Aptitude in American Alligators: ecological factors affecting cognition and behavior

American alligators (Alligator mississippiensis) possess flexible cognitive abilities. Given these cognitive abilities, we hypothesized the capacity for flexible learning is significantly affected by ecological factors such as environmental temperature and contaminants. Additionally, we reasoned that the same ecological modifying inherent cognitive capacities also shape other traits in adaptive ways. In order to explore these relationships we utilized three separate studies divided between four chapters. In the first three chapters our research employed serial-reversal experiments to quantify the total number of errors in a series of discrimination problems to investigate the effect of temperature and environmental contaminates on inherit cognitive abilities. Chapter 4 is devoted to an investigation of the effect of these same environmental contaminates on thermoregulatory behavior and metabolism. Specifically, experiments presented in chapter one investigated the visual learning abilities of juvenile American alligators in order to determine the degree of complex learning. Chapter 2 describes the effect of temperature on the performance of juvenile American alligators in a spatial discrimination task by tasking animals to complete a series of ten reversal at two environmentally relevant temperatures. The successful development of both protocols led us to ask the question of whether or not we could apply a similar behavioral assessment of learning and memory abilities in animals exposed to environmental contaminants. Recognizing that organochlorines still contaminate the waters inhabited by American alligators, Chapter 3 explores the affect in ovo exposure to DDE has on learning and behavior in hatchling American alligators. Specifically, we investigated if DDE affects cognition of American alligators by comparing the performance of individuals that were exposed as embryos to the performance of control individuals in a spatial discrimination task. Organochlorines, such as DDE, can bioaccumulate and are therefore particularly problematic for top predators. Therefore, it is important to know if standard metabolism or preferred body temperatures are perturbed by DDE exposure because these changes may affect the overall health of animals, their reproductive success, and the health and growth rates of hatchlings. Therefore, in Chapter 4 we designed a study to measure changes in the thermoregulatory system and metabolism of animals exposed to DDE.

APTITUDE IN AMERICAN ALLIGATORS: ECOLOGICAL FACTORS AFFECTING COGNITION AND BEHAVIOR by Jennifer Lauren Yowell A dissertation submitted to the faculty of The University of Utah in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Biology The University of Utah August 2011 Copyright © Jennifer Lauren Yowell 2011 All Rights Reserved The University of Utah Graduate School STATEMENT OF DISSERTATION APPROVAL The dissertation of Jennifer Lauren Yowell has been approved by the following supervisory committee members: Colleen G. Farmer , Chair 06/06/2011 Date Approved Raymond P. Kesner , Member 06/03/2011 Date Approved Todd Uriona , Member 06/06/2011 Date Approved Franz Goller , Member 06/10/2011 Date Approved David R. Carrier , Member Date Approved and by Neil Vickers , Chair of the Department of Biology and by Charles A. Wight, Dean of The Graduate School. ABSTRACT American alligators (Alligator mississippiensis) possess flexible cognitive abilities. Given these cognitive abilities, we hypothesized the capacity for flexible learning is significantly affected by ecological factors such as environmental temperature and contaminants. Additionally, we reasoned that the same ecological modifying inherent cognitive capacities also shape other traits in adaptive ways. In order to explore these relationships we utilized three separate studies divided between four chapters. In the first three chapters our research employed serial-reversal experiments to quantify the total number of errors in a series of discrimination problems to investigate the effect of temperature and environmental contaminates on inherit cognitive abilities. Chapter 4 is devoted to an investigation of the effect of these same environmental contaminates on thermoregulatory behavior and metabolism. Specifically, experiments presented in chapter one investigated the visual learning abilities of juvenile American alligators in order to determine the degree of complex learning. Chapter 2 describes the effect of temperature on the performance of juvenile American alligators in a spatial discrimination task by tasking animals to complete a series of ten reversal at two environmentally relevant temperatures. The successful development of both protocols led us to ask the question of whether or not we could apply a similar behavioral assessment of learning and memory abilities in animals exposed to environmental contaminants. Recognizing that organochlorines still contaminate the waters inhabited by American alligators, Chapter 3 explores the affect in ovo exposure to DDE has on learning and behavior in hatchling American alligators. Specifically, we investigated if DDE affects cognition of American alligators by comparing the performance of individuals that were exposed as embryos to the performance of control individuals in a spatial discrimination task. Organochlorines, such as DDE, can bioaccumulate and are therefore particularly problematic for top predators. Therefore, it is important to know if standard metabolism or preferred body temperatures are perturbed by DDE exposure because these changes may affect the overall health of animals, their reproductive success, and the health and growth rates of hatchlings. Therefore, in Chapter 4 we designed a study to measure changes in the thermoregulatory system and metabolism of animals exposed to DDE. iv This work is dedicated to my grandfather, Albert Araneo, for all his love and support. He is the foundation of my family and without him we would not be who we are. TABLE OF CONTENTS ABSTRACT......................................................................................................…..iii LIST OF FIGURES ..........................................................................................…vii ACKNOWLEDGMENTS...............................................................................….…ix Chapter 1 INTRODUCTION.........................................................................................1 References......................................................................................……...10 2 "RED LIGHT, GREEN LIGHT" A VISUAL REVERSAL STUDY EXPLORING LEARNING AND MEMORY IN JUVENILE AMERICAN ALLIGATORS…………………………………………………………………..13 Introduction….……………..……………………………………………...........14 Materials and Methods..……………………………………………...……….19 Results.……………………………………………………..…………….........24 Discussion...…………………………………………...……………..…..........24 References……………………………………………………...…...…...........30 3 THE EFFECT OF TEMPERATURE ON REVERSAL LEARNING IN JUVENILE AMERICAN ALLIGATORS (ALLIGATOR MISSISSIPPIENSIS).………………………………………………………….33 Introduction……………..………………………………………………...........34 Materials and Methods..………………………………………………………41 Results.………………………………………………………………..….........48 Discussion...…………………………………………………………..…..…...50 References………………………………………………………...……...……61 4 EFFECT OF DDE ON LEARNING IN AMERICAN ALLIGATORS ………64 Introduction….……………..……………………………………..……..……..66 Materials and Methods..…………………………………………...……...….71 Results.………………………………………………………..………..………79 Discussion...…………………………………………………..………...….….81 References………….………………………………………………..………...91 5 EFFECT OF DDE ON METABOLISM AND THERMOREGULATION IN AMERICAN ALLIGATORS..……………………………………………….....99 Introduction….……………..……………………………..…………………..100 Materials and Methods..…………………………………..………………...105 Results.………………………………………………………….……….……109 Discussion...………………………………………………..………………...111 References……………………………………………………...…………….130 6 CONCLUSIONS......................................................................................135 viii LIST OF FIGURES Figure Page 2.1. Arial view of visual reversal apparatus..........................................................27 2.2 Mean total errors per reversal with SEM........................................................28 2.3 Full width (number of reversals) at half the difference between the initial number of errors and the maximal number of errors………………………………29 3.1 Linear regression of spatial temperature data showing a decrease in the number of errors committed by an animal as days within a reversal increased, red line includes all days regardless of whether all animals completed ten trials, black line includes only those days were all animals completed all ten trials…...56 3.2 Mean total errors per reversal, first set of reversal was run at one temperature, then for the second set of 10 reversal the temperature treatment had been switched, a clear trend of decreasing errors per days spent in each reversal was observed, alligators made fewer total errors at 22°C than at 32°C regardless of temperature regime (p=0.0282)……………………………………...57 3.3 Logistic regression analyzing the relationship between days in reversal and successful trials.……………………………………………………………………….58 3.4 Logistic regression analyzing the relationship between reversal number and successful trials…………………………………………………….………………….59 3.5 Logistic regression analyzing the relationship between temperature and successful trials…………………………………………………………………..……60 4.1 Aerial view of T-maze…………………………..…………………………....…...87 4.2 Total errors per blocks of five trials in Reversal 0 and Reversal 1……….….88 4.3 Total errors per trial in Reversal 0………………………………………....……89 4.4 Total errors per trial in Reversal 1…………….…………………………...……90 5.1 Quantile quantile plot of %O2 use for animals naturally exposed to DDE versus control animals from a clean environment……….………………….……120 5.2 Pooled residual plot of %O2 use for animals naturally exposed to DDE versus control animals from a clean environment……………………….………….…….121 5.3 Quantile quantile plot of % CO2 use for animals naturally exposed to DDE versus control animals from a clean environment………………………………..122 5.4 Polled residual plot of % CO2 use for animals naturally exposed to DDE versus control animals from a clean environment……………………………….123 5.5 Quantile quantile plot of heat production (°C) for animals naturally exposed to DDE versus control animals from a clean environment…………………….……124 5.6 Mass corrected logarithmic plot for SMR (ml 02/min) for control animals from a clean environment…………………………………….……………..……………125 5.7 Mass corrected logarithmic plot for SMR (ml 02/min) for animals naturally exposed to DDE……………………………………..……………………………….126 5.8 and 5.9 Residual values for mass corrected logarithmic plot for SMR (ml 02/min) for control animals from a clean environment and animals naturally exposed to DDE……..……………………………………………………………….127 5.10 Histogram of the preferred body temperature (°C) modes for control animals from a clean environment and animals naturally exposed to DDE……………………………………………………………………………………128 5.11 Relationship between mass (g) on preferred body temperature (°C) of control animals from a clean environment and animals naturally exposed to DDE……………………………………………………………………………………129 viii ACKNOWLEDGMENTS The authors would like to thank R. Elsey of the Rockefeller Wildlife preserve and L. Guillette of Florida State University, Gainesville for all her guidance and for providing animals. We would also like to acknowledge J.T. Olds for writing the reversal and data collection program. Finally, we would like to thank D. Bain, S. Ingebretsen, M. Miller, A. Thompson and E. Taylor for their assistance with training and data collection. Animals were housed under the protocol number 06-12008 and 08-06003. CHAPTER 1 INTRODUCTION Several features of cognition have evolved in similar ways in both primates and several phyletically-independent species of birds (Lefebure et al., 2004). Furthermore, in birds and mammals specialized behaviors such as tool use and parental care are linked to the evolution of cognitive abilities (Lefebure and Sol, 2008). Such observations lead one to wonder if there are ecological and life-history factors that have influenced the convergent evolution of cognitive abilities in two such divergent lineages. Examples of such ecological pressures or life history patterns include food type, social groups, and climate. Species that take advantage of a variety of feeding sources, or have to use more complicated strategies to locate or handle food, will rely on more sophisticated adaptability patterns. Furthermore, food type may force such species to rely more heavily on innovative behaviors in order to access and survive on complex food sources. For example, caching behaviors require extensive spatial memory in order to store and retrieve food (Harvey et al., 1980). Additionally, predatory behaviors require a greater ability to pursue, detect and manipulate prey (Glitterman, 1986; Huber et al., 1997). A second life history pattern that has influenced the evolution of cognitive abilities in birds and mammals is sociality. The size of a social group can influence cognitive abilities 2 because larger groups require greater cognitive abilities in order to keep track of group members and the social interactions and relationships between group members (Dundar, 1998). Furthermore, sociality requires group members to process and correctly respond to other group members using either visual or vocal communication signals. Such communication requires higher brain function and cognitive abilities. Finally, climate has been hypothesized to play a role in the evolution of complex behavior and cognitive abilities. Specifically, species from temperate environments have evolved to cope with an environment that can be dramatically different from one season to the next. For example, these species deal with winter