peterson and Rowe, 1980).

7 Table 1: Comparison of elasmobranch species

Environmental Factors Behavioral Visual Streak Factor Illumination Habitat and Foraging Width Length Location depth style Tiger shark Fair Pelagic Biting/ Narrow Not across Ventral to 0-140m bump and entire retina retinal bite meridian Epaulette Not Benthic N/A Broad Across most Central shark recorded 0-50m of retinal meridian Small- Dim Benthic N/A Narrow Across most Just dorsal of spotted 50-400m of retinal the meridian dogfish meridian shark Black- Dim Bentho- N/A Broad Across entire Central mouth pelagic retinal dogfish 300-2,000m meridian shark Velvet- Dim Meso- N/A Narrow Across entire Central belly shark pelagic retinal 500-2,000m meridian Lemon Fair to Benthic Ram-feeder Broad Across most Central shark Good 0-90m of retinal meridian Bigelow’s Not Benthic Suction Narrow Across most Dorsal of the ray recorded 650-2200m feeder of retinal meridian meridian (table adapted from Bonazzo et al., 2000; Hueter, 1991; Compagno, 1984; Motta, Tricas, Hueter, and Summer, 1997)

There are multiple factors which appear to relate to the width, length, and location of the visual streak (Bozzano and Collin, 2000; Hueter, 1991). These factors include: (1) the amount of light available to an animal, (2) the habitat of the animal (e.g., whether the horizon is completely open or partially obstructed), and (3) how an animal utilizes its visual streak, which includes foraging and prey detection strategies.

8 Illumination

In regards to the variability of illumination, this factor appears to relate to the density of the ganglion cells within the visual streak and the extent of the visual streak horizontally across the retina. This variability in the ganglion cell density within the visual streak has been evident between nocturnal and diurnal species, including ungulates and primates (Hughes, 1977; Lima,

Silveira, and Perry, 1996). These studies revealed that the visual streak of nocturnal animals, if they possess a visual streak, generally contains lower cell densities than that of diurnal animals

(Hughes, 1977; Lima et al., 1996).

Within the realm of aquatic animals, Oliver et al. (2000) revealed that green sea turtles, inhabiting areas with clear water and a high level of illumination, have strong visual streaks (e.g. visual streaks with the highest number of retinal ganglion cells and longest horizontal extent). In contrast, the loggerhead and leatherback sea turtles, both inhabiting areas with highly varied illumination, possess weaker visual streaks with lower cell densities and shorter horizontal extents.

Habitat

Referring to habitat, the visual streak generally extends the farthest horizontally in species living in open habitats with an unobstructed view of the horizon (Collin and Pettigrew,

1988b; Oliver et al., 2000). For instance, Collin and Pettigrew (1988b) investigated several teleost species inhabiting coral reefs and found that species with completely unhindered views of the horizon, such as the clown triggerfish and blue tuskfish, possess horizontal visual streaks extending across most of their retinal meridians. Whereas the Australian frogfish, inhabiting a more “closed environment” (e.g. with an obstructed view of the visual horizon) possess a weaker visual streak extending a much shorter distance across the retinal meridian (Collin and Pettigrew,

1988a).

9 Behavior

The third factor which may relate to an animal’s visual streak concerns how an animal behaviorally utilizes its visual system. Collin and Pettigrew’s (1988b) study of coral reef fishes revealed that the differences in location of peak cell density within the visual streak and actual location of the visual streak may vary due to the way these fish species utilize them. More specifically, the topography of the visual streak may vary in association with the behavioral needs of the teleost (e.g. the visual streak may be more important for predatory behavior and/or for predator surveillance).

For example, the blue tuskfish and painted flutemouth both possess a horizontal visual streak along their retinal meridians (Collin and Pettigrew, 1988b). However, the visual streak of these fish species differs in regards to whether a temporal area of increased ganglion cell density is separate from (blue tuskfish) or extends into (painted flutemouth) the visual streak. The variations in retinal topography possessed by these fish species may indicate whether their visual streak is useful primarily for predator surveillance or for predatory behavior (Collin and

Pettigrew, 1988b). The blue tuskfish forages by searching through and moving coral debris on the substrate in search of food. The authors believe that the visual streak of this fish species may be useful for predator detection, as possessing a visual streak congruent with the environmental horizon may help it to watch for predators while it forages (Collin and Pettigrew, 1988b). The temporal area of increased cell density in the blue tuskfish appears better suited to foraging for prey (Collin and Pettigrew, 1988b). The temporal area subtends the visual region directly in front of the blue tuskfish, and should increase its resolving power in the area where the fish would be searching for invertebrates within the substrate (Collin and Pettigrew, 1988b).

The painted flutemouth, however, shadows other fish species to approach its prey (swift- moving fishes) by surprise (Collin and Pettigrew, 1988b). Possession of a visual streak correlated with the environmental horizon may be more useful to detect and approach unsuspecting prey,

10 rather than guard for predators in this fish species. The authors also state that this fish’s retinal topography (e.g. having the temporal retinal specialization “extend” into a visual streak across the retinal meridian) is similar to the retinal topography of other species where vision is more valuable for prey detection than predator surveillance (Collin and Pettigrew, 1988b).

The study conducted by Oliver et al. (2000) also supports the idea that differences in feeding behaviors may be correlated with visual streak length. Of the three species examined in the study, green turtles possessed the strongest and longest visual streak, likely due to their well- illuminated habitat containing an unobstructed view of the visual horizon (Oliver et al., 2000).

Both the loggerhead and leatherback turtles possess weaker visual streaks than the green turtle, with the leatherback turtle possessing the weakest visual streak of the three (Oliver et al., 2000).

Although the visual streak of the loggerhead turtle was wider than that of the green turtle, its visual streak was less horizontally extensive and contained a lower ganglion cell density (Oliver et al., 2000). Loggerhead turtles forage for prey such as snails, sea anemones, and crustaceans contained within and around the mats of sea grasses or algae this turtle hides amongst (Oliver et al., 2000). The sea grasses and algae mats loggerhead turtles feed within would hinder much of their vision, with the exception of objects located directly in front of them. Therefore, it is likely they would not need to possess a visual streak extending across their entire retinal meridian to provide them with increased sampling of visual targets in their lateral/peripheral visual fields

(Oliver et al., 2000).

Of all three turtle species in the study, the leatherback turtle possessed the weakest (least elongated across the retinal meridian and containing the lowest density of retinal cells) visual streak (Oliver et al., 2000). The leatherback turtle was also the only species in the study to possess an area centralis (e.g. a small area of increased retinal cell density) separate from their visual streak (Oliver et al., 2000). Leatherback turtles feed primarily on jellyfish and other jellylike prey they capture via diving. The weakness of the leatherback turtles’ visual streak

11 indicates that this type of visual adaptation may not be as beneficial to detect prey as their well- developed area centralis (Oliver et al., 2000). The lack of a lengthy and strong visual streak in the leatherback turtle could also be a result of their predilection for capturing prey in open water where the environmental horizon may be vague or even absent (Oliver et al., 2000).

Other examples regarding how the location of the visual streak may reveal its importance in a species daily survival, and how the visual streak may relate to an animal’s behavior (e.g. foraging or predator surveillance behaviors) are the striped panchax (Apolcheilus lineatus) and

Graham’s Hechtling (Epiplatys grahami), two freshwater fish species. Both the striped panchax and Graham’s Hechtling possess two “band-shaped” thickenings extending across their retinal meridian. One band traverses the retinal meridian, and the other “band” extends across the retina, just ventral to the retinal meridian (Collin and Pettigrew, 1988b). These band-shaped thickenings are the equivalent of visual streaks (Collin and Pettigrew, 1988b). These fish species feed on insects and other small prey items living upon or just below the water surface where higher resolution power in the upper visual fields would aid these fish species in locating prey

(Collin and Pettigrew, 1988b). Therefore, the authors believe that the ventral thickening is likely useful for detection of prey located just above the fish and the thickening of the central retinal meridian is congruent with the lateral visual field of these fishes and useful for predator detection

(Collin and Pettigrew, 1988b).

Organization of retinal ganglion cells in sharks

The visual surroundings of many shark species relate well to the terrain theory, as their environments are composed of a two-dimensional setting containing a sand-water boundary or a boundary containing a “horizontal gradation of light within the water column in the clear waters of the open ocean” (Bozzano and Collin, 2000). Therefore, most shark species, especially those living in relatively well-illuminated benthic, pelagic, or mesopelagic habitats, could possess a

12 visual streak. This appears to be true (see Table 1) as shark species investigated thus far (e.g., the tiger, epaluette, black-mouthed, velvet-belly, lemon, small-spotted dogfish, and California horn shark) all possess a visual streak (Bozzano and Collin, 2000; Hueter, 1989; Hueter, 1991).

Shark species that have been investigated also show species-specific variations in the width, length, and location of their visual streaks, similar to that found in teleosts and terrestrial vertebrates (Bozzano and Collin, 2000; Hueter, 1991; Peterson and Rowe, 1980). These variations in visual streak organization appear to be associated with environmental factors as well as predatory or surveillance behavior, and not with phylogenetic relationships between shark species.

Illumination

In contrast to sea turtle hatchlings, there appears to be no apparent difference regarding overall ganglion cell density within the visual streak between deep-sea (low illumination) and shallow-water (higher illumination) shark species (Bozzano and Collin, 2000). However, there does appear to be a difference in the percentage of giant ganglion cells contained within the retina of shallow water versus deep-sea sharks (Bozzano and Collin, 2000). Giant ganglion cells are characterized by a larger soma (e.g. two-to-three times the soma size of other ganglion cells) and are thought to possess larger receptive fields than the normal ganglion cells (Bozzano and Collin,

2000). Shallow water species, (e.g. the tiger and epaulette sharks), regardless of whether they are benthic or pelagic, have the lowest percentage of giant ganglion cells, whereas deep-sea species

(e.g. the blackmouth dogfish and velvet belly sharks) have the highest percentage (Bozzano and

Collin, 2000). This difference in the overall percentage of giant ganglion cells may actually result from differences in the distinctiveness of the visual horizon between deep-sea and shallow water species (Bozzano and Collin, 2000).

13 Habitat

Bozzano and Collin (2000) suggest that more pelagic than benthic species should possess a broad visual streak. This may be generally true, as the small-spotted dogfish and Bigelow’s ray are benthic and possess a narrow visual streak, whereas the black-mouth dogfish shark is benthopelagic and possesses a broader visual streak (Bozzano and Collin, 2000). However, the lemon shark, a shallow-water benthic species with a broad visual streak, may be an exception.

The visual streak of the lemon shark forms a fairly wide horizontal band of increased cell density running along the retinal meridian (Hueter, 1989). This form of visual streak may aid the lemon shark in both prey and predator detection. Both diurnal and nocturnal in activity, the lemon shark pursues swift-moving crustaceans and fish (Hueter, 1991). It hunts via patrolling over a sandy substrate or sea grass flats sweeping its body from side-to-side (Hueter, 1991; Oliver et al., 2000). Possession of a broad and lengthy visual streak may allow this shark to detect movement of visual objects both directly in front of it and within the lateral periphery of this visual field (Hueter, 1991). Therefore, the visual streak of the lemon shark may play a role in detection of prey. Although hunters themselves, lemon sharks are also occasionally predated upon by larger sharks. The presence of a sizeable visual stimulus has been found to elicit a

“rapid withdrawal response” in lemon sharks (Hueter, 1991). Therefore, it is also likely that the visual streak of the lemon shark may be useful for detection of predators as well.

Behavior

Whether a visual streak is broad or narrow may also be associated with whether detection of predators or prey is visually important to a shark species. The location of the visual streak either across the retinal meridian, or just dorsal or ventral to the retinal meridian, may vary in relation to the different types of predatory behaviors employed by shark species.

14 All shark species examined possessed rather centrally located visual streaks with the exception of the tiger shark (Bozzano and Collin, 2000). The visual streak of this shark is located ventrally to the retinal meridian (Bozzano and Collin, 2000), and would subtend vision in the upper visual field. The tiger shark is a large predator generally found in shallow water, near the surface (Bozzano and Collin, 2000). This shark likely has the most varied diet of all shark species, as it feeds on bony fish, sea turtles, sea snakes, mollusks, and mammals, among other items (Compagno, 1984). As this shark generally attacks via a bump-and-bite or ambush-style predatory technique (Compagno, 1984), a visual streak allowing it to swim unnoticed underneath potential prey, such as one subtending the upper portion of the shark’s visual field, would be advantageous.

In regards to predatory behavior and increases in ganglion cell density, an increased area of retinal ganglion cells was found in the center of the epaulette shark’s visual streak (Bozzano and Collin, 2000). This benthic shark inhabits relatively shallow waters and preys upon benthic invertebrates such as crustaceans and mollusks (Compango, 1984). An increase in the resolving power within the central area of their frontal visual field may aid them in locating their prey

(Bozzano and Collin, 2000).

The benthopelagic blackmouth dogfish shark possesses two areas of increased cell density within the nasal and temporal areae of their visual streak (Bozzano and Collin, 2000).

This shark consumes swift-moving prey such as bony fishes and cephalopods via sweeping its head and body from side-to-side (Bozzano and Collin, 2000). The two areas of increased cell density should increase the sampling of visual targets within the shark’s frontal and caudal visual fields and are likely useful for detection of prey (Bozzano and Collin, 2000). This sharks’ retinal topography is also congruent with the idea proposed by Collin and Pettigrew (1988b) that species possessing temporal areas of increased cell density which extend into a visual streak are more likely to utilize these areas for prey detection.

15 Possible retinal ganglion cell topography of the bonnethead shark

The bonnethead shark is likely to possess a visual streak extending across the entire length of their retinal meridian, congruent with the visual horizon, due to its potentially well-illuminated shallow-water habitat which should possess a rather distinct visual horizon.

Similar to the lemon shark, the visual streak of the bonnethead shark may be useful for both detection of predators and prey. However, the primary use of this shark’s visual streak should not be prey detection, as it would appear from the location of this shark’s eyes on its broad, elongated head, that this shark may lack a frontal visual field. This potential blind spot would negate the bonnethead shark’s ability to visually locate prey directly in front of it, though it may possess the ability to locate prey within its lateral visual fields. The sweeping side-to-side motion of this shark’s head while it patrols for prey may also aid it to detect prey within its peripheral visual fields.

Even though pelagic sharks may, in general, possess wider visual streaks than benthic species, the bonnethead shark, like the lemon shark, may be an exception. Due to the location of this shark’s eyes within the extreme edges of its broad, shovel-shaped head, the bonnethead shark may also possess an elongated lateral visual field. This potential increase in the lateral visual field should allow the bonnethead shark to develop a broad horizontal visual streak extending across their entire retinal meridian. This type of visual streak should aid the bonnethead shark in avoiding predation by larger shark species, as it would provide the bonnethead shark with full visual access to the areas along its sides. Prey species of this shark are primarily crustaceans with the ability to change direction rapidly, thus possessing a broad visual streak may also aid them in locating potential prey within their lateral visual fields.

16 Chapter Two

Methods

Eight retinas, from eight individual sharks, were utilized for this study. The sharks were obtained with help from Mote Marine Laboratory in Sarasota, Florida. Each shark was caught within the Tampa Bay region (Charlotte Harbor) using gill nets. Measurements of length and weight were taken before the sharks were placed into a cooler containing ice. A preservative (4% paraformaldehyde solution in 0.5 M phosphate buffer, 1.00 cc per eye) was injected intraocularly to prevent disturbing the retina within ten minutes of the sharks expiring to preserve the eyes.

The eyes were then removed and immersed in a 4% paraformaldehyde solution in 0.1 M PB solution, Ph 7.4; Huxlin and Goodchild, 1997) in small containers and placed into a cooler for transport back to the lab. Excess tissue (e.g. connective tissue) was removed from the eye before the eye was placed into the preservative solution. The retinas were removed from the eyes and wholemounted within 24 hours of collection. Eight of the retinas were used for Nissl staining and two eyes (without the retina removed for the procedure) were used for the retrograde tracing with

DiI. Though eyes were to be counterbalanced between left and right for this project, seven were right eyes and one was from the left eye. This discrepancy was due to selection of the best wholemounts from the retinas available for topographic analysis of ganglion cells.

Nissl Staining

Before removal of the retina from the eyecup, each eye was marked with a small indentation, using a # 11 scalpel (Hueter, 1988), to maintain the dorsal/ventral and anterior/posterior orientations of the eye. The eyes were then removed and placed in a deep petri dish containing the preservative solution (4% paraformaldehyde solution in 0.1 M phosphate buffer). The preservative solution covered each eye to prevent them from becoming dehydrated.

After fixation, an adaptation of Hueter’s (1988) retinal wholemount technique was utilized. To

17 remove the retina, the eye was placed in a petri dish filled with the 4 % paraformaldehyde solution. Using a scalpel, the eyes were cut open at the choroidal-scleral boundary to gain access to the retina. Cutting ceased when there was a small amount of tissue, forming a ‘lid’, left on the dorsal portion of the eyecup. The lens and vitreous humor were then lifted from the eyecup and removed. Eye orientation (dorsal vs. ventral) was maintained while the retina was removed from the eyecup using small indentations made with a scalpel blade on the dorsal and ventral margins of the retina. Using a soft brush (camel hair, #0), the retina was then transferred unto a glass slide.