temperatures that are much lower than the temperatures experienced in the summer months. Furthermore, these environments may be much harsher due to short days and food shortages that also accompany low temperatures. Innovative behavior and the ability to adapt to novel situations could improve the ability of these animals to survive and therefore increase species persistence. While these patterns are observed in birds and mammals, ectotherms live in these same environments and are sensitive to some of the same evolutionary pressures as their endothermic neighbors. Ecological factors such as food type and social group size could influence the cognitive abilities of ectotherms in the same manner as they influence birds and mammals. For example, predators will need to be able to learn how to track and manipulate sparse and evasive prey in order to be successful, whether the predator is an ectotherm or endotherm. Similarly, social groups will still require a greater amount of neural ability to process the interactions between group 3 members regardless of whether the group members are ectotherms or endotherms. However, ecological factors such as environment and climate may have a very different effect on the cognitive abilities of ectotherms. Studies on ectotherms may provide new insights into the evolution of cognitive abilities because certain ecological and environmental aspects will affect ectotherms in ways that birds and mammals are insensitive. Crocodilians make a good study organism for cognitive studies because they share certain life history factors and specialized behaviors with birds and mammals (i.e., parental care, social groups and predatory behavior). However, as ectotherms they will offer unique insights into the evolution of complex behavior, adaptability to novel situations and cognitive abilities. Crocodilians display a repertoire of complex vocal and behavioral communication cues (Modha, 1967; Garrick and Lang, 1977; Garrick et al., 1978). Both vocal and behavioral communication cues are highly developed and important for social interactions including sexual competition, territory establishment, mate selection and copulation (Garrick and Lang, 1977). Such complex interactions lend support to the idea that this species poses flexible learning abilities. Additionally, female alligators return to the same nest site at the beginning of each reproductive cycle (Elsey et al., 2008), requiring the ability to learn and remember the location of these nest sites. With a complex social structure, long-term territory establishment, parental care that includes nurturing and rearing young and adaptability to their environment, it is clear that alligators 4 have the capacity to learn and remember. However, the degree of this learning ability has not been suitably tested in the wild or in laboratory captive animals. The overarching hypothesis of this thesis is that American alligators (Alligator mississippiensis) possess flexible cognitive abilities. Specifically, American alligators rely on visual and spatial discrimination abilities in their behavior and learning abilities, even when reared in captivity. We approached our investigation with three separate studies, described in the four chapters below. In the following three chapters, our research employed reversal problems to quantify the total number of errors in a series of discrimination problems with chapter five devoted to an investigation of thermoregulatory behavior and metabolism. Progressive improvement in the number of errors committed during a series of discrimination tasks can be taken as a measure of learning, and is observed in a variety of species (Bitterman, 1965a; 1965b). Reversal problems are a valuable and established tool for evaluating and comparing the learning abilities of different species (Stettner et al., 1967). Futhermore, reversal problems lend insight into general problem-solving abilities "that transcends behavioral domains and different ecological demands" pg.136 (Lefebure et al., 2004). The observation that various species perform differently in these tasks lends insight into phylogenic differences in behavior and learning abilities (Stettner et al., 1967). Across several orders of birds it has been shown that the total number of errors committed in a series of visual discrimination problems declines as an individual's experience with the problem increases. Chickens (Gallus gallus domesticus), pigeons (Columba livia domestica) and 5 crows (Corvus americanus) all progressively improve in the number of errors committed during a series of visual discrimination tasks. However, in quail (Colinus virginianus) experience does not seem to affect the number of errors and therefore, quail do not reflect the pattern of decreasing errors seen in other bird species. We hypothesize that American alligators (Alligator mississippiensis) display a pattern of progressive improvement in reversal problems, similar to the pattern seen in many bird species including Corvids. Experiments presented in Chapter 2 investigated the visual learning abilities of juvenile American alligators reared in the laboratory. Specifically, the research established optimal conditions for training and reversal using a food reward, ascertained the visual discrimination ability of the American alligator, and finally, determined the degree of complex learning using a serial reversal approach. Very little is known about the ability of American alligators to participate in these types of tasks, and the laboratory setting permits testing of the innate ability of the subject. We believe that the visual discrimination task explained in Chapter 2 closes this gap. Chapter 3 describes the effect of temperature on the performance of juvenile American alligators in a spatial discrimination task. Based on information obtained in other species (Reid, 1957; Warren et al., 1960; Eskin and Bitterman, 1961; Northcutt and Heath, 1973), we investigated the effect of temperature on the spatial discrimination ability of American alligators. This assessment of spatial learning incorporated two different temperature treatments - one at the lower end and another at the upper end of the American alligator's preferred 6 activity range. The results of this study present new data and a novel approach to quantifying learning in a predatory reptile. American alligators are capable of learning to perform a lever pressing action in order to receive a food reward (Araneo and Farmer, unpubl.). Furthermore, American alligators are capable of discriminating two stimuli on the basis of visual cues (Araneo and Farmer, unpubl.). As Krekorian et al. (1968) demonstrated, desert iguanas perform better in a learning task when close to their preferred body temperature. Additionally, learning in this species appeared to be less effective at cooler temperatures (Krekorian et al., 1968). American alligators have a preferred body temperature, after feeding, of 30°C (Farmer et al., 2008). We predicted that American alligators would show differences in a spatial discrimination study based on different temperature regimes. Temperature affects various aspects of learning in a variety of species (Roussel et al., 1982). Even with extensive training sessions, both hypothermia and hyperthermia perturb memory acquisition in the rodent (Roussel et al., 1982). Furthermore, in rats a core body temperature increase of 2 or 3°C can cause amnesia (Misanin et al., 1979). In hummingbirds, it is hypothesized that the associated drop in temperature during torpor is incompatible with memory consolidation (Roth et al., 2010). Such a trade-off, between energy conservation and memory consolidation, implies that in the hummingbird, memory consolidation may not be temperature compensated. Therefore, if juvenile American alligators resemble the pattern observed in mammals and the 7 hummingbird, we might expect to see an effect of temperature on the spatial discrimination ability of juvenile American alligators. Whereas in Chapter 2 we successfully designed a learning protocol that allowed us to investigate the visual discrimination ability of juvenile American alligators, in Chapter 3 we applied this knowledge to a second learning assay that allowed us to investigate the effect of temperature on the spatial discrimination ability of this species. The effective establishment of both of these protocols led us to ask the question of whether or not we could apply a similar behavioral assessment of learning and memory abilities in animals exposed to environmental contaminants in ovo (Chapter 4). The synthetic pesticide DDT [1,1,1-trichloro-2,2-di(4-chlorophenyl)ethane] is used throughout the world, exposing humans and wildlife to this organochlorine (Kleinow et al., 1987). This pollutant and its breakdown products are linked to a variety of morphological, developmental, and physiological abnormalities (Schantz and Widholm, 2001). Even though it is known that these compounds derail normal cerebral function in birds and other species, little is known about the effects of organochlorines on the crocodilian brain (Hunt and Hunt, 1977; Luoma, 1992; Iwaniuk et al., 2006)). Recognizing that organochlorines still contaminate the waters inhabited by American alligators, the effect of these pollutants on alligator development, survival and species preservation needs addressing. Chapter 4 explores the effect in ovo exposure to DDE (1,1-dichloro-2, 2-bis[p-chlorophenyl] ethylene) has on learning and behavior in hatchling American alligators. Specifically, we 8 investigated if organochlorines, such as DDE, affect cognition of American alligators by comparing the performance of individuals that were exposed as embryos to an organochlorine, to the performance of control individuals in a spatial discrimination task. Understanding the effects of organochlorine exposure on crocodilian cognition and behavior is important in a number of ways. Crocodilians have complex social behaviors including, territory defense, parental care and nest site fidelity. These behaviors may be critical for the health of hatchlings and persistence of the species. An individual's fitness depends on a suite of traits that interact with the environment. How integrated phenotypes evolve that are complex and multifunctional is a central question at the frontier of evolutionary biology. We hypothesized that there would be a relationship between the capacity for flexible learning, and environmental temperature and contaminants, reasoning that the multiple ecological drivers of cognitive capacities will also shape other traits in adaptive ways. For example, if enhanced cognitive abilities increase the amount of food an individual procures, there may be selective benefits to the co-evolution of more rapid rates of growth and higher body temperature set-points. The way these phenotypes intertwine could be derailed by exposure in ovo to environmental contaminants. Therefore, in Chapter 5 we designed a study to measure changes in the thermoregulatory system and metabolism of animals exposed to DDE. Organochlorines, such as DDT, can bioaccumulate and are therefore particularly problematic for top predators. Furthermore, many crocodilians live in regions of the world where DDT continues to be used to 9 combat malaria or in areas where the breakdown products of DDT, DDE and DDD (1,1-Bis(p-chlorophenyl)-2,2-dichloroethane), remain in the ecosystem. Numerous aspects of poikilotherm metabolism are affected by environmental temperature (Rome, 1990;; Logue et al., 2000; Somero, 2004; Guschina and Harwood, 2006; Bicego et al., 2007). Therefore it is important to know if standard metabolism or preferred body temperatures are perturbed by DDE exposure because these changes may affect the overall health of animals, their reproductive success, and the health and growth rates of hatchlings. 10 References Bicego, K.C., R.C.H. Barros, and L.G.S. Branco. 2007. Physiology of temperature regulation: comparative aspects. Comparative Biochemistry and Physiology, Part A 147: 616-639. Bitterman, M.E. 1965a. Phyletic differences in learning. American Psychologist 20: 396-410. Bitterman, M.E. 1965b. The evolution of intelligence. Scientific American 212: 92-100. Dally, J.M., N.J. Emery, and N.S. Clayton. 2005. The social suppression of caching in western scrub-jays (Aphelocoma californica). Behaviour 142: 961-977. Dundar, R.I.M. 1998. The social brain hypothesis. Evolution and Anthropology 6: 178-190. Elsey, R.M., P.L. Trosclair III, and T.C. Glenn. 2008. Nest-site fidelity in American alligators in a Louisiana coastal marsh. Southeastern Naturalist 7(4):737-743. Farmer, C. G., Uriona, T. J., Steenblik, M., Olsen, D. and K. Sanders. 2008. The right-to-left shunt of crocodilians serves digestion. Physiological and Biochemical Zoology 81:125-137. Garrick, L. D., and J.W. Lang. 1977. Social signals and behaviors of adult alligators and crocodiles. American Zoologist 17:225-239. Garrick, L. D., J.W. Lang, and H.A. Herzog Jr. 1978. Social signals of adult American alligators. Bulletin of the American Museum of Natural History 60(3):153-192. Glitterman, J.L. 1986. Carnivore brain size, behavioral ecology, and phylogeny. Journal of Mammalogy 67:23-36. Guschina, I.A., and J.L.Hardwood. 2006. Mechanisms of temperature adaptations in poikilotherms. FEBS Letters 580:5477-5483. Harvey P.H., Clutton-Brock, T.H. and G.M. Mace. 1980. Brain Size and ecology in small mammals and primates. Proclomation of National Academy of Science USA 77: 4387-4389. 11 Huber, R., M.J. van Staaden, L.S. Kaufman, and K.F. Liem. 1997. Microhabitat use, trophic patterns, and the evolution of brain structure in African cichlids. Brain, Behavior and Evolution 50:167-182. Hunt, G. L. and M.W. Jr., Hunt. 1977. Female-female pairings in western gulls (Larus occidentalis) in southern California. Science 196:1466-1467. Iwaniuka, A. N., T. Koperskia, K.M. Chengc, J.E. Elliottd, L.K. Smithe, L.K. Wilson, L.K., and D.R.W. Wyliea. 