The retina was then flattened against the glass slide. If the retina did not lie flat against the glass slide, small incisions were made around the retina’s circumference to help it to lie flat.

Any additional preservative solution and vitreous material was then carefully removed by touching filter paper to the coverslide to absorb them. Each retina was then placed into a dust- free container for at least 12 hours to fully adhere to the slide before being taken out and gently washed with DH2O to prepare it for Nissl staining.

Each retina was stained for 10-15 minutes in 0.05% Cresyl Violet. Each slide was then dehydrated, cleared, and coverslipped. Once the retinas were stained, the retinal ganglion cell layer was examined microscopically.

Topographic Mapping

Before taking pictures, each retina was traced into the Canvas7 computer software program using a computerized Wacom drawing tablet, then divided into 1mm sections. A starting point for pictures was pinpointed, then the coordinates for each 1mm section were labeled on the retinal drawings from readings taken from a Nikon microphot-FXA microscope using an

X,Y grid system on its stage micrometer. A micro-photograph was taken of the lower right corner intersection between each 1 mm square with a Lucida camera attached to the Nikon 18 microscope. Retinal ganglion cell counts were taken from a 200 micron square area in the center of each microphotograph. Cell counts were not taken from the very edges of the retina, as retinal shrinkage (usually between 2 and 20%; Collin, 1988; Oliver et al., 2000) can occur in these areas

(Bozzano and Collin, 2000). Retinal ganglion cell counts were noted for each 1mm section of each retina. These raw counts were then converted to the number of cells per 1mm² and topographic maps were composed to represent the visual topography.

In the ganglion cell layer, ganglion cells were differentiated using morphological standards of Collin (1988). The author found retinal ganglion cells to be large, irregularly shaped with darkly stained somas (Collin, 1988). Ganglion cell morphology was to be assured by labeling ganglion cells within the retina using DiI for retrograde tracing.

Retrograde ganglion cell labeling

In order to assure correct identification of ganglion cell morphology, four retinas were to be examined using retrograde labeling of ganglion cells. Two eyes from one bonnethead shark were used to test whether the crystal form of 1,1’ –dioctadecyl-3,3,3,’,3’ – tetramethylindocarbonocyanine perchlorate (DiI) would be suitable for retrograde labeling of ganglion cells. DiI is a lipophilic carbanocyannine which attaches to and stains the plasma membrane. It is then is diffused laterally through the cell, eventually staining the entire cell.

The optic nerve ending of the eyes utilized for retrograde labeling were cut to make the ends level and a small incision was made at into the end of the optic nerve. A small crystal of DiI was then placed into the incision. As DiI is sensitive to light, this procedure was performed under minimal light conditions in the lab. Once the DiI crystal was securely placed into the incision at the end of the optic nerve, the eye was returned to the container of preservative solution, with the optic nerve tip containing the DiI crystal supported above the preservative to keep it dry. In both cases, the eyes were placed into small containers filled with the 4% paraforamldehyde in 0.1M

19 PB buffer solution, Ph 7.4. The containers were then wrapped in aluminum foil to protect them from light contamination and placed in a dry, dark area. DiI was allowed to absorb into the cells of the eyes in one case for two weeks and for four weeks for the second case. In both cases, after the allotted time for DiI absorption, the retinas were removed and wholemounted using the aforementioned procedure with the exception of the retinas being removed under low light conditions. Once the retinas were wholemounted, several drops of glycerin were placed on the retinas and a coverslip applied. The retinas were then examined under a Nikon microscope under fluorescent lighting for evidence of DiI staining. Unfortunately, the DiI did not completely stain the entire ganglion cell bodies, making it impossible to use DiI to confirm ganglion cell morphology. In lieu of the DiI retrograde labeling, morphology of ganglion cells was confirmed using the descriptions from Collin (1988) and Hueter (1991).

Data Analysis

Ganglion cells

Wholemounted retinas were used to identify ganglion cells. The morphological criteria for identification of ganglion cells by Collin (1988) and Hueter (1991) was used. An average of

333 regions were sampled from each retina (see Table 1 in the results section for individual sampling data). A visual streak was to be defined in this study as any area of the retina where a significant increase in ganglion cell density was found.

Nissl stains

Eight retinas, from both male and female bonnethead sharks were whole mounted, stained with Nissl substance, and photographed. After completion of retinal counts, each retina was mapped. The resulting maps were divided into four quadrants (dorso-nasal and temporal and ventro-nasal and temporal) to aid the descriptive process. Retinal counts were divided into four

20 categories, low (under 25%), medium (26-50%), high (51-75%), and highest (over75%) number of retinal ganglion cells. These categories were based on dividing the maximum and minimum counts averaged across the eight retinas into four equal quartiles and used to measure differences in retinal ganglion cell numbers across each retina.

Expected results of these experiments

The bonnethead shark was expected to possess a wide horiztonal visual streak (covering at least one-third of the longitudinal retinal diameter) across the entire length of the retinal meridian.

This streak should mediate vision in their panoramic lateral visual field. If this visual streak was not found, then vision may not be important to the daily survival of this shark species.

If retinal specializations were found (increased areas of peak retinal ganglion cell density) in the nasal region of the retina, then the bonnethead shark may use this area of increased resolution to detect predators coming from behind, and therefore, predator detection would likely be the primary function of their visual system, and its subsequent organization. If a retinal specialization was found within the temporal region of the bonnethead shark retina, then this shark species may utilize their visual sense to detect prey in front of or to the sides of the shark’s head. A visual streak in this retinal area would likely be aided by the head movements of the bonnethead shark, as they sweep their heads from side to side while patrolling for prey.

21 Chapter Three

Results

For the eight retinas utilized in this study pictures (0.5mm by 0.5mm² each) were taken at

1mm intervals across the entire area of each retina. This cumulated in a total of 2,660 photographs. The number of pictures for each retina varied between 266 (case # IMF2) and 413

(case # 2625-2). The average number of photographs taken per retina was 332.5. For each picture, retinal ganglion cells were identified and counted and this information was used to create topographic maps of ganglion-cell density across each individual retina (Figures 2 through 9).

See Table 2 for individual data regarding the number of counted areas for each retina. A typical photographic image utilized for analysis is shown in Figure 1.

Figure 1. Typical photographic image showing retinal ganglion cells Darkly stained irregularly shaped cells (solid arrows) were counted as ganglion cells. The smaller circular cells (open arrow) were considered possible amacrine cells and not counted.

In general, topographical maps revealed heterogeneous characteristics of retinal ganglion cell distribution. Therefore, some areas within each retina showed higher ganglion cell density than the rest of the retina. In most cases, these areas of higher density formed a band-like shape

22 across the retinal meridian. The term “band” was used to describe them instead of “visual streak” for the results section. For the reasoning behind this terminology, see the discussion section.

Of the eight retinas used for this study, six were from adult animals and two were from immature individuals that are referred to as IMF1 and IMF2. However, since no apparent differences regarding the cell distribution pattern were revealed between the adult and juvenile sharks, all results, including those from both immature and adult specimens, were combined and analyzed together.

DiI crystals were placed into the optic nerve to stain and identify ganglion cells.

However, the DiI substance did not fully stain the ganglion cells, making it impossible to discern ganglion cell morphology. Descriptions of ganglion cell morphology from Hueter (1991) and and

Collin (1988) were used to identify and count ganglion cells in lieu of DiI staining.

General Results

Cell Numbers/Density

In all cases combined (from the 2,660 individual photographs), a total of 402,372 cells were identified as retinal ganglion cells. The minimum number of ganglion cells found was 127 cells per mm² (case # 2625-2), whereas the maximum number equaled 1571 cells per mm² (case #

2626-8). The mean ganglion cell density was 693 cells per mm² with a standard deviation of 43.

Figures 2 through 9 show the frequency distribution for each individual case.

For the results section, cell numbers were categorized into quartiles in order to demarcate the general trend of retinal ganglion cell distribution as well as utilize all data collected from the eight cases. The quartiles were obtained by averaging the minimum and maximum cell counts from each individual retina. These averaged minimum and maximum counts were then used to establish the quartiles utilized for this study. These quartiles are defined as follows: Low (0-480 cells per mm²), Medium (481-747 cells per mm²), High (748-1010 cells per mm²) and Highest

23 (1011+ cells per mm²). Information for each individual case is found in Table 2 (counts are in cells per mm²).

High Density Band

Out of the eight cases, two retinas revealed no clear heterogeneous pattern of ganglion cell distribution (cases 2625-2 and 2626-6). However, the remaining six cases (2627-1, 2626-4,

2626-5, 2626-8, IMF1 and IMF2) possessed the area of increased ganglion cell density categorized as “higher density” along the retinal meridian (see Figures 2, 3, 4, 5, 6, and 7).

These bands, though, varied in length between the individual cases. Four of the retinas (cases

2626-5, 2626-8, IMF1, and IMF2) contained areas of “higher” cell density running at least two- thirds of the length of the entire retinal meridian. In two cases (2626-4 and IMF2) the band extended across the entire length of the retinal meridian, and in one case (2627-1) the area of

“higher” cell density covered at least half the length of the retinal meridian.

Variations in the width of this “higher” density band were also observed. In four cases

(2626-4, 2626-5, IMF2 and 2626-8; see Figures 3, 4, 5, and 8), the “higher” density band covered approximately one-third of the dorsal-to-ventral expanse of the retina. The width of the “higher” density band was widest in one case, 2627-1(see Figure 2), where it covered at least half of the dorsal-to-ventral expanse of the retina.

High density dorso-temporal area

In addition to the band of “higher” density, there was a distinct area of increased ganglion cell density in the dorso-temporal retina in six cases (2627-1, 2626-4, 2626-5, 2626-8, IMF1 and

IMF2). Shape and density of this dorso-temporal increase varied between all six cases (see

Figures 2, 3, 4, 5, 6, and 7). Three cases, (2627-1, 2626-5, and 2626-8) had fairly small dorso- temporal areas compared to the rest of the cases. In all five cases, density within the dorso-

24 temporal area was categorized as “High”. In three cases (2627-1, IMF1 and IMF2), this dorso- temporal higher-density area appeared to be an extension of the “higher” density band, whereas in the other four cases (2627-1, 2626-4, 2626-5, and 2626-8) this area was not connected to the

“higher” density band.

Table 2 Retinal ganglion cell counts and statistics (all cell counts in cells/mm²)

Case 2627-1 2626-4 2626-5 2626-8 IMF1 IMF2 2625-2 2626-6 Overall Number Mean Mean 702 742 634 751 720 801 575 620 693

Median 688 742 634 720 706 797 566 620 684

Minimum 136 285 127 353 208 371 127 140 218

Maximum 1317 1322 1036 1571 1249 1457 1118 1086 1270

Counted 285 328 317 339 368 266 413 344 322.5 Areas

25 Individual Results

Dorsal

Cells/mm² Low Medium High Highest Nasal Temporal OD

Ventral

Figure 2. Topographic map of case 2627-1

Retina 2627-1

This retina came from the right eye of an adult female shark. The band of “higher” density (outlined in figure), starting in the mid-nasal portion of the retina, was found running along most of the retinal meridian in this retina (Table 2, Figure 2).

The dorso-nasal portion of the retina contained both low and medium cell counts. The dorso-temporal portion of the retina also contained mostly medium cell counts, with the exception of an area of high counts located just above the retinal meridian in the extreme temporal edge of the retina. The band of “higher” cell density started just nasally of the optic disc (along the equator of the retina) and ran from this area across to the temporal portion of the retinal meridian

26 as well as down toward the ventro-central portion of the retina. The ventro-temporal portion of the retina contained mostly low and medium cell counts.

Dorsal

143 194 138 171 162 142 e 131 135 e 182144 107 e e e nc 149132 178 156124 162 118 nc 165 167138 150 170 c 136101 172106 151168123187 12916999 164176 166 e e 183 nc e 175 113139 141173 157158145178 170 182 118 171195 159 143157 143128 139 79 143161131125140162 209182175 179 205190213 178150168 125181197 e/c 222 161 136106 14685 128151116 157155157 141165 171 154 218 229 142123146 152174 202 149 198 210 172153 83 117 10015516311581510148 139 156146 176161 223203 169 166115126 130155116 e e176186 175 131182175181141188164 180155 180167200222 225 142

152 161192 218220201218211 166155 140 198 Nasal 159 92 193171169186258174 245 203216 222 Temporal 164 171 e 63c 161152 176168164 169175 192 198209 125 156 206 198 241 161 127198on 217 223195 173 145 136220 205 137 218 172 139173 150 174 OD 195198 244170177 e 176171172 157179196 123175173191 270198180165 Cells/mm² 202 195 174162 229 72 157161 163159 161 e 238e/c e 203174172 Low e 182 150161 140 109 149178 191 187162129 164183 161 121 Medium 201152 81 184 134 e 139 77 146 133 145 190154190 189 176 High 183255 142131 85 84 91 143175 159179174166 190 158 Highest e 164 156 195 180115 157 160 166172 153 183292 198199180 148 163162 e 161187 nc 169185 106139 152138141 140204 133

Figure 3. Topographic map of case 2626-4 (areas marked with e= edge, nc= not countable)

Retina 2626-4

This retina came from the right eye of an adult female shark. A band of “higher” cell density was observed along part of the retinal meridian, starting at the nasal edge of the retina and running across the entire retinal meridian (Table 2, Figure 3).

The dorso-nasal portion of the retina contained mostly medium cell counts with a few randomly interspersed high counts. The dorso-temporal portion of the retina also contained some medium and mostly high cell counts, as well as an area of High cell counts located just above the retinal meridian running from the mid-retina to the extreme temporal edge of the retina. The ventro-nasal portion of the retina contained mostly medium cell counts with a few low and high

27 counts at the extreme ventro-nasal edge. The ventro-temporal portion of the retina contained

mostly high cell counts.

Dorsal 121 89 130 e 120 135 nc 113 118 132133 62 86 126144 97 80 80 64 111114 124 109134 1mm 95 147 119143168 142 150 93 198133 108 87 118 206 127181 150155 91 187 196173129 135 85 101 74108 99 108 142 125 209159 175182 196173134 147 187 nc 147 92 139 147 238197 217 116107195 125 114 e 193 127137 190 95 162 71 e 55 93 77 185172 139153154 123 107131159129 139142 119 114117112 93 154 180 228 157 180 123176 155158118143 131110139 112107 118 127 94 12091 100 e 160178 162154 171 80 40 59 136 141152 148 153132175131 156 147 86 115e151147 191 250 192 142 153 177225 95 63 180236 242 244229 246 239256245203 243 194 207189155 116193 165 202 164 146 124 155157 132104 63 229133 152 146162 173 175156 165 175185186 169 199187 225194 167 153161 146 133 142 130156179 119110 98 160 158 175168 176 210209170 181 165169 217177 165 206177

nc 83 104 125174 174 183207 198 207166158 181 176165 194 Nasal nc 90 102 111138 139 147177 148 209208 58 51 189185 nc 109 118163 182 170 171217 nc 168 193 165181 196 Temporal 146 172 142171181156 171 171170 140162 164 140 164154142131 163 208178 166172 160 137 171168159127126 169149 167136 89 164 169 147171161142 107 133144 163144137 Cells/mm² 151 98 157 159 168186153 80 133 135144 168124148 0-480 Low 170147182 102 481-747 Medium 124 147 nc 748-1010 High 1011+ Highest

Ventral Figure 4. Topographic map of case 2626-5 (areas marked with e= edge, nc= not countable)

Retina 2626-5

This retina came from the left eye of an adult female shark. A band of “higher” ganglion cell density (outlined in figure) began in the mid-nasal portion of the retina and ran across the rest of the retinal meridian (Table 2, Figure 4).

The dorso-nasal portion of the retina contained mostly medium and low cell counts. The dorso-temporal portion of the retina contained mostly medium cell counts. The ventro-nasal portion of the retina contained mostly medium cell counts whereas the ventro-temporal portion of the retina contained mostly medium and some high cell counts in the most dorsal part of this area.