2006. The effects of environmental exposure to DDT on the brain of a songbird: changes in structures associated with mating and song. Behavioral Brain Research 173(1):1-10. Kleinow, K. M., Melancon, M.J. and J.J. Lech. 1987. Biotransformation and induction: Implications for toxicity, bioaccumulation and monitoring of environmental xenobiotics in fish. Environmental Health Perspective. 71:105-119. Krekorian, C.O., V.J. Vance, and A.M. Richardson. 1968. Temperature-dependent maze learning in the desert iguana, Dipsosaurus dorsalis. Animal Behavior 16:429-436. Lefebure, L., S.M. Reader, and D. Sol. 2004. Brains, innovations and evolution in birds and primates. Brain, Behavior and Evolution 63:233-246. Lefebure, L. and Sol D. 2008. Brains, lifestyles and cognition: are there general trends? Brain, Behavior and Evolution 72:135-144. Logue, J.A., A.L. De Vries, E. Fodor, and A.R. Cossins. 2000. Lipid compositional correlates of temperature-adaptive interspecific differences in membrane physical structure. The Journal of Experimental Biology 203:2105- 2115. Luoma, J. R. 1992. New effect of pollutants: hormone mayhem. New York Times. May 24. Misanin, J.R., R.E. Vonheym, S.W. Bartelt, W.L. Boulden, and C.F. Hinderliter. 1979. The effect of hyperthermia on memory in rats. Physiology and Psychology 7:339-344. Modha, M. L. 1967. The ecology of the Nile crocodile on central island, Lake Rudolf. East African Wildlife Journal 5:74-95. Northcutt, R.G. and J.E. Heath. 1973. T-maze behavior of the Tuatara (Sphenodon punctatus). Copeia 1973(3):617-620. Reid, R.L. 1957. Discrimination-reversal learning in pigeons. Journal of Comparative Physiology and Psychology 51:716-720. 12 Rome, L.C. 1990. Influence of temperature on muscle recruitment and muscle function in vivo. American Journal of Physiology 259:210-222. Roth, T.C., N.C. Rattenborg, and V.V. Pravosudov. 2010. The ecological relevance of sleep: the trade-off between sleep, memory and energy conservation. Philosophical Transactions of the Royal Society B 365:949-959. Roussel, B., P. Turrillot, and K. Kitahama 1982. Effect of ambient temperature on learning. Journal of Physiology and Behavior 28:991-993. Schantz, S. L., and J. J. Widholm. 2001. Cognitive effects of endocrine-disrupting chemicals in animals. Environmental Health Perspectives 109(12):1197-1206. Somero, G.N. 2004. Adaptation of enzymes to temperature: searching for basic strategies. Comparative Biochemistry and Physiology, Part B 139:321-333. Stettner, L.G., W.J. Schultz and A. Levy. 1967. Successive reversal learning in the bob-white quail (Colinus virginianus). Animal Behaviour 15(1) :1-5. Warren, J.M. 1960. Reversal learning by paradise fish (Macropodus opercularis). Journal of Comparative and Physiological Psychology 53(4):376-378. CHAPTER 2 "RED LIGHT, GREEN LIGHT" A VISUAL REVERSAL STUDY EXPLORING LEARNING AND MEMORY IN JUVENILE AMERICAN ALLIGATORS Abstract Learning abilities of juvenile American alligators (Alligator mississippiensis) were studied by measuring performance in a serial reversal experiment under controlled laboratory conditions. Ten juvenile American alligators were trained to push a colored Plexiglas target. Once captured, this behavior was reinforced with a food reward. After this behavior was established as a stable and reliable pattern, it was utilized in a learning problem. Alligators were tasked to discriminate between two visually distinct stimuli. The positive stimulus was rewarded, while a response to the negative stimulus was neither rewarded nor punished. Upon reaching a pre-Determined criterion, the signs of the discriminanda were reversed in a series of additional trials. A clear trend of decreasing errors per reversal was observed, indicating mastery of the learning task and behavioral flexibility when utilizing an acquired physical task. We 14 conclude that juvenile alligators will train successfully to perform a learned visual discrimination task when offered a food reward. Introduction Scientists and the general public have always been interested in how different species learn and what anatomical and physiological attributes correspond to learning differences or similarities. Unfortunately, there is no adequate way of ordering species in terms of "intelligence," simply because, to date no one benchmark can be relied upon to measure or represent an index for intelligence (Bitterman, 1965a). Attempts to discover such a benchmark trait have failed, most likely because such studies have relied on simple memory tasks that are unable to elucidate details of problem solving and any phylogenetic hierarchies (Bitterman, 1965a). However, more complex dynamic learning tasks (i.e. serial reversal problems) can provide insights when the performance of different species is compared (Bitterman, 1965b). A common method used to assess the ability of an animal to learn is to train the individual to perform a task and then repeatedly change the task that is required of the animal. Such repeated changes allow one to observe how quickly the animal learns a new behavior. A habit-reversal experiment is one such assay, as it relies on the same basic methodology (Bitterman, 1965b). In a habit-reversal experiment animals are presented with two stimuli that are either visually or spatially distinguishable. One of the two stimuli is assigned to be the "correct" choice and consistently produces a food reward when selected by the animal. Visual problems reward the correct stimulus regardless of position. Spatial 15 problems reward a correct location regardless of what stimuli are present at the location. Throughout the experiment, animals are rewarded for choosing the pre-determined "correct" choice. Once the animal reaches a predetermined criterion of correct choices, the discriminanda are reversed and the negative stimulus is now given a positive sign and rewarded while the previously positive stimulus is now given a negative sign and if selected no longer produces a reward. Experimenters collect data on the total number of errors committed during each reversal. Animals of various taxa, including pigeons, some fish species, turtles, chickens, and rats show progressive improvement and a decrease in the number of errors committed during each reversal (Gatling, 1951; Reid, 1957; Wodinsky and Bitterman, 1957; Bitterman et al., 1958; Warren et al., 1960; Eskin and Bitterman, 1961; Gonzalez et al., 1964; Stearns and Bitterman, 1965; Setterington and Bishop, 1967; Mackintosh and Cauty, 1971). The typical trend is an initial mastery of the problem, a dramatic increase in errors during early reversals and a steady improvement with additional reversals. Errors may increase during early-reversals, as animals tend to persist in selecting the stimuli that previously produced a reward, but as the animal's experience continues, its selection habit becomes more flexible (Bitterman, 1965b). However, the results in fishes are complicated because studies of a number of species have failed to show progressive improvement while other species show improvement (Wodinsky and Bitterman, 1957; Bitterman et al., 1958; Warren, 1960; Behrend et al., 1965; Behrend and Bitterman, 1967; Setterington and Bishop, 1967; Mackintosh and Cauty, 1971). These contradictory studies in fishes may imply 16 that the progressive improvement observed in some fish species only appears in fishes under a narrow range of restrictive experimental or reward conditions (Engelhardt et al., 1973). Crocodilians are interesting organisms for studying learning and memory because of their phylogenetic relationship with birds; a group well characterized for learning and memory (Bitterman, 1965b). The crocodilians appeared in the late Cretaceous Period and they are the sole surviving lineage of a prominent clade of archosaurs known as the Crurotarsi (Brusatte, 2009). The Crurotarsi radiated widely in the Early and Middle Triassic and contained many morphologically diverse forms, such as the armored herbivorous aetosaurs, the ostrich-like Effigia and Shuvosaurus, large terrestrial carnivorous forms such as the "rauisuchians," as well as semi-aquatic crocodile-like forms such as the phytosaurs. The sister lineage of the Crurotarsi, the Avemetatarsalia, also radiated widely in the Triassic and includes the remarkable number of morphologically diverse and successful animals such as pterosaurs and dinosaurs. Thus, an investigation of the learning ability of modern crocodilians and their specialized brain function can be compared to the learning capacity and brain function in their sister taxon, birds, to provide clues about learning capacities in the basal archosaur order and about how these capacities changed as the lineage radiated. Direct fossil evidence of learning capacities is extremely rare and should be substantiated with a thorough understanding of the capacities of the extant lineages. For example, studies of an endocast of an allosaurus brain suggest that 17 the neuroanatomy of this lineage of dinosaurs was more similar to that of modern crocodilians than to their close relatives the avian dinosaurs (Rogers, 1999). A major difference between the brains of modern birds and crocodilians is that birds have greatly reduced olfactory organs and an enlarged telencephalon, which is suggested to impart greater neuronal complexity such as a capacity for somatosensory processing, whereas allosaurus retained the large olfactory organs and small forebrain (Rogers, 1999). The expanded forebrain is suggested to have enabled birds to have more plastic behavior and thus, if allosaurus had complex behaviors, these behaviors would have been highly structured and not especially flexible (Rogers, 1998). However, although crocodilians have a small telencephalon relative to hindbrain structures, it is not fully known how plastic or structured the behavior of crocodilians is. Crocodilians display a repertoire of complex vocal and behavioral communication signals (Modha, 1967; Garrick and Lang, 1977; Garrick et al., 1978). Both vocal and behavioral cues are used during sexual competition, territory establishment, mate selection and copulation (Garrick and Lang, 1977). This complex social interaction may select for flexible learning abilities. Furthermore, female alligators show nest fidelity from year to year, returning to the same nest site at the beginning of each reproductive cycle (Elsey et al., 2008), which may require the females to learn where the good sites are located and then remain faithful to them. Very little is known about the ability of crocodilians to participate in learning tasks. The spatial discrimination ability of the spectacled Caiman (Caiman crocodiles) has been explored (Williams, 1967; Williams, 1968; 18 Northcutt and Heath, 1971). While details of these studies differed slightly, the basic design was constant and utilized a modified T-maze to determine the ability of the spectacled Caiman to reverse a habit formed in a spatial problem. Williams (1967 and 1968) and Northcutt and Heath (1971) observed a decrease in the number of trials required for Caimans to learn a new arm after each reversal. Williams (1968) explored the dominance of spatial versus visual cues with spectacled Caimans. Williams determined that when Caimans were trained to respond to a spatial cue and then presented with a choice point where visual and spatial cues were both available, animals tended to persist in a habit based on the spatial cues instead of incorporating the use of visual cues into their performance in the spatial learning task (Williams, 1967). Gossette and Hombach (1969) compared the spatial ability of Alligator mississippiensis to Crocodylus acutus. Both species were asked to discriminate based on location and both species showed progressive improvement, although the sample size used in this study (N=4) was small. Furthermore, Davidson studied the spatial discrimination ability of American alligators utilizing a spatial T-maze and the animals were motivated by escape from heat (Davidson, 1966). The floor of a metal T-maze was heated to 40.56°C, a temperature exceeding the upper lethal temperature of 38°C of alligators (Colbert et al., 1946). Although core body temperature was not measured, it is possible this protocol severely stressed and adversely affected the performance of animals in a learning task. Another potential problem with the Davidson study is that internal body temperature could affect the performance of these animals in such a task (Heath, 19 1965). The current work explores the ability of juvenile American alligators to perform a visual, rather than spatial, discrimination task when tested without the use of heat or other punishments. To our knowledge, the visual discrimination ability of the American alligator has not previously been studied. Materials and Methods All animals were part of a single clutch procured from the Rockefeller Wildlife Preserve, Grand Chenier LA., at approximately 3 weeks of age in September 2006. Animals were housed in plastic tanks, containing an equal depth of water as the reversal apparatus. Previous to reversal training, heaters were introduced into animal tanks to maintain water temperature between 30°C 􀀀 5°C. Tanks were housed in an environmentally controlled room with an air temperature ranging from 25 °C to 28° C. At 2 years of age, seven Alligators from this clutch were randomly divided into two groups, one group of four individuals and one group of three individuals (Groups A and Group B respectively). Group A and B began reversal training on September 8th, 2008 and October 10th, 2008 respectively. Group A was trained to respond positively to the color red during Reversal 0 while Group B was trained to respond positively to the color green during Reversal 0. These individuals ranged in size from 1.7 kg to 1.30 kg and were trained to perform a simple task when presented with two visually distinguishable discriminanda in order to receive a food reward. The reversal apparatus was based on the design Bitterman (1965a) used to explore the visual discrimination ability of Tilapia macrocephala, with modifications making it more appropriate for the modality of the alligators. The 20 reversal apparatus (Figure 2.1) consisted of a 75-gallon glass aquarium divided into a start box and a choice point. At the choice point two Plexiglas targets were present, each target had been outfitted with a Tricolor- LED and could be illuminated with one of three different colors. Each target was attached to a single magnetic reed switch. By pushing the target with its nose, the alligator tripped the magnetic switch and completed a circuit causing an automatic feeder to distribute a food reward. The food reward was distributed in the back of the start box, not near the targets. The aquarium was filled with water to a depth of 12.7 cm and maintained at a temperature of 32°C 􀀀 5°C by a 250 Watt submersible glass aquarium heater. Reversal training was divided into three progressive steps, each step representing a more difficult task, culminating in reversal-learning and data collection. The goals of the first step of training were to introduce study animals to the experimental routine and to train them to come to expect food only when inside the glass habit-reversal aquarium and nowhere else. During this first step of training, a single individual was selected at random and placed in the glass habit-reversal aquarium. During step one, each Plexiglas target was illuminated with the color blue, a neutral stimulus. This individual was then given 2 Mazuri brand commercial alligator chow pellets at a time until ten pellets were consumed or an hour had elapsed. Individuals were only fed what they consumed in the glass aquarium. This was repeated 3 days a week until the alligator was consuming 10 pellets within the hour. Once that criterion was met, the second step of training began during the next scheduled day of training. 21 In step two, as in step one, animal housing and handling procedures were kept consistent. Each target continued to be illuminated with the neutral stimulus. However, now each target was baited with a single pellet of food. Each alligator was given a half-hour to consume 10 pellets. If 10 pellets had not been consumed within 30 minutes the alligator was removed from the glass aquarium and returned to the home cage. The experimenter checked on the animal every 5 minutes, if a pellet had been consumed, the experimenter reset the trial by encouraging the alligator to walk back to the start box, opaque dividers were then put in place, the target was baited with an additional pellet of food, and the experimenter replaced the aquarium lid and removed the dividers. This continued until the individual had met the criterion of consuming 10 pellets in a half-hour. Once that criterion was met, the third step of training began during the next scheduled day of training. The goal for the third step of training was to train the alligators to actually push the Plexiglas targets. When actively searching for food, juvenile alligators move their heads from side to side. If food is not easily found, this movement becomes more aggressive as the individual appears to become frustrated. By training them to expect food to be present in front of the targets, we had hoped that this exploratory behavior could be captured and utilized to push the targets. Just as before, both targets were illuminated with the neutral stimulus. Once in the tank individuals immediately approached the target. As they discovered food was not easily found, the alligators began actively searching for food, which eventually became forceful enough to displace the Plexiglas target. The 22 experimenter rewarded the individual in the start box. A single press of target defined a trial. After each trial the experimenter reset the apparatus just as in step two. After a few repetitions it appeared the alligator began to associate a food reward with the action of pushing the target. During step two both targets were baited in an attempt to train the alligators to come to expect food from both targets. In step three, both targets were rewarded for the same reason, the training in step two carried over into step three. Alligators showed very little spatial preference and actively explored and engaged both the left and right target with equal frequency. A single push of a target was defined as a trial. Once the criterion of completing ten trials in a half hour was met, data collection began during the next scheduled day of training. In step three of the training, and during data collection, when the target was pushed an automatic feeding apparatus was briefly activated. An LED and magnetic switch on each target was wired to a circuit board. An Arduino (Ivrea, Italy) processing chip installed on the circuit board allowed the LED and magnetic switch to communicate with a computer. An automated data collection program was written using Python software. This program was used during each trial and recorded which magnetic switch was activated, left or right, and what color the LED was projecting and when the switch was activated. During reversals, when the LEDs were illuminated with red or green, the computer randomly selected which color would be located at the left and right target with the same color never appearing in the same location more than three times in a row. Data collection was also automated in the same manner. 23 After completing the training protocol, reversal learning began on the next scheduled day. Individuals from Group A were trained to respond to the color red, while Group B was trained to respond to the color green during reversal learning. During reversal learning each individual was given a total of 10 trials, 3 days a week, resulting in a total of 30 trials a week. At the beginning of each 10 trial-set, an animal was placed in the start box of the habit reversal aquarium with the opaque dividers in place. The dividers were then removed, revealing the Plexiglas targets. Throughout the series of visual discrimination tasks, one target would be illuminated with the green LED while the other would be illuminated with the red LED. The box was covered with black upholstery velvet so that the animals could not see the experimenter prior to pushing the target and receiving the reward. Once the animal had selected the "correct" color 7 out of 10 trials on a single day of data collection, the color of the positive stimulus was reversed. This procedure was repeated until the color of the positive stimulus had been changed ten times. The total number of trials and the total number of errors per reversal was collected for each individual. The total number of errors for each reversal for all seven animals was tested for normality. While some of the reversals contained a normal distribution of errors, not all the reversals were found to contain a normal distribution of errors. For each reversal the mean was calculated of the total errors made by all animals (mean total of errors). This mean total of errors was used to assess the performance of each group in each 24 reversal. A one-way ANOVA was used to test for a significant difference in the performance between early and later reversals. Results Results are plotted for the seven individuals (Figure 2.2) who completed ten reversals (Reversal 1-10) plus the original color reversal problem (Reversal 0). No difference was seen in the performance between Group A and Group B, therefore the results of these groups were pooled. The mean total errors decreased as reversal number increased. The difference in mean total errors (Figure 2.2) between early and later reversal was highly statistically significant (P=0.0024). We also saw a decrease in the variation of performance between individuals. Discussion Reversal experiments such as these present two problems. Each problem offers unique insights into learning processes and memory formation. The first problem allows an animal to demonstrate an ability to associate a learned response with a food reward. However, with continual reversal of the positive and negative discriminanda, a second problem arises where the animal is allowed to demonstrate an ability to learn and remember that the responses acquired in the first problem must be flexible. The general performance pattern observed across numerous species in learning reversal tasks is as follows. In an initial reversal (e.g. reversal 0), a certain number of errors are made, and then in successive reversals (e.g. 25 reversal 1-3) errors increase to a maximum value. After this peak, errors begin to decrease and eventually plateau. Once this maximum is reached, the rate of decrease in the number of errors with successive reversals differs among species. This rate was quantitatively compared by finding the full width (number of reversals) at half the difference between the initial number of errors and the maximal number of errors of this waveform (FWHM). Figure 2.3 illustrates this methodology. The values for species other than alligators were approximated from published graphs. The FWHM measurement allows comparison of the shape of the performance function and quantitatively represents how quickly each species performance in a learning task improved. In rats (Gatling, 1951) animals returned to the FWHM at reversal 2, pigeons (Reid, 1957) returned at reversal 5, chickens (Bacon et al., 1962) returned at reversal 3. Alligators in the current study returned to the FWHM at reversal 2. Therefore, the performance of juvenile American alligators in a visual discrimination task is similar to that observed in other species including mammals and birds. The overarching objective of this work was to investigate learning abilities in juvenile American alligators reared in the laboratory. Specifically, the goals of this research were one, to establish the optimal conditions for training and reversal using a food reward, two, ascertain the visual discrimination ability of the American alligator and three, determine the degree of complex learning by testing the flexibility of a learned response using a serial reversal approach. Very little is known about the ability of American alligators to participate in these types of tasks. In order to close this gap, fundamental studies are required to establish 26 the baseline performance. We believe that a study of the ability of juvenile lab reared American alligators to utilize an acquired physical response in novel situations, such as those presented in serial reversal experiments, will provide insight into the basal learning in Archosaurs. Reversal problems, such as the one utilized in the current study, are a valuable paradigm for evaluating and comparing the learning ability of different species (Stettner et al., 1967). Furthermore, performance differences of species in these tasks can lend insight into phylogenic differences in behavior and learning ability (Stettner et al., 1967). Birds are the sister group to crocodilians and by comparing the learning ability of these two groups in reversal problems we may gain insight into basal learning in Archosaurs. A variety of bird species, that represent several orders, have all been shown to exhibit a decreasing number of total errors in a series of visual discrimination problems. Chickens (Gallus gallus domesticus), pigeons (Columba livia domestica), crows (Corvus americanus), all showed progressive improvement in the number of errors committed during a series of visual discrimination tasks. However, quail (Colinus virginianus) do not show progressive improvement in the number of errors committed in a serial reversal experiment and therefore do not reflect the pattern seen in other bird species. Alligators show a pattern of decreasing errors in a series of successive visual discrimination problems, the pattern observed in alligators is similar to the pattern seen in many bird species including Corvids. This shared pattern may lend insight into the role of behavioral flexibility and the evolution of learning in basal Archosaurs. 27 Figure 2.1. Arial view of visual reversal apparatus 28 Figure 2.2. Mean total errors per reversal with SEM Visual Reversal at 30!C: Mean Total Errors ± SEM 0 1 2 3 4 5 6 7 8 9 10 0 10 20 30 40 50 60 70 80 90 100 Visual Reversal at 30!C Reversal 29 Figure 2.3. Full width (number of reversals) at half the difference between the initial number of errors and the maximal number of errors Initial Errors Reversal Peak Total Errors Waveform Maximum Full Width of Waveform Half Maximum of Waveform 30 References Bacon, H.R., J.M. Warren, and M.W. Schein. 1962. Non-spatial reversals learning in chickens. Animal Behavior 10:239-243. Behrend, E.R., and M.E. Bitterman. 1967. Further experiments on habit reversal in the fish. Psychonomic Science 8:363-364. Behrend, E.R., V.R. Domesick, and M.E. Bitterman. 1965. Habit reversal in the fish. Journal of Comparative and Physiological Psychology 60:407-411. Bitterman, M.E., J. Wodinsky, and D.K. Candland. 1958. Some comparative psychology. The American Journal of Psychology 71(1):94-110. Bitterman, M.E. 1964. An instrumental technique for the turtle. Journal of Experimental Analysis of Behavior 7(2):189-190. Bitterman, M.E. 1965a. Phyletic differences in learning. American Psychologist 20:396-410. Bitterman, M.E. 1965b. The evolution of intelligence. Scientific American 212:92-100. Brusatte, R.E., M.J. Benton, M. Ruta, and G.T. Lloyd. 2009. Superiority, competition, and opportunism in the evolutionary radiation of dinosaurs. Science 321:1485-1488. Colbert, E.H., R.B. Cowles, and C.M. Bogert. 1946. Temperature tolerances in the American alligator and their bearing on the habits, evolution, and extinction of the dinosaurs. Bulletin of the American Museum of Natural History 86:327-374. Davidson Jr., R.S. 1966. Operant stimulus control applied to maze behavior: heat escape conditioning and discrimination reversal in Alligator mississippiensis. Journal of Experimental Analysis of Behavior 9:671-676. Elsey, R.M., Trosclair III, P.L. and T.C. Glenn. 