28

nc Dorsal 122 140152 ne 136135 166 e 78 154 120 nc 134 159 200170 79107 142 184109 135 71 nc nc nc 97101 133 303 248267 202183 171 180 168149 178 nc 99 138 140144 127174211 186 1mm 191 181165 180 190 183 nc 112 nc 139 13695 119139144116 117143 150 203 146128170 146 nc nc 139 62 159 125132 132138 134 136124170 170153146120 134106 9694 126 133 nc 175177 180 161 171 181 188 148163 187159163 163 e 187170 140 128133 97 108 153 154 123 158 141 168203235 236 187 192197 228 160 164 130130101 118132 138 170 nc ck 195205 185 170 ck 114 175 172140 177179 171 192166 174242255257 17022021120616121911213985 146129131 137120106133 135149124 123 264176 150193 211200199 224 nc 281200193 208 124135138 228 238264 248 Nasal 175 193 253 248281229 226219 176198224 202181 174144145 167 146184 Temporal 141 175176 226 nc213 303 222149 243274 266 209210170214 205 186176198 203204

142181 163 186232 251 218 227201 nc 152 248 234 251 286 219213193136 177 181 OD 178189 171 173 225249 216 262 203196244 240250 263233222287 218215 173164 165 184 172158152 145 nc 166180 158176222 233 292264347 264230180 184 159173 nc 179154 142 199169163 168186 191 181 223248 265217 241191 185134 Cells/mm² 175 183 211 206 211 121 Low 163 155167 160181 135 155 216 212 137140 0-480 170 135156 146142161 148176171 186 nc 129153 481-747 Medium 140 168 125 139136 144 nc 162 nc e 139 748-1010 High 109 128 126144 131 117 118128 1011+ Highest e 79 e 140 111128 nc nc 140 121 122 nc 98 Ventral

Figure 5. Topographic map of case 2626-8 (areas marked with e= edge, nc= not countable)

Retina 2626-8

This retina came from the right eye of an adult female shark. A band of “higher” cell density (outlined) was located along the retinal meridian. This band started in the nasal portion of the retina and ran across the majority of the retinal meridian as well slightly into the ventro- temporal portion of the retina (Table 2, Figure 5).

The dorso-nasal portion of the retina contained mostly medium cell counts. The dorso- temporal portion of the retina contained mostly medium cell counts with a small area of high counts within the dorsal portion of this area. The ventro-nasal portion of the retina contained mostly medium cell counts interspersed with a few high counts. The ventro-temporal portion of

29 the retina contained mostly high cell counts with the very extreme ventral edges containing medium cell counts.

Dorsal

152 131 101 133 162 18

126 192 192 152 150 156 92 180 146 173 158 271 162 178 217 179 126 109 184 145 203 157 145 153 131 158 123 127 186 179 248 182 111 93 143 146 169 119 181 106 91 140 174 132 135 118 216 164 191 143 196 206 204 145 134 100 108 116 140 166 143 86 89 137 156 135 176 121 231 226 149 206 142 135 139 230 133 78 86 122 92 159 111 113 104 153 153 153 159 146 169 190 235 196 116 209 217 149 156 122 133 131 106 200 175 157 155 e 142 169 133 272 159 202 217 224 276 129 147 124 109 59 183 e 182 224 230 184 169 240 232 140 e 119 56 1mm 167 200 46 e 104 178 240 210 138 121 133

176 190 53 e 210 96 145 197 172 183 241 196 228 Nasal 191 232 149 204 221 257 183 245 226 202 Temporal

238 230 173 e 220 210 178 219 157 170 175 142 228 247 159 150 206 134 145 152 108 141 206 204 182 175 139 133 151 120 128 Cells/mm² 222 163 157 110 120 130 152 158 202 148 120 0-480 Low 155 170 156 107 151 127 107 185 143 85 481-747 Medium 125 173 166 197 155 146 140 140 117 120 184 128 117 748-1010 High 139 126 173 158 185 227 157 163 129 143 168 188 195 1011+ Highest 130 176 173

Ventral

Figure 6. Topographic map of case IMF1

Retina IMF1

This retina came from the right eye of an immature female. A band of “higher” cell density (outlined) was located along the retinal meridian and started at the nasal edge of the retina and ran into both the dorso-temporal and ventro-nasal portions of the retina (Table 2, Figure 6).

The dorso-nasal portion of this retina contained mostly medium cell counts with some a few high cell counts located at within the central portion of this area. The dorso-temporal portion of the retina contained high density cell counts within its mid-to-dorsal portion and medium and low counts at the extreme temporal edge of the retina. The ventro-nasal portion of the retina

30 contained mostly high cell counts, whereas the ventro-temporal portion contained mostly medium cell counts with a small area of high counts in the most ventro-temporal portion of this area.

e Dorsal 200 161141nc 167214 nc nc123172111183248 202201196199192 137147132168170130140 e 206179194 218198151169227 215164160135149169159 nc 215117176224172163160241 nc 201209183218174157148112169175130 188 263170195186203146 nc 190255192200173156168150171178166168nc137 224175 nc 144 198267287270154171166165177175173 nc 175154 e e 226 e 180 291 228197188167182186211197167144256196nc e 203198207 295217 e 243221228 189216 224231234197 193238219223264297e 200 e 260273278240236221203211190182199210200217207 187 120 218183215190170195217164168200127193169203163174 196 206221253296 82177169200199123218116e 195163204 225220322 e 12799 9613911313816095 108201103186106 Nasal 154127nc 92 164136171145128192164195185143178 Temporal 189 e 92171207127128205153126175139158176183172189 185e 167 237152164202113175124203135 157182160137180180158 e 182154 150151201175154 150178197182119187138178 Cells/mm² 167 157130217 107138 158226251147201126 117 125128172 1mm 0-480 Low 175nc 168 231222 nc 150130 114147183179 194 481-747 Medium e e 4513944nc nc nc 119 196 e 187 151190 748-1010 High 133128145 nc134 e 205 187 1011+ Highest 189232 199 Ventral

Figure 7. Topographic map of case IMF2 (areas marked with e= edge, nc= not countable)

Retina IMF2

This retina came from the right eye of an immature female. A narrow band of “higher” cell density, starting at the nasal edge of the retina and running across the entire retina, was located along the retinal meridian, just above the optic disc (Table 2, Figure 7).

The dorso-nasal portion of the retina some medium cell counts with an area of high cell counts. The dorso-temporal portion of the retina contained mostly high cell counts. The ventro- nasal portion of the retina contained mostly medium and a scattering of high cell counts whereas

31 the ventro-temporal portion of the retina contained a mixture of high and medium cell counts with no distinguishing pattern.

e 127 172 e 116 Dorsal 149135146 e nc e e 119121 142122111130 171 135 e 102 143125170 82 105 99 107109135 146 154 108 e 134 104132 108 98 120 92 119 137 109 133 113 e 99 116106 89 108 102112119 108 123 144 11291 e 86 101 93 e 97 108102 134 134 124 122 116 102106 108 78 112120 97 108104109 119 147 121 139140 121 94 e 102 126102113 78 121 108 94 122 92 110 120 Cells/mm² e 102128 74 126105 110102105 79 62 123 109125 86 110 144 143 130150145 Low e nc nc 74 156 82 87 114124113 125127 51 92 110163 202 222 190173193 214 136e Medium 119 e 156 127122 116 155 68133 180 83 106 90 65 51 62 82 155 nc 154 156 132 High 12 Nasal 131151184177 72 85 138 125128 107 94 139 159193138 117 Temporal Highest 123 e 135137122148 127175 184 121 173156 141 99 158 198247207 136 127 184169 155 e 110155159138145 127173 28 55 191137 171 164153169 173 134199 164 164 86 187 117 ON87223 107 192159172 139 137 142152 105152 148201 179152153 185132 113 134 168132152 144 168 115 153134 132 139160 113149 101 e 100100 143 141129142 138 137 111129

113165 148 156116 122 72 99 103126130 136 139 108116 127148 102114

85 129136 145 116111 109130105 126 e/c 98157 146 121 123155 125 111140 96 e 135139 122101 127 58138 122 130 139114 146 172 188157 162 148151 1mm 167143 Ventral

Figure 8. Topographic map of case 2625-2 (areas marked with e= edge, nc= not countable)

Retina 2625-2

This retina came from the right eye of an adult male. This retina did not contain any distinct pattern of cell distribution. However, there were scattered small areas containing high cell counts located in areas along the retinal meridian (Table 2, Figure 8).

Both the dorso-nasal and dorso-temporal portions of the retina contained mostly medium and low cell counts without any definite pattern to their distribution. Both the ventro-nasal and

32 ventro-temporal portions of the retina contained mostly medium cell counts. There were scattered areas of high counts along the retinal meridian, however these counts were not distributed in a way that would allow for the establishment of an observable cell distribution pattern.