2008. Nest-site fidelity in American alligators in a Louisiana coastal marsh. Southeastern Naturalist 7(4):737-743. Engelhardt, F., W.T. Woodward, and M.E. Bitterman. 1973. Discrimination reversal in the goldfish as a function of training conditions. Journal of Comparative and Physiological Psychology 85(1):144-150. Eskin, R.M., and M.E. Bitterman. 1961. Partial reinforcement in the turtle. Quarterly Journal of Experimental Psychology 13:112-116. Gatling, F. 1951. The effect of repeated stimulus reversal on learning in the rat. Journal of Comparative Physiology and Psychology 45:347-351. 31 Gossette, R.L., and A. Hombach. 1969. Successive discrimination reversal (SDR) performances of American alligators and American crocodiles on a spatial task. Perceptual and Motor skills 28:63-67. Heath, J.E. 1965. Temperature regulation and diurnal activity of horned lizards. University of California Publications in Zoology 64:97-136. Herman, H., M. Jouvet, and M. Klein. 1962. Analyses polygraphique du sommeil de la tortue. C.R. Academy of Science (Paris) 258:2175-2178. Mackintosh, N.J., and A. Cauty. 1971. Spatial reversal learning in rats, pigeons, and goldfish. Psychonomic Science 22:281-282. Northcutt, R.G., and J.E. Heath. 1971. Performance of caimans in a t-maze. Copeia 1971 (3):557-560. Reid, R.L. 1957. Discrimination-reversal learning in pigeons. Journal of Comparative Physiology and Psychology 51:716-720. Rogers, S.W. 1998. Exploring dinosaur neuropaleobiology: computed tomography scanning and analysis of an Allosaurus fragilis endocast. Neuron 21:673-679. Rogers, S.W. 1999. Allosaurus, crocodiles, and birds: evolutionary clues from spiral computed tomography of an endocast. The Anatomical Record (New Anat.) 257:162-173. Setterington, R.G., and H.E. Bishop. 1967. Habit reversal improvement in the fish. Psychonomic Science 7:41-42. Somero, G.N. 2004. Adaptation of enzymes to temperature: searching for basic "strategies." Comparative Biochemistry and Physiology Part B 139:321-333. Stephens, P.J., P.A. Frascella, and N. Mindrebo. 1983. Effects of ethanol and temperature on a crab motor axon action potential: a possible mechanism for peripheral spike generation. Journal of Experimental Biology 103:289-301. Walker, T.J. 1975. Effects of temperature on rates in poikilotherm nervous systems: evidence from the calling songs of meadow katydids (Orthoptera: Tettignoniidea: Orchelimum) and reanalysis of published data. Journal of Comparative Physiology 101:57-69. Warren, J.M. 1960. Reversal learning by paradise fish (Macropodus opercularis). Journal of Comparative and Physiological Psychology 53(4):376-378. Warren, J.M., K.H. Brookshire, G.G. Ball, and D.V. Reynolds. 1960. Reversal learning by white leghorn chicks. Journal of Comparative Physiology Psychology 53(4):371-375. 32 Williams, J.T. 1967. A test for dominance of cues in the spectacled caiman. Psychonomic Science 8:280. Williams, J.T. 1968. Reversal-learning in the spectacled caiman. The American Journal of Psychology 81(2):258-261. Wodinsky, J., and M.E. Bitterman. 1957. Discrimination-reversal in the fish. American Journal of Psychology 70:569-576 CHAPTER 3 THE EFFECT OF TEMPERATURE ON REVERSAL LEARNING IN JUVENILE AMERICAN ALLIGATORS (ALLIGATOR MISSISSIPPIENSIS) Abstract Ectotherms are sensitive to environmental temperatures. Aspects of poikilotherm metabolism, sensory and motor neuron physiology and the action potentials of muscles have all been shown to be affected by environmental temperature. Recognizing that neuronal physiology is temperature sensitive prompted the question of whether learning and memory are also temperature sensitive. American alligators (Alligator mississippiensis) are capable of performing both visual (Araneo and Farmer, unpubl.) and spatial discrimination tasks (Gossette and Homback, 1969). However, it is not known how temperature affects their ability to learn or perform in such discrimination tasks. The effect of temperature on spatial learning in juvenile American alligators was investigated. Eight naive alligators were divided between two temperature regimes. Group A completed a series of 10 habit reversals at 32°C and then another set at 22°C. 34 Group B completed a series of 10 habit reversals at 22°C and then 10 more at 32°C. Performance was measured using mean total error per reversal. A clear trend of decreasing errors was observed. In addition, alligators made fewer total errors at 22°C than at 32°C regardless of temperature regime. We conclude that alligators perform better in a spatial learning task at 22°C than they do at 32°C. Introduction Studies in learning are aided by development of specialized methods to classify and compare cognitive abilities in several species (Bitterman, 1965). Investigators around the globe have spent time developing techniques and methods to reliably test the adaptability of learning behavior in many species. However, development of methods to test learning in reptiles, especially crocodilian species, has been limited. Large reptiles such as alligators offer the opportunity to study a species capable of adaptation to climatic change throughout millions of years of evolution. Furthermore, crocodilians represent one of two remaining Archosaur lineages. An investigation of its learning ability and specialized brain function may provide clues about basal learning in the Archosaur order. The current research investigated the resilience or susceptibility to temperature changes, of a specialized form of learning in the American alligator. Temperature has been found to affect synaptic transmission, postsynaptic transmission, postsynaptic integration, spike initiation and conduction of neurons 35 (Montgomery and Macdonald, 1990). Murray et al. (2007) found that the phospholipid composition of membranes can be affected by temperature in C.elegans. Many of these neuron changes could be the result of a change in membrane fluidity similar to that found in C.elegans. Furthermore, Somero (2004) investigated the effect of temperature on protein and again found a positive correlation, meaning protein activity increased with temperature. Brain activity is based on neuron activity that has been shown to be affected by temperature (Montgomery and Macdonald, 1990). Therefore, all these temperature sensitive physiological parameters may play a role in learning and memory. Temperature effects various aspects of learning in a variety of species (Roussel et al., 1982). The effects of temperature on the performance ability in rodents is not clear. Even with extensive training sessions, both hypothermia and hyperthermia perturb memory acquisition (Roussel et al., 1982). A severe learning deficit can be produced in mice with even a moderate degree of hypothermia (Essman and Sudak, 1962; Sudak and Essman, 1962; Sudak and Essman 1963). Contrasting results indicate that mice were still able to learn a brightness discrimination task with a rectal temperature of 27 or 20 °C, but latency increased due to impaired motor activity (Boyd and Caul, 1979). Others have found that hypothermic rats, with a core body temperature decrease of 3- 6°C, were still able to learn a memory task when heat was used as a reward (Weiss and Laties, 1961; Panuska and Popovic, 1963). Furthermore, in rats an observed rectal temperature increase of as little as 2 or 3°C can cause amnesia 36 with the severity of symptoms being related to the magnitude of hyperthermia (Misanin et al. 1979). Roussel et al. (1982) observed that when mice were exposed to extreme temperatures (10 °C or 34 °C) prior to learning an avoidance task, these extreme temperatures greatly reduced a subject's learning ability (Roussel et al., 1982). Ectotherms are sensitive to environmental temperatures. Aspects of poikilotherm metabolism, sensory and motor neuron physiology, and the action potentials of muscles are affected by temperature (Krekorian et al., 1968; Zerbolio, 1973; Reeves, 1977; Stephens et al., 1982; French, 1985; Montgomery and Macdonald, 1990; Rome, 1990; Somero, 1995; Gray, 1998; Hosler, 2000; Hosler et al., 2000; Logue et al., 2000; Somero, 2004; Guschina and Harwood, 2006; Guderley, 2004; Bicego, et al., 2007; Murray et al., 2007). Recognizing that neuronal physiology is temperature sensitive prompted the question of whether learning and memory are also temperature sensitive; the implication being learning and memory would be most efficient within a specific temperature range. However, one could make an argument that the ability to create and store new memories and associations regarding one's environment at any temperature could offer a substantial fitness advantage. Only a few studies have been done to examine the effect of temperature on memory and learning in poikilotherms (Krekorian et al., 1968; Riege and Cherkin, 1972; Zerbolio, 1973; Borsook et al., 1977; Roussel et al., 1982; Stephens et al., 1982). In Goldfish, Riege and Cherkin (1972) attempted to assess the effects of temperature on memory and learning. At a given shock 37 level, temperature was found to have an effect on avoidance scores, with avoidance of shock treatments decreasing as temperature decreased (Riege and Cherkin, 1972). However, while memory of punishment was dependent on temperature, it was also found to be dependent on shock level and test interval after shock (Riege and Cherkin, 1972). Zerbolio (1973) conducted a conditioned avoidance response in Goldfish and found that temperature, days of training and the interaction between the two had an effect on avoidance rates. Zerboilio (1973) found a positive relationship between temperature and avoidance rates, in other words avoidance rates increased with increasing temperature. Borsook et al. (1977) attached floats to the ventral surface of fish and observed the rate of compensation with in three temperature regimes to determine the learning ability of fish. From their results, Borsook et al. (1977) concluded that acclimation temperature influences the learning ability of fish (Borsook et al., 1977). A performance index was used to determine how much time a fly spent in a non-preferred temperature compared to a preferred temperature, with more reversals each fly spent less time in the negative reinforcing temperature (Zars and Zars, 2006). The study found that avoidance of the higher temperature, in the 24/30, 24/33 and 24/37 temperature pairings, increased as temperature increased. Furthermore, low temperatures were not shown to reinforce place memory as strongly as high temperatures (Zars and Zars, 2006). Honeybee workers maintain the brood within a narrow temperature range around 34.5 􀀀 1.5 °C. (Jones et al., 2005). Jones et al. (2005) found that short- 38 term memory in Honeybees, Apis mellifera, was affected by temperature of pupation. However, long-term memory was not affected by temperature of pupal development (Jones et al., 2005). Tautz et al. (2003) and Von Frisch (1993) show that bees reared at 36°C preformed normal communication dances, while bees at 32°C preformed fewer dances and the dances that are preformed tend to be shorter than those of bees that pupated at 36°C. This study investigated the effect of temperature on learning and memory and fluctuating asymmetry, and reared bees within the normal temperature with in a hive as well as temperatures that could be experienced along the hive's margins. Short-term memory and learning was significantly affected by rearing temperatures. Rearing temperature was not found to have an effect on long-term memory (Jones et al. 2005). Krekorian et al. (1968) found that maze learning is temperature dependent in the desert iguana, Dipsosaurus dorsalis. Specifically, they found that animals tested at their preferred body temperature outperformed their contemporaries tested at cooler temperatures. Desert iguanas have a preferred body temperature around 42°C (Krekorian et al., 1968). Twenty lizards were randomly assigned to one of four experimental groups. Heat was used as a reward and was supplied through tactile contact of the lizard's body with the floor of the goal box. Experimental groups all had the same difference in body temperature from goal box, but the body temperature and goal box temperatures were different. Learning curves for the two groups are similar in shape, while not statistically significant, the group that experienced the lower temperatures took longer to run the maze. Lizards closest to their preferred temperature learned faster on two 39 separate mazes. Furthermore, Krekorian found that more individuals met the learning criterion at 32°C than 27°C (Krekorian et al., 1968). Many of these studies may be compromised by utilization of a time component, i.e. ,trial length, (Borsook et al., 1977) in their learning criteria, or use of a punishment, such as shock (Reige and Cherkin, 1972) that may be more intense at higher temperatures. Other studies may also be compromised by approaches that can create potential artifacts by the use of temperature as a reward; such gradients can act to differentially reinforce behavior (Krekorian et al., 1968). Where temperatures introduce a stronger reward value, performance values would appear to increase due to the strength of this reinforcement not because memory acquisition is enhanced. In other words, if we accept that animals prefer one temperature to another and that preference is enough of a reward to increase performance values then the temperatures are not actually improving the physiological components of memory. Performance is merely increasing due to the reinforcement (Krekorian et al., 1968). An alternative and more robust approach to address the question of temperature and memory/learning is the use of reversal studies. These offer a useful method to explore temperature and learning for several reasons. One, the performance of individuals is not based on a time component. Two, alligators have been shown to learn effectively without the use of a punishment. Three, reversal studies do not present the opportunity for differential reinforcement because such studies do not rely upon the use of temperature gradients. 40 A common method used to assess the ability of an animal to learn is to train the individual to perform a task and then repeatedly change the task that is required of the animal. Such repeated changes allow one to observe how quickly the animal learns a new behavior. A serial-reversal experiment is one such assay, as it relies on the same basic methodology (Bitterman, 1965b). In a serial-reversal experiment animals are presented with two stimuli that are either visually or spatially distinguishable. One of the two stimuli is assigned to be the "correct" choice and consistently produces a food reward when selected by the animal. Visual problems reward the correct stimulus regardless of position. Spatial problems reward a correct location regardless of what stimuli are present at the location. Throughout the experiment, animals are rewarded for choosing the pre-determined "correct" choice. Once the animal reaches a predetermined criterion of correct choices, the discriminanda are reversed and the negative stimulus is now given a positive sign and rewarded while the previously positive stimulus is now given a negative sign and if selected no longer produces a reward. Experimenters collect data on the total number of errors committed during each reversal. Animals of various taxa, including pigeons, some fish species, turtles, chickens, and rats show progressive improvement and a decrease in the number of errors committed during each reversal (Gatling, 1951; Reid, 1957; Wodinsky and Bitterman, 1957; Bitterman et al., 1958; Warren et al., 1960; Eskin and Bitterman, 1961; Gonzalez et al., 1964; Stearns and Bitterman, 1965; Setterington and Bishop, 1967; Mackintosh and Cauty, 1971). The typical trend is an initial mastery of the problem, a dramatic increase in errors during early 41 reversals and a steady improvement with additional reversals. Errors may increase during early-reversals, as animals tend to persist in selecting the stimuli that previously produced a reward, but as the animal's experience continues, its selection habit becomes more flexible (Bitterman, 1965b). American alligators are capable of performing both visual and spatial discrimination tasks (Gossette and Homback, 1969). However, it is not known how temperature affects their ability to learn or perform in similar discrimination tasks. One possibility is that performance values decrease, indicating animals are less effective learners at certain temperatures, or do the time components used to measure performance values simply increase, indicating that learning is just as effective but simply takes longer. Specifically, learning may not actually be delayed at specific temperatures (if one could directly measure neuronal association), but rather reflects the delayed muscle activity used to assess learning. This study will attempt to answer the first question and lay groundwork for the second. Materials and Methods The activity level of poikilotherms is set by environmental temperature. However, using a food reward means that alligators are motivated by hunger level, which is also influenced by temperature. This level of motivation will also be set by temperature and how quickly they digest. To keep activity levels at an efficient level, the temperature of the home cages was maintained at 30°C. Animals were housed individually in 54-gallon cattle troughs each heated with an 42 aquarium heater; all cattle troughs were housed in a temperature-controlled room. All experiments were conducted in an environmentally controlled room with a temperature range of 26-28°C. The experimental apparatus consisted of a 75-gallon aquarium (measuring 1.22 m, 0.47 m wide and 0.79 m deep) filled with approximately 6.35 cm of water at temperature of 30 ± 1.5°Celsius. A 250-Watt aquarium heater was used to maintain the temperature of the water with in the aquarium. The sides of the aquarium were covered with heavy weight black upholstery velvet, limiting the amount of light within the aquarium. All attempts were made to ensure that the only light available during data collection come from the reversal apparatus (see below). A lid for the aquarium was also constructed from the same black velvet. Two black plastic trays were used as a partisan to divide the tank into two regions a start box and a goal box. The goal box was further divided into two alleyways by a quarter in thick pieces of opaque acrylic sheeting. Each alleyway led to a single Plexiglas target. Each target was constructed from a single half inch thick piece of frosted Plexiglas cut into 11.43 cm by 11.43 cm square. A Radio Shack technology Plus, 5mm High Brightness Full-color LED was then installed in the piece of Plexiglas and silicone was used to keep it in place. A Full-color LED is manufactured to produce three colors, blue (470 nm), red (624 nm) and green (525 nm). This made a convenient light source because a single LED could be used during both training and data collection phases of the experiment. Each target was then equipped with a Guard N.O./N.C. magnetic reed switch. As the target was 43 depressed by an animal, the switch would come in close proximity to the magnet closing a circuit activating an automatic feeding apparatus. The LED and magnetic switch of each target were wired to a circuit board. An Arduino processing chip installed on the circuit board allowed the LED and magnetic switch to communicate with a computer. The computer monitored and recorded what magnetic switch was activated, left or right, and what color the LED projected when the switch was activated. During reversals, the LEDs were illuminated with green and the computer randomly selected which target, left or right, would produce a food reward. At the end of the first set of reversal a Chi-squared was run on the computer selection to ensure that the computer program was indeed randomly selecting both the left and right targets with the same frequency. Attempts were made to distribute a food reward using a Fish Mate P21 automatic pond fish feeder that had been modified with a small rotary motor. In previous reversal studies we have found that when the motor was activated, by a correct selection, alligators would walk to the back of the tank to the area where the food reward was delivered. Therefore, even though the fish feeder is not used to distribute a reward the sound of the motor appears to facilitate learning. Training was divided into three progressive steps, where each step involves a more difficult task, culminating in reversal learning and data collection. Each training step was fully automated by a computer program; the experimenter simply had to select which training protocol was appropriate. 44 The goal of the first step of training was first to introduce study animals to the experiment's routine and second to train them to come to expect food only when they are in the glass habit reversal aquarium and nowhere else. On training days, all individuals were transferred to a holding container. The holding container was a Rubbermaid container filled with water held at a temperature of 30°C. The holding tank is important because it allows individuals to be kept warm while waiting for other individuals to finish their trails. During this first step of training a single individual was selected at random, removed from the holding tank and placed in the glass habit reversal aquarium. During step one, each Plexiglas target was illuminated with the color blue, a neutral stimulus. This individual was then given two pellets at a time until 10 pellets were consumed or an hour had past, whichever came first. Individuals were only fed what they consumed in the glass aquarium. This was repeated 3 days a week until the alligator routinely consumed 10 pellets within an hour. Once this criterion was met, the second step of training was begun on the next scheduled day of training. In step two, each target was illuminated with the neutral stimulus. However, now each target was baited with a single pellet of food. Each alligator was given a half-hour to consume 10 pellets. The experimenter was responsible for checking on the alligator every 5 minutes, if a pellet had been consumed, the experimenter would then reset the trial by encouraging the alligator to walk back to the starting box, the dividers were put in place, the target was baited with an additional pellet of food, the experimenter would then replace the lid and pull the 45 dividers. This continued until the individual had met the criterion of consuming ten pellets in a half-hour. Once that criterion was met, the third step of training was begun on the next scheduled day of training. The goal for the third step of training was to train the alligators to actually push the Plexiglas targets. When actively searching for food, juvenile alligators move their rostrum from side to side. If food is not easily found this movement becomes more aggressive as the individual appears to become frustrated. By training them to come to expect food to be present in front of the targets we have been able to capture this exploratory behavior and utilize it for target training. Just as before both targets were illuminated with the neutral stimulus. Once in the tank, individuals immediately approached the targets. As they discovered food is not easily found the alligators began to actively search for food. In previous studies, this searching has become forceful enough to displace the Plexiglas target at which time an experimenter rewarded the individual. A single press of a target defined a trial. After each trial the experimenter reset the apparatus just as in step two. In previous experiments the alligator began to associate a food reward with the action of pushing the target after only a few repetitions. During step two both targets were baited in an attempt to train the alligators to come to expect food from both targets. In step three, both targets were rewarded for the same reason. Alligators showed very little spatial preference and actively explored and pressed the left and right target with equal frequency. Once the criterion of completing 10 trials in a half-hour was met, data collection began during the next scheduled day of training. 46 Once an individual had completed each step to criterion, data collection was begun. During reversal learning each individual was given a total of 10 trials per day, 3 days per week resulting in a total of 30 trials in a 7-day interval. Group A completed a series of ten habit reversals at 32°C and then another set at 22°C. Group B completed a series of 10 habit reversals at 22°C and then 10 more at 32°C. Group A was trained with the right target as the positive discriminanda while Group B was trained to respond positively to the left target. For each consecutive reversal, individuals were reversed, on the next scheduled day, once an individual achieved a performance of 7 correct trials out of 10 trials. This was repeated for a total of 10 reversals. The temperature régimes were then switched and a second set of 10 reversals was completed. Previous studies have found individual variation in the performance values of animals in a learning task, with some individuals excelling while others were never able to meet criteria even on an initial reversal (Araneo and Farmer, Fig. 2.2, unpubl.). Therefore, a comparison between groups may not be effective. Instead it may be more effective to compare the performance during an initial set of reversals to the performance of those same individuals in an additional set of reversals after a new temperature regime is established. For the group of animals completing reversals at 22°C an additional procedure was utilized to ensure that the animals would be completing reversals at 22°C. Prior to the completion of the third step of the training protocol, the smallest and the largest animals were placed in a holding tank containing water 47 at a temperature of 22°C. The cloacal temperature of these individuals was then taken every 10 minutes. The smallest individual reached a temperature of 22°C after 20 minutes, while the largest individual reached a temperature of 22°C after 45 minutes. It was determined that it would take an average of 30 minutes for the majority of the animals to reach a temperature of 22°C. Every animal was then placed in the holding tank one at a time for 30 minutes to ensure enough time for their cloacal temperature to reach 22°C. Therefore, on each day of data collection, were an animal would be completing trials at 22°C, the animal was removed from its home cage and placed in the holding tank for 30 minutes prior to data collection. The experiment began by establishing a baseline ability of alligators in a spatial problem. With these baselines established, we repeated the experiment with the same individuals completing an additional set of reversals in a second temperature regime. Such an experimental setup allowed for a comparison between the results of the two temperature experiments. Performance was measured using mean total error per reversal. A comparison was achieved by evaluating the mean total errors per reversal at the two different temperature regimes. However, two additional techniques were also utilized as a way to gauge temperature effects on learning and memory. One, a linear regression was utilized in order to shine light on the relationship between days spent in each reversal and the number of successful trials. Two, logistical regressions were used to look at the relationship between covariant values. This spatial experiment 48 has a number of variants. For analysis of these variants we broke the data into "yes" and "no". In this way we were able to look at the effect (if any) the variant has on the outcome individually. By utilizing logistical regressions we were able to look at the following variants: days spent in each individual reversal, reversal number and temperature. Logistical regressions produce an estimate