Dorsal

145

109

147 nc nc

136 160 168 164 108

145 163 107

182

e 153

178 153

120 160

123

Cells/mm² 159

128

174

132

177

138 97 126

e 170 101 e

140

134

145 132 141

162

0-480 e 73

137 123 nc nc

98 Low 153

130

111 122 138 92 170

109 68

138

126 140

95 130

119

e 107

129

132

115 36 e

116 ck 73 481-747 149

Medium 85 54 156

45

67

129 161 114

nc nc 82

136 89

79 90 74 131

57

748-1010 High 137 40

100

132

120 174

118 110 189

136 50

121

57 65

e nc nc 68 36 124

123 107

170 1011+ 195

Highest 112

112

53

41

163

126

128 nc nc e nc 82 76 151 133

197 177 127

140 131 147 140 6 nc 169 121

171 83 177 138

nc 103

nc nc 106

170

116 127

105 100

106 125 141 150 162 129 135 134 93 91 154 101 131 122 165 134 180 183

e nc nc 95 e nc

155 67

115 147

199

154

167 119

180 142 181 123

161 132 138 137 160 137 178 167

125 95 154 116 119 173

84

167

92

181

204

240

90 e 175

nc 166 161

131 158 129 158

107 121 130 165 163 167 107 156 136 160 138 159 167 175 134 109 135

147

127

139 175 69

142

207

150 124 167

142 155 142 173 163 131 164

160 171

168 177 134 182 126 100

168 128 131

154

130 31

190

112 140 145 102

163 161 108 nc

138

173

193 203

159 218 115

149 141 160 137 178 143

e e 152

85 158

e 134

156 107

130 154 63 137 1mm

141 132 146

145 192 136

165 198 106 110 146

Nasal 173 198

e on

117 168

152 162 132

127

136 nc

93 201

130 144 131 142 135 ck

197 174 137

168 154

e 179 Temporal

188 134

82 140

94 128 176 136

115 132

143

128

159 166 208

199 168 ck 140 142 109

141

159 104 151 136 133 111 131 106

168 172 141 92 149 135 147

129 e e e nc 182 171 218 163 153 120 153 Ventral

Figure 9. Topographic map of case 2626-6 (areas marked with e= edge, nc= not countable)

Retina 2626-6

This retina came from the right eye of an adult female shark. This retina showed absolutely no distinguishable pattern of cell distribution along the retinal meridian (Table 2,

Figure 9).

The dorso-nasal and dorso-temporal portions of the retina contain mostly medium cell counts. Both the ventro-nasal and ventro-temporal portions of the retina contained mostly

33 medium cell counts with some high and low counts interspersed across the retina in no distinguishable pattern.

34 Chapter Four

Discussion

The current results revealed that bonnethead sharks had some heterogeneity in retinal ganglion cell distribution. The results also showed that the bonnethead possessed a higher- density “band” of ganglion cells traversing the central potion of the retinal meridian. Finally, small areas of higher cell density in the dorso-temporal area were found in several cases.

Although a higher density “band” of retinal ganglion cells was found within the retinal meridian of this species, the ratio between cell counts within the retinal meridian and other areas outside of the retinal meridian were not appreciably different enough to warrant calling this area a visual streak. The term “band” was used in the present study, even though there was a high degree of individual variation in its overall shape between cases, because of its fairly elongated shape in all cases.

The results of this study may be explained by two factors (habitat openness and predatory behavior) that appear to be related to ganglion cell topography in sharks as well as other species.

Although bonnethead sharks used in this study lived in fairly well-lit, shallow water habitat, their surroundings were likely obstructed, to some extent, due to particulate matter in the water. This factor may be related to the relatively short length and low ratio between cell density within and outside of the “band” of higher cell density within their retinal meridian. The low ratio between cell counts within and outside of the “band” as well as the low overall density of ganglion cells within the bonnethead retina may also be explained by behavior. Though the diet of these animals is well-known, when they are most active during a 24-hour cycle is not. Therefore, the results from this study may also be explained by these sharks being predominately nocturnal in nature. If the bonnethead shark is a predominantly nocturnal predator, then a visual streak may not be necessary to aid them in prey detection. Sensitivity to light, rather than visual acuity, would likely be more important to a nocturnal predator. Thus, if this species is predominately

35 nocturnal, this may explain their lack of a visual streak. The dorso-temporal area of increased ganglion cell density found in several cases could also be potentially associated with the predatory behavior of this species.

Habitat openness

Habitat openness may influence both the length of areas of higher ganglion cell density as well as the width of these areas. Hughes terrain theory (1977) predicts that animals with a distinct visual horizon should possess a visual streak or areas of higher ganglion cell density along their retinal meridian. More specifically, species with unobstructed views of their surroundings generally possess lengthy visual streaks which extend across the entire length of their retinal meridian (Collin and Pettigrew, 1988b). Species inhabiting areas with partially obstructed views of their habitat generally possess visual streaks that do not traverse the entire length of their retinal meridian (Oliver et al., 2001).

Results of this study revealed that the “band” of increased cell density in the bonnethead shark was rather narrow in width and did not traverse the entire length of the retinal meridian.

The increase in overall ganglion cell density across the retinal meridian was expected because of the fairly-well illuminated, shallow water habitat of the bonnethead shark. Additionally, several other shark species (lemon, tiger, epaulette, small-mouth dogfish, etc.) all possessed a visual streak. Although it does not meet the requirements of a visual streak, the existence of a short and weak “band” found in the present study is consistent with predictions from Hughes terrain theory.

Why did the bonnethead sharks in the present study not possess a visual streak? One possibility may be their activity cycle. If this species is predominately nocturnal in nature, that could explain the lack of a strong visual streak as well as why their ganglion cell density is rather low considering their habitat. Nocturnal predators require a visual system that is more sensitive to light (and would possess lower ganglion cell densities across their retinal meridians) than able

36 to resolve visual images with high acuity. Another possibility may be related to the water quality of their habitat. Sharks used in this study were collected from the waters of Charlotte Harbor, which contains rather cloudy water (Humphreys and Grantham, 1995; Tomasko, 2001). Because of murkiness of the water, the view of the visual horizon was likely obstructed, to some extent, for this shark species, potentially making possession of a visual streak not useful to them.

Therefore, sharks in this habitat may not have developed a visual streak, dependent on the amount of time they spend within this habitat during the year.

A potential answer as to whether the water quality of the bonnethead sharks collected for this study or their behavioral patterns over a 24 hour-cycle affected the ganglion cell density within their retinas could be found from a comparison of bonnethead sharks living in the waters of Charlotte Harbor to those inhabiting Florida Bay. Florida Bay is located near the Florida Keys and contains less particulate matter which can interfere with overall water clarity (Cortes et al.,

1996; Cortes and Parsons, 1996). Preliminary results from mitochondrial DNA testing have also revealed that there is no significant difference between the mitochondrial DNA of bonnethead sharks inhabiting either region (Lombardi-Carlson, Cortes, Parsons, and Manire, 2003). This comparison would be necessary to ascertain whether a relatively narrow and short area of increased ganglion cell density is found species-wide or is due to nocturnal behavior or even potentially regional differences in habitat.

Behavior

Behavior may also be an influential factor affecting retinal ganglion cell topography.

This factor may influence both the width as well as the location of the visual streak and any associated specialized areas within the retina (Bozzano and Collin, 2001; Collin and Pettigrew,

1988a and 1988b).

37 In the bonnethead shark, their higher density “band” was found in a fairly central location within the retina. This higher density “band” did not transverse the entire retinal meridian for the majority of cases. The higher density “band” likely subtends vision for this species visual horizon. Nonetheless, ganglion cell density may not be high enough, within the higher density

“band”, to provide this species with anything more than increased motion detection in this area of their visual environment.

However, the areas of higher ganglion cell density located within the temporal retina were not an expected finding, and could be related to the predatory behavior of this shark species.

Species who have specialized methods of capturing prey may have specializations that provide them with a visual advantage in locating prey (Bozzano and Collin, 2001; Collin and Pettigrew,

1988a and 1988b; Oliver et al., 2001).

Six of the bonnethead sharks from this study possessed a dorso-temporal area of increased ganglion cell density. According to Collin and Pettigrew (1988b), species possessing areas of increased ganglion cell density in their temporal retina likely use these areas for prey detection. Bonnethead sharks primarily predate upon swift moving blue crabs (Cortez et al.,

1996). These crabs may be located moving along the substrate in front and slightly below the visual horizon of this shark. Therefore, an area of higher ganglion cell density located within the dorso-temporal retina may be advantageous to this shark species. An increased sensitivity to motion detection along the substrate would likely aid the bonnethead shark in locating potential prey items. Capture of these prey items would possibly then fall to other sensory systems, such as movement detected by the lateral line and electroreception of prey location detected through use of the ampullae of Lorenzini.

Although results from the lone male retina utilized in the study did not reveal any topographic pattern, possibly due to methodological issues, the lack of a definitive pattern could also be related to sexually dimorphic differences in the head shape of these sharks between males

38 and females. Further investigation regarding the retinal topography of male bonnethead sharks is necessary to establish whether this is a possibility.