coefficient or coefficient value. Furthermore, each coefficient value can have a positive or negative sign. This value is a measure of the relationship between the variant and the outcome. A higher number on the graph represents a higher correlation between the variant and the outcome, in our cases successful trials. Results American alligators can learn a physical task when training is associated with a positive reward. Statistical values, of number of successful trials, taken over all alligators are significant and show an improvement in the number of successes achieved by each alligator as days within any given reversal increases (Figure 3.1, p=0.0003). As the days progress within a reversal, animals performed more successfully and therefore made fewer errors. A linear regression of number of successes per days in any reversal standardizes the day of each reversal. The grey line includes all trials. The black line only takes into account those days were the animals completed all ten trials. The darkness of the circles represent the number of times that value occurred, with grey circles the value occurred less often and black circles indicating the value occurred far more frequently. When one looks at the relationship between days spent in each 49 reversal and the number of success a clear trend of decreasing errors per day in reversal is observed (Figure 3.1). Temperature was also found to have an effect on the spatial learning ability of juvenile American alligators. There was a statistical significance between the number of successful trials at the lower temperature (22°C), p=0.0282, when compared to the number of success at the higher temperature (32°C) when the performance of every animal across every reversal is taken into account. In addition, alligators made fewer total errors at 22°C than at 32°C regardless of when they experienced each temperature (Figure 3.2). By utilizing Logistical regressions we looked at the following variants: days in each individual reversal (Figure 3.3), reversal number (Figure 3.4) and temperature (Figure 3.5). Logistical regressions produce an estimate coefficient or coefficient value. Additionally, as the number of days within each reversal increased all alligators made fewer errors (Figure 3.3), based on the observation that none of the coefficient values are negative. The coefficient value of alligators numbered 22, 11 and 24 is very small, indicating these alligators showed little improvement. The greatest improvement, and highest coefficient value, was seen in the alligators numbered 12, 1 and 14, while alligators numbered 5 and 6 represent intermediate coefficient values. After running a Logistical regression for the reversal number variant, we see a positive relationship between reversal number and number of successful trials for alligators numbered 11, 12, 1, 24 and 14. A negative coefficient value is seen for alligators numbered 22, 5 and 6. However, 50 none of these negative values are as large as the positive coefficient values of alligators numbered 11, 12 and 14. Finally, a logistical regression of the temperature variant was also run. Here 32°C was assigned a positive value while 22°C was assigned a negative value. Alligators numbered 22, 5,1, 24 and 14 all show a strong relationship between the lower temperature of 22°C and the number of successful trials. Three of the individuals (alligators numbered 11, 12 and 6) show a relationship between 32°C and the number of successful trials. However, overall a significant effect was observed towards 22°C. We conclude that alligators perform better in a spatial learning task at 22°C than they do at 32°C, however the performance of individuals is highly variable. Discussion The effect of temperature on spatial learning in juvenile American alligators was investigated. Temperature affects a variety of neuronal processes both in endotherms and ectotherms. However, whether temperature is a significant environmental modifier of the ability of crocodilians to master these problems was unknown and not well tested. The results of our research indicate that American alligators perform better in a spatial discrimination task at 22 °C than they do at 32 °C. A subject's ability to discriminate between two stimuli based upon positional or spatial cues is a spatial experiment. When the spatial cues are presented in a series of reversals, it is called a spatial reversal experiment and our approach tasked Alligators to perform two series of ten 51 reversals each. The reversals embody several learning problems, the first learning to associate a specific physical task with a reward and the second learning an association between a specific location and a reward. Finally, once the stimulus signs are reversed an animal must master an additional problem of learning and demonstrating successful flexible responses - to different stimuli. Aspects of ectotherm physiology are particularly sensitive to environmental temperature, including the cognitive abilities of ectotherms (Krekorian et al., 1968; Zerbolio, 1973; Borsook et al., 1977; Zars and Zars, 2006). The relationship between temperature and cognitive performance has been investigated in the laboratory with the common prediction that the higher the temperature, the better the performance. Similarly, we predicted the performance of juvenile American alligators in a spatial learning task would be enhanced at higher temperatures due to the greater activity of proteins and overall increased metabolic rate observed in poikilotherms at higher temperatures. Such results would resemble those seen in the desert iguana, were we would anticipate the best results for alligators to occur at 32°C, the temperature closest to the preferred body temperature of American alligators (Krekorian et al., 1968; Farmer, 2008). Our results showed enhanced performance occurred at the cooler experimental temperature of 22°C. French (1985) saw that the performance of fish was enhanced when animals were tested at a temperature lower than the temperature experienced during memory formation. In our experiment we saw that overall both Group A and B made fewer errors at 22°C. If our animals 52 followed the pattern seen by French, only the group originally trained at 32°C would have preformed better at 22°C. Therefore, the temperature our animals experienced during memory acquisition did not affect memory utilization because we saw that individuals from both temperature treatments made fewer errors at 22°C. Additionally, in fish individuals reach an upper limit where higher temperatures are found to be detrimental to learning. Therefore we can say that memory formation and utilization are modulated by temperature and that temperature extremes can perturb functional memory and learning processes. Alligators prefer a temperature of 30°C after feeding (Farmer et al., 2008). Therefore, 32°C may represent an upper limit of functional memory formation accounting for the increased number of errors at higher temperatures. However, due to the temperature extremes these organisms endure in their natural habitat, we did not expect 32°C to be detrimental to learning. Thirty-two degrees C is a temperature that these animals potentially experience on a daily basis. Furthermore, the thermal maximum of this species is 38°C. Therefore, it is unlikely that 32°C is going to be detrimental to learning. A far more likely, but unconventional, explanation for a better performance at a cooler temperature surrounds the alligator's stress and hunger level. Because it is unlikely the animals in the current experiment reached the same satiation point that Farmer's animals reached, it is possible that the animal's hunger level was responsible for setting up behavioral patterns that account for the increased performance at the cooler temperature. Based on the laboratory results seen by Farmer et al. (2008), we can predict a similar pattern in wild populations. In the 53 wild, animals at 30°C, whose hunger is not completely satisfied, may seek out cooler temperatures to reduce the stress of hunger and the temperature-induced metabolic ramp up. We suggest the housing and animal care arrangement was potentially stressing our animals. In other words, because our experiment relied on motivating the animals with food, we assume our animals always experienced a certain level of hunger. Therefore, when the animal was placed in the reversal arena at the warmer temperature the animal's stress level increased, causing the animal to make a greater number of errors. However, when the animal was placed in the reversal apparatus at the cooler temperature, the temperature corresponding to the preferred body temperature for our animal's hunger level, the animal's stress level decreased and they were able to make fewer errors. While American alligators perform better in a spatial discrimination task at 22 °C than they do at 32 °C, the performance of individuals is highly variable. The variables responsible for this individual variation were explored using logistical regressions. From these regressions it is clear that some animals seemed to be more sensitive to one variable, such as reversal number, while other individuals appeared to be sensitive to another variable such as trial number. In other words, in every individual a relationship between number of success and at least one of the variables we considered was observed. However, the same relationship was not seen across all alligators. For example, alligator 14 showed a strong relationship between days in reversal, reversal number and 22°C. Alligator 22 showed the strongest relationship between reversal number and 22 °C, but not for days in reversal. Therefore, our logistical 54 regressions demonstrate the individual variation in performance. Additional variables such as incubation temperature and individual responses to hunger stress could also help to explain the variation in performance values we observed. In honeybees, it is seen that temperature of pupation differentially effects short-term and long-term memory. Rearing temperature has been shown to affect a bee's ability to perform communication dances. Specifically, bees reared at a low temperature perform fewer dances than bees reared at higher temperatures (Jones et al., 2005). Therefore, just as in honeybees, it is possible that the temperature of incubation may be the source of this variation. Incubation temperature is known to affect sex, body size, energy reserves and metabolic rate in American alligators (Allsteadt and Lang, 1995; Western et al., 2000). Incubation temperature may also be affecting neuronal processes that are responsible for learning and memory. Therefore, it is possible that incubation temperature could account for the individual variation we observed in the spatial discrimination task. An alternative explanation that may account for the individual variation we saw in our covariant analysis surrounds an individual alligator's response to hunger stress. Our covariant analysis explored a number of variables that help to quantify an individual's performance in a learning task, in other words each covariant can be thought of as a measure of performance. The stress of being hungry could have had a differential effect on each one of these measurements of performance. For example, the stress of being hungry may affect the relationship between days in reversal and success for one animal and 55 the stress of being hungry may affect the relationship between reversal number and success for another animal. American alligators are a useful study organism for cognitive studies because they share certain life history characteristics with birds and mammals. For example all three groups exhibit parental care, social groups and predatory behavior. Furthermore, certain ecological factors such as food type could affect the evolution of cognitive abilities in a similar manner in both ectotherms and endotherms. However, certain ecological variables, like temperature and climate, may have differential effects on neuronal processes in ectotherms and endotherms. The current study explored the effect of environmentally relevant temperatures on the spatial discrimination ability of American alligators. These environmental temperatures are temperatures these animals could potentially experience on a daily basis. However, in this study of this ectothermic species we saw a significant difference in the number of errors committed at the cooler temperature. Therefore, these results indicate that the flexible learning abilities in ectotherms may be influenced by environmental pressure not seen in endotherms. Furthermore, such pressure may imply these abilities are under a different evolutionary pressure as well. 56 Figure 3.1. Linear regression of spatial temperature data showing a decrease in the number of errors committed by an animal as days within a reversal increased, grey line includes all days regardless of whether all animals completed ten trials, black line includes only those days were all animals completed all ten trials. 57 Figure 3.2. Mean total errors per reversal, first set of reversal was run at one temperature, then for the second set of 10 reversal the temperature treatment had been switched, a clear trend of decreasing errors per days spent in each reversal was observed, alligators made fewer total errors at 22°C than at 32°C regardless of temperature regime (p=0.0282). 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 Group D at 22!C then 32!C Group C at 32!C then 22!C Visual Reversal at 32!C: Mean Total Error± SEM Reversal 58 Figure 3.3. Logistic regression analyzing the relationship between days in reversal and successful trials 59 Figure 3.4. Logistic regression analyzing the relationship between reversal number and successful trials 60 Figure 3.5. Logistic regression analyzing the relationship between temperature and successful trials 61 References Allsteadt J. and J.W. Lang, 1995. Incubation temperature affects body size and energy reserves of hatchling american alligators (Alligator mississippiensis). Physiological Zoology 68(1):76-97. Bitterman, M.E. 1964. An instrumental technique for the turtle. Journal of Experimental Analysis of Behavior 7(2):189-190. Bitterman, M.E. 1965a. Phyletic differences in learning. American Psychologist 20:396-410. Bitterman, M.E. 1965b. The evolution of intelligence. Scientific American 212:92-100. Bitterman, M.E., J. Wodinsky, and D.K. Candland. 