Technical problems in the present study and possible solutions

No definitive pattern of retinal ganglion cell distribution was found in two of the retinas

(2625-2 and 2626-6) and considerable individual differences were observed in the rest of the retinas. This absence and variation of the retinal ganglion cell pattern may be due to methodological problems. In particular, in the field, it is possible that the eyes did not receive enough preservative to fix them properly or too much time passed before preservative was added to the eyes to conserve them. If the eyes did not receive preservative in a timely matter, this could have resulted in ischemic damage which could explain the absence of a topographic pattern in two of the cases as well as the considerable individual differences between the “band” of higher cell density as well as location and size of the dorso-temporal area between all cases.

Retinal counts in this study may have also been underestimated. The tracer DiI was to be utilized as a medium with which ganglion cell morphology in the bonnethead shark could be documented. However, DiI does not appear to be compatible with elasmobranch body chemistry and was unable to be used to document retinal ganglion cell morphology. Ganglion cell counts were then based on descriptions from both Hueter (1991) and Collin (1988). All ganglion cell counts in this study were based on darkness of the Nissl stain, cell-body shape, and presence of axon bodies. It is possible that the actual number of ganglion cells were undercounted.

To correct for both problems, it may be better to watch the sharks after collection and inject preservative into them immediately after the shark expires, instead of attempting to inject preservative into the eyes within 15 minutes of the shark expiring. As of now, sacrificing the shark and immediately harvesting the eyes still appears to be the most effective way to conduct a

39 study of retinal topography. It would also be helpful to find a suitable tracer that works well with the body chemistry of elasmobranches.

Future directions

To further elaborate upon and better understand the significance of retinal ganglion cell topography in the bonnethead shark, an investigation of the visual threshold and visual capabilities of this shark species would be beneficial. These types of studies should also take the differences in head shape between males and females into account. If this species is unable to detect targets or objects located in front and slightly below them, then vision is likely not as important to their daily survival as other sensory systems.

Conducting an analysis of retinal ganglion cell topography on bonnethead sharks living in

Florida Bay may also help to confirm the findings from this project. Findings from the same species of shark inhabiting a more illuminated and open habitat may shed light on whether or not the overall ganglion cell concentration as well as width and length of the higher density “band” is common to all bonnetheads living in Florida or varies according to location (i.e. is influenced by environmental factors). If retinal topography between the two populations of bonnethead sharks is the same, then a study investigating whether or not these sharks are nocturnal in nature would also aid in better understanding of their retinal topography. Physically measuring and behaviorally testing the extent and limits of the visual field of the bonnethead shark (from either region) would also help to ascertain whether the retinal topography of this shark could be related to its predatory behavior.

Conclusions

Even though a higher density “band” was revealed within the retinal meridian of this species, the ambiguity of the shape of this “band” as well as the low ratio between the density of

40 this “band” may signify that vision is not as important to this species as other sensory systems or that this shark could be more nocturnal than diurnal in its activity patterns. This region could also be left over from before the evolution of this species unique head shape or could be used to lower the threshold for detection of disturbances within the shark’s visual horizon. Further research regarding the visual capabilities of this species could reveal how, if at all, this area of increased cell density is utilized. Comparisons between this and other hammerhead species could also reveal whether or not retinal topography differs between the bonnethead shark and other hammerhead species. If there is a difference in retinal ganglion cell topography within the other species of hammerhead sharks, it could signify that behavioral utility is the most influential factor behind retinal ganglion cell topography in these sharks.

41 References

Ali, M.A. and M. Anctil (1974a) Retinas of the electric ray (Narcine brasiliensis) and the freshwater stingray (Paratrygon motoro). Vision Research 14: 587-588.

Ali, M.A. and M. Anctil (1974b) Giant ganglion cells in the retina of hammerhead shark

(Sphyrna lewini). Vision Research 14: 903-904.

Ali, M.A. and M. Antcil, eds. (1976) Retinas of fishes: an atlas. Spriner-Verlag, Berlin.

Arrese, C., Dunlop, S. A., Harman, A. M., Braekevelt, C. R., Ross, W. M., Shand, J., and L. D.

Beazley (1999) Retinal structure and visual acuity in a polyprotodont marsupial, the fat-tailed dunnart (Sminthopsis crassicaudata). Brain, Behavior, Evolution 53: 111-126.

Bonazzo, and Collin, S.P. (2000) Retinal ganglion cell topography in seven species of elasmobranch. Brain Behavior Evolution 55: 191-208.

Collin, S.P. (1988) The retina of the shovel-nosed ray, Rhinobatos batillum (Rhinobatidae): morphology and quantitative analysis of the ganglion, amacrine, and bipolar cell populations.

Experimental Biology 47: 195-207.

Collin, S.P. and J. D. Pettigrew (1988a) Retinal topography in reef teleosts: some species with dell-developed area but poorly developed streaks. Brain, Behavior, and Evolution. 31:269-282.

42 Collin, S. P., and J. D. Pettigrew (1988b) Retinal topography in teleosts: some species with prominent horizontal visual streaks and high-density areae. Brain, Behavior, Evolution. 31: 283-

295.

Compango, L.V. J. (1984) Sharks of the world. Carcharhiniformes. FAO Fish Synops.

Volume 4, Part2.

Cortes, E. and G.R. Parsons (1996) Comparative demography of two populations of the bonnethead shark (Sphyrna tiburo) in southwest florida. Canadian Journal of Fisheries and

Aquatic Science 53: 709-718.

Cortes, E., Manire, C.A., and R.E. Hueter, (1996) Diet, feeding habits, and diel feeding chronology of the bonnethead shark, Sphyrna tiburo, in southwest florida. Bulletin of Marine

Science 58: 353-367.

Graeber, R.C. (1978) Behavioral studies correlated with central nervous system integration of vision in sharks. Sensory Biology of Sharks, Skates, and Rays. E.S. Hodgson and R.F.

Mathewson, eds. Office of Naval Research, Arlington, Va. Pg. 11-105.

Gruber, S.H. (1977) The visual system in sharks: adaptations and capability. American

Zoologist. 17: 453-470.

Gruber, S.H., Gulley, R.L., and J. Brandon (1975) Duplex retina in seven elasmobranch species.

Bulletin of Marine Science. 25: 353-358.

43 Gruber, S.H., Hamasaki, D.I., and C.D. Bridges (1963) Cones in the retina of the lemon shark

(Negaprion brevirostris). Vision Research 3: 397-399.

Hamasaki, D.I., and S.H. Gruber (1965) The photoreceptors of the nurse shark, Ginglymostoma cirratum, and the stingray Dasyatis sayi. Bulletin of Marine Science. 15: 1051-1059.

Hebel, R. (1976) Distribution of retinal ganglion cells in five mammalian species (pig, sheep, ox, horse, dog). Anatomy and Embryology (Berl.). 150: 45-51.

Hoese, H.D. and R.B. Moore (1958) Notes on the life history of the bonnetnose shark, Sphyrna tiburo. The Texas Journal of Science 10: 69-71.

Hueter, R.E. (1988) Retinal Topography and the retinotectal projection pattern in the juvenile lemon shark (Negaprion brevirostris). Society of Neuroscience Abstracts. 14:1119.

Hueter, R.E. (1989) The organization of spatial vision in the juvenile lemon shark (Negaprion brevirostris): retinotectal projections, retinal topography, and implications for the visual ecology of sharks. Dissertation Abstracts International Part B: The Sciences and Engineering . 50: 138.

Hueter, R.E. (1991b) Adaptations for spatial vision in sharks. The Journal of Experimental

Zoology Suppliment 5: 130-141.

Hueter, R.E. and S.H. Gruber (1982) Recent Advances in studies of the visual system of the juvenile lemon shark (Negaprion brevirostris). Florida Scientist 45: 11-28.

44 Hughes, A. (1977) The topography of vision in mammals of contrasting lifestyle: comparative optics and retinal organization. Handbook of Sensory Physiology VII/5: The visual system in vertebrates. F. Crescitelli, ed. Springer-Verlag, Berlin. Pg. 613-756.

Humphreys, J. and S.B. Grantham (1995) Charlotte Harbor. Fanthom Magazine, Winter edition.

Huxlin K.R. and Goodchild A.K. (1997) Retinal ganglion cells in the albino rat: revised morphological classification. Journal of Comparative Neurology 385: 309 - 323.

Johnsen, P.B. and J.H. Teeter (1985) Behavioral responses of the bonnethead (Sphyrna tiburo) to controlled olfactory stimulation. Marine Behavior and Physiology 11:283-291.

Kajiura, S.M. and K.N. Holland (2002) Electroreception in juvenile scalloped hammerhead and sandbar sharks. Journal of Experimental Biology 205: 3609-3624.

Lima, S.M., Silveira, L.C., and V.H. Perry (1996) Distribution of M ganglion cells in diurnal and nocturnal New World monkeys. Journal of Comparative Neurology 368: 538-52.

Lombardi-Carlson, L.A., E. Cortes, G.R. Parsons, and C.A. Manire. 2003. Latitudinal variation in life-history traits of bonnethead sharks, Sphyrna tiburo, (Carcharhiniformes: Sphyrnidae) from the eastern Gulf of Mexico. Marine & Freshwater Research 54(7):875-884.

Martin, A. (1993) Hammerhead shark origins. Nature 364: 494.

45 Mass, A. M. and A. Y. Supin (2000) Ganglion cell density and retinal resolution in the sea otter,

Enhydra lutris. Brain, Behavior, Evolution. 55: 111-119.

Motta, P.J., Tricas, T.C., Hueter, R.E., and A.P. Summer (1997) Feeding mechanism and functional morphology of the jaws of the lemon shark, Negaprion brevirostri. Journal of

Experimental Biology 200: 2765-2780.

Motta, P.J. and C.D. Wilga (2000) Durophagy in sharks: feeding mechanics of the hammerhead

Sphyrna Tiburo. The Journal of Experimental Biology 203: 2781-2796.

Myrberg and Gruber (1974) The behavior of the bonnethead shark. Copeia. 2: 358-374.

Nakaya, K. (1995) Hydrodynamic function of the head in the hammerhead sharks

(Elasmobranchii: Sphyrnidae). Copeia 1995: 330-336.

Oliver, L.J., Salmon, M., Wyneken, J., Hueter, R.E., and T.W. Cronin (2000) Retinal anatomy of hatchling sea turtles: anatomical specializations and behavioral correlates. Marine and

Freshwater Behavavior and Physiology 33: 233-248.

Peterson, E.H., and M.H. Rowe (1980) Different regional specializations of neurons in the ganglion cell layer and inner plexiform layer of the California horned shark, Heterodontus francisci. Brain Research. 201: 195-201.

46 Shimizu, T., Cox, K., Karten, H., and L.R.G. Britto (1994) Cholera toxin mapping of retinal projections in pigeons (Columbia livia), with emphasis on retinohypothalamic connections.

Visual Neuroscience 11: 441-446.

Stell, W.K. (1972) The structure and morphologic relations of rods and cones in the retina of the spiny dogfish, Squalus. Comparative Biochemistry and Physiology 42A: 141-151.

Stell, W.K., and P. Witkovsky (1973) Retinal structure in the smooth dogfish, Mustelus canis: general description and light miscoscopy of giant ganglion cells. Journal of Comparative

Neurology 148: 1-32.

Stone, J. and P. Halasz (1989) Topography of the retina in the elephant Loxodonta africana.

Brain, Behavior, Evolution. 34: 84-95.

Strong, W.R., Gruber, S.H., and F.F. Snelson (1990) Hammerhead shark predation on stingrays: an observation of prey handling by Sphyrna mokkarran. Copeia. 1990: 386-340.

Tester, A.L., and S. Kato (1966) Visual target discrimination in blacktip sharks (Carcharrhinus melanopterus) and grey sharks (C. menisorrah). Pacific Science 20: 461-471.

Tomasko, D.A. (2001) Seagrass Restoration Varies in Southwest Florida's Estuaries. ERF

Conference, Tradewinds Conference Center, St. Pete Beach, Florida November 4-8.

47 Wong, R. O., Wye-Dvorak, J., and G. H. Henry (1986) Morphology and distribution of neurons in the retinal ganglion cell layer of the adult tammar wallaby – Macropus eugenii. Journal of

Comparative Neurology. 253: 1-12.

Wright, T. and R. Jackson (1964) Instrumental conditioning of young sharks. Copeia. 1964:

409-412.

48

APPENDICES

49 Appendix A: Raw data retinal counts (in cells per mm²)

Dorsal

Nasal Temporal

Ventral

Figure 10. Raw data retina 2627-1

50 Dorsal

Nasal Temporal

Ventral

Figure 11. Raw data retina 2626-4

Dorsal

885 785

724 806 733

747

Nasal Temporal

Ventral

Figure 12. Raw data retina 2626-5

51

nc 552 634 688 353 ne Dorsal 697 616 611 543 752 e nc 358 484 653 607 833 493 720 611 905 770

321 nc nc nc 439 457 602 1372 1123 1209 1mm 914 828 774 815 761 675 806 nc 448 625 634 652 575 788 955 842 693 865 819 747 815 860 828 nc nc 507 629 616 430 539 629 652 525 530 647 679 919 661 580 770 661 nc nc 629 281 720 566 598 598 625 607 616 561 770 770 661 543 607 480 435 426 570 602 nc 792 801 815 729 774 819 851 670 738 847 720 738 738 e 847 770 638 580 602 439 489 693 697 nc 761 919 1064 1068 ck 847 557 869 892 1032 715 883 928 724 838 742 770 589 589 457 534 598 516 625 770 792 779 634 801 810 774 869 752 788 1096 1163 770 996 955 933 991 507 629 593 620 543 480 602 611 561 1195 797 679 874 955 905 901 10141123 1154 1037 nc 1182 905 874 942 385 661 584 729 561 611 625 675 792 874 1032 1077 11951145 1123 1272 1023 991 797 896 634 1014 914 809 788 652 656 756 661 833 638 792 797 1023 nc 964 12721005 6751100 1240 1204 946951770969928 842 797 896 919 924 643 819 819 738 842 1050 1136 987 1028 910 nc 720 1123 1059 1136 1204 991 964 874 616 801 Nasal 806 856 774 783 1019 1127 978 1186 919 887 1105 1087 1132 1191 1055 10051299 987 973 783 742 747 833 779 715 688 656 nc 752 815 715 797 1005 1010 1055 1322 1195 15711195 1041 815 833 720 783 Temporal nc 810 697 643 901 765 738 761 742 865 955 819 1123 1200 982 1091 865 838 607 792 738 702 756 724 819 828 611 702 933 978 960 955 620 634 548 770 611 706 661 643 729 670 797 774 842 nc 584 693 566 629 616 652 nc 733 634 761 nc e 629 493 580 593 570 652 530 534 580 503 580 e 358 nc e nc 634 634 548 552 nc 435

Ventral

Figure 13. Raw data retina 2626-8

Dorsal

Nasal Temporal

Ventral

Figure 14. Raw data retina IMF1 52

Dorsal

e 905 729 638 na 756 969 nc na 557 779 503 838 1123 914 910 887 901 869 878 620 665 598 761 770 589 634 933 810 878 987 896 684 765 1028 973 742 724 611 675 765 720 na e 973 530 797 1014 779 738 724 1091 nc 910 946 828 987 788 711 670 507 765 792 589 851 1191 770 883 842 919 661 nc 860 1154 869 905 783 706 761 679 774 806 752 761 na 620 1014 792 na 652 896 1209 1299 1222 698 774 752 747 801 792 783 na 792 697 e 1023 e Temporal 815 1317 1032 892 851 756 824 842 955 892 756 652 e 887 na e 919 896 937 e 1100 1001 1032 1014 1046 1059 1159 874 1077 991 1010 11951345 e e 1177 1236 1259 1087 1068 1001 892 955 860 824 901 951 905 982 937 1336 982 Nasal 987 828 973 86 770 883 982 919 761 905 575 874 765 919 738 788 905 856 978

933 1001 1145 1340 371 801 765 905 901 557 987 742 e 883 738 924 543 847 1019 996 1458 e 575 448 435 629 512 625 724 525 489 910 466 842 480 887 697 575 na 417 742 616 774 656 430 869 742 883 838 647 806 856 e 417 774 937 575 580 928 693 580 792 629 715 797 828 779 856 1073 688 742 914 512 792 561 919 611 570 711 824 724 620 815 815 715 824 697 e 679 684 910 792 697 838 756 806 892 824 539 847 625 806 530 711 589 982 566 580 779 484 625 679 715 1023 1136 665 910 570 756 1mm 679 589 792 na 516 665 828 810 761 1046 1005 878 ne 629 nc nc nc 539 684 860 887 e e 847 602 580 656 na 607847 e e 928 856 1050 901 403 e Ventral

Figure 15. Raw data retina IMF2

Dorsal

Temporal Nasal

1mm

Ventral 53

Figure 16. Raw data retina 2625-2

Dorsal

Nasal Temporal

1mm Ventral

Figure 17. Raw data retina 2626-6

54