1958. Some comparative psychology. The American Journal of Psychology 71(1):94-110. Borsook, D., C.J. Woolf, and A.D. Vellet. 1977. Temperature acclimation and learning in fish. Experientia 34(1):70-71. Boyd S. C. and W.F Caul. 1979. Evidence of state dependent learning of brightness discrimination in hypothermic mice. Physiology and Behavior 23:147- 153. Broughton, R. 1972. Phylogenetic evolution of sleep studies, pp. 2-7. In: The Sleeping Brain, M. Chase M. (ed.). Brain Information Service, Los Angeles. Crosby, W.C. 1917. The forebrain of Alligator mississippiensis. Journal of Comparative Neurology 27:325-402. Eskin, R.M., and M.E. Bitterman. 1961. Partial reinforcement in the turtle. Quarterly Journal of Experimental Psychology 13:112-116. Essman, W.B. and F.N. Sudak, 1962. Effect of body temperature reduction on response acquisition in mice. Journal of applied Physiology. 17:113-116. Flannigan, W.F., R.H. Wilcox, and A. Rechtschaffen. 1972. The EEG and behavioral continuum of the crocodilian, Caiman sclerops. Electroencephalography and Clinical Neurophysiology 34:521-538. French, A.S. 1985. The effects of temperature on action potential encoding in the cockroach tactile spine. Journal of Comparative Physiology A 156:817-821. Gatling, F. 1951. The effect of repeated stimulus reversal on learning in the rat. Journal of Comparative Physiology and Psychology 45:347-351. Gonzalez, R.C., Roberts, W.A. and M.E. Bitterman 1964. Learning in adult rats with extensive cortical lesions made in infancy. American Journal of Psychology 77:547-562. 62 Gossette, R.L., and A. Hombach. 1969. Successive discrimination reversal (SDR) performances of American alligators and American crocodiles on a spatial task. Perceptual and Motor Skills 28:63-67. Guderley, H. 2004. Locomotor performance and muscle metabolic capacities: impact of temperature and energetic status. Comparative Biochemistry and Physiology, Part B 139:371-382. Guschina, I.A., and J.L. Harwood. 2006. Mechanisms of temperature adaptation in poikilotherms. FEBS letters 580:5477-5483. Jones, J.C., Helliwell, P., Beekman, M., Maleska, R., and B.P. Oldroyd. 2005. The effect of rearing temperature amd memory in the honey bee, Apis mellifera. Journal of Comparative Physiology(A):191:1121-1129. Krekorian, C.O., V.J. Vance, and A.M. Richardson. 1968. Temperature-dependent maze learning in the desert iguana, Dipsosaurus dorsalis. Animal Behavior 16:429-436. Mackintosh, N..J. and A. Cauty, 1971. Spatial reversal learning in rats, pigeons, and goldfish. Psychonomic Science 22:281-282. Meglasson, M.D., and S.E. Huggins. 1978. Sleep in a crocodilian Caiman sclerops. Comparative Biochemistry and Physiology 63(A):561-567. Montgomery, J.C., and J.A. Macdonald. 1990. Effects of temperature on nervous system: implications for behavioral performance. American Journal of Physiology Regulatory Integrative Comparative Physiology 259:191-196. Murray, P., S.A.L. Hayward, G. Govan, A.Y. Gracey, and A.R. Cossins. 2007. An explicit test of the phospholid saturation hypothesis of acquired cold tolerance in Caenorhabditis elegans. Proceedings of the National Academy of Science 104(13):5489-5494. Northcutt, R.G., and J.E. Heath. 1973. T-maze behavior of the Tuatara (Sphenodon punctatus). Copeia 1973(3):617-620. Panuska J.A. and V.P. Popovic, 1963. The critical temperature for learning in hypothermic rats. Preceedings of the International Congress of Zoology, 16th. Washington D.C. Platel, R., H.J.A. Beckers, and R. Nieuwenhuys. 1973. Le champs corticaux chez Testudo hermanni (Reptile Chelonien) et chez Caiman crocodylus (Crocodilian). Acta morph. Neerl-scand. 11:121-150. Rechtschaffen, A., M. Bassan, and S. Ledeccky-Janecek. 1968. Activity patterns in Caimen sclerops (crocodilia). Psychophysiology 5:201. Riege, W.H., and A. Cherkin. 1972. One-trial learning in goldfish: temperature dependence. Behavioral Biology 7 (1-44R):255-263. 63 Reid, R.L. 1957. Discrimination-reversal learning in pigeons. Journal of Comparative Physiology and Psychology 51:716-720. Rome, L.C. Influence of temperature on muscle recruitment and muscle function in vivo. American Journal of Physiology, Regulatory Integrative and Comparative Physiology 259:210-222. Roussel, B., P. Turrillot, and K. Kitahama. 1982. Effect of ambient temperature on learning. Journal of Physiology and Behavior 28:991-993. Somero, G.N. 2004. Adaptation of enzymes to temperature:searching for basic "strategies". Comparative Biochemistry and Physiology, Part B 139:321-333. Stephens, P.J., P.A. Frascella, and N. Mindrebo. 1983. Effects of ethanol and temperature on a crab motor axon action potential: a possible mechanism for peripheral spike generation. Journal of Experimental Biology 103:289-301. Setterington R.G. and H.E. Bishop, 1967. A comparison of key-pecking with an ingestive technique for the study of discriminative learning in pigeons. American Journal of Psychology 78:48-56. Sudak W.B. and F.N. Essman, 1962. Effect of body temperature reduction on response acquisition in mice. Journal of Applied Physiology 17:113-116. Sudak W.B. and F.N. Essman 1963. Effect of hypothermia on the establishment of a conditioned avoidance response in mice. Journal of Comparative and Phsiological Psychology 56(2):366-369. Vasilescu, E. 1970. Sleep and wakefulness in the tortoise. Review Roumain de Biologie Animals 15:177-179. Walker, T.J. 1975. Effects of temperature on rates in poikilotherm nervous systems: evidence from the calling songs of meadow katydids (Orthoptera: Tettignoniidea: Orchelimum) and reanalysis of published data. Journal of Comparative Physiology 101:57-69. Warner, B.F. and S.E. Huggins. 1978. An electroencephalographic study of sleep in young caimans in a colony. Comparative Biochemistry and Physiology 59 A:139-144. Warren, J.M. 1960. Reversal learning by paradise fish (Macropodus opercularis). Journal of Comparative and Physiological Psychology 53(4):376-378. Warren, J.M., K.H. Brookshire, G.G. Ball, and D.V. Reynolds. 1960. Reversal learning by white leghorn chicks. Journal of Comparative Physiology Psychology 53(4):371-375. Western P.S., Harry J.L., Marshall Graves, J.A. and A.H. Sinclair. 2000. Temperature-dependent sex determination in the American alligator: expression of SF1, WT1 and DAX1 during gonadogenesis. Gene241(2):223-32. 64 Williams, J.T. 1967. A test for dominance of cues in the spectacled caiman. Psychon. Science 8:280. Wodinsky J. and M.E. Bitterman, 1957. Discrimination-reversal in the fish. The American Journal of Psychology 70(4):569-576. Zars M. and T. Zars, 2006. High and low temperature have unequal reinforcing properties in Drosophila spatial learning. Journal of Comparative Physiology A 192:727-735. Zerbolio, D.J., 1973. Temperature-dependent learning in goldfish: a multi-trial active avoidance situation. Behavioral Biology 8:755-761. CHAPTER 4 EFFECT OF DDE ON LEARNING IN AMERICAN ALLIGATORS Abstract The current study explores the effects of in ovo DDE (1,1-dichloro-2, 2- bis[p-chlorophenyl] ethylene) exposure on learning and behavior of hatchling American alligators (Alligator mississippiensis). Specifically, we investigated if organochlorines, such as DDE, affect cognition of American alligators by comparing the performance of individuals exposed as embryos to an organochlorine to the performance of control individuals in a spatial learning task. Both field and laboratory evidence show that alligator populations are susceptible to the contaminant DDE. However, little is known about the effects DDE has on the behavior and learning ability of these organisms. Learning and memory play a crucial role in a variety of essential behaviors such as territory establishment and parental care. In order to address this question, a modified T-maze was used to elucidate the effect of DDE exposure on the learning and memory ability of this species. Hatchling American alligators, exposed to organochlorines, through maternal transfer, made substantially more errors in a learning task than control individuals who had not been exposed to organchlorines. We conclude that DDE affects the acquisition of a learning task. 66 However, it does not seem to affect the utilization of acquired memory. Because learning and memory are essential elements to these behaviors, if indeed perturbed by organochlorines, exposed populations could suffer a substantial fitness cost detrimentally affecting the dynamics of exposed populations. Introduction The worldwide use of the synthetic pesticide DDT (1,1,1-trichloro-2,2,-bis p-chlorophenyl ethane) has exposed both humans and wildlife to this pollutant and it's breakdown products (Kleinow et al., 1987). A variety of morphological, developmental, and physiological abnormalities in both humans and wildlife have been linked to DDT exposure (Schantz and Widholm, 2001). Furthermore, DDT continues to be used to combat malaria in sub-Saharan Africa and other warm, humid regions of the world, and thus many wildlife populations continue to be exposed to DDT (Wu et al., 2000). Numerous adverse effects and behavioral changes have been observed in birds, the sister group to crocodilians exposed to organochlorine compounds. For example, environmental DDT exposure alters the size of the forebrain and changes the song control system and nuclei necessary for song production and normal sexual behavior in the American robin (Turdus migratorius) (Iwaniuk et al., 2006). DDT is lipoophilic and has been linked to deaths in Western grebes (Aechmophorus occidentalis) due to accumulation in fatty intracerebral tissue (Dolphin, 1959; Hunt and Bischoff, 1960). DDT and its breakdown products also have estrogenic activity in birds as well and can cause alterations to the sexual characteristics, similar to those seen in alligators (McLachlan, 1993; Kelce et al., 67 1995; Jobling et al., 1996). Sex organ development, behavior and fertility are affected by exposure to estrogenic chemicals. For example, male sea gulls living in contaminated ecosystems ignore nesting colonies (Hunt and Hunt, 1977). Furthermore, female sea gulls may pair and nest together (Luoma, 1992). Additionally, when gull eggs are treated with DDT male female sex reversals are produced (Fry and Toone, 1981). Furthermore, abnormal secondary sex characteristics are seen in roosters (Gallus gallus) exposed to DDT (Burlington and Lindeman, 1950). Even though it is known that these compounds can derail normal cerebral function in birds and other species, little is known about the effects of organochlorines on the crocodilian brain. American alligators (Alligator mississippiensis) exposed to organochlorine contaminants, including DDE, show multiple reproductive abnormalities, low clutch viability, reduced phallus size, and altered plasma hormone concentrations (Woodward et al. 1993; Guillette et al., 1994, 1996b, 1997, 1999b, 2000; Crain et al., 1998; Pickford et al., 2000). The hepatotoxic and other effects of high levels of exposures to organochlorines are relatively well understood but the effects of lower doses, while not immediately lethal, may impact species through effects on behavior and cognition. Although organochlorines affect behavior and learning in children (Keifer and Mahurin, 1997; Eskenazi, 2006) and other animals (Eriksson et al., 1990), a literature search was unable to turn up any studies that have specifically assessed the effects of these molecules on learning and behavior in crocodilians, where low doses known to exist today may critically impact survival and complex social behaviors. Critical behaviors that may be adversely affected 68 by the persistence of these contaminants are breeding, nurturing and rearing. Crocodilians exhibit extensive parental care (Hunt 1975; Pooley 1977). Female parental care starts with nest building and nest defense throughout the incubation period (Mcilhenny 1935; Joanen, 1969; Joanen and McNease, 1972; Hunt and Watanabe, 1982). However, parental behavior extends beyond incubation. Crocodilian mothers have been observed scraping open nests to retrieve new hatchlings as well as using their teeth to gently open any un-hatched eggs (Watanabe, 1980). Additionally, females will continue to defend their pods, groups of hatchlings, through the summer and into the following spring (Deitz, 1979). In addition to parental care behavior, crocodilians display a repertoire of vocal cues starting as hatchlings and continuing throughout the life of the adult (Modha, 1967; Garrick and Lang, 1977; Garrick et al., 1978). The early vocal cues are thought to be critically important to the mother-offspring interaction. After placing a hatchling at a roadside, a researcher observed a mother carrying a vocalizing hatchling back to the nest from which it had been removed (Kushlan, 1973). Pods will remain together for the first year (Woodward et al., 1987) and individuals within the pods will vocalize to each other at the approach of an intruder. Crocodilians display a repertoire of behavioral cues as adults as well (Modha, 1967; Garrick and Lang, 1977; Garrick et al., 1978). Both vocal and behavioral cues are used during sexual competition, territory establishment, mate selection and copulation (Garrick and Lang, 1977). Both male and female crocodilians establish and defend territories (Garrick and Lang, 1977). Furthermore, female alligators show nest fidelity from year to year, 69 returning to the same nest site at the beginning of each reproductive cycle (Elsey et al., 2008). We investigated the effects of in ovo DDE (1,1-dichloro-2, 2-bis[p-chlorophenyl] ethylene) exposure on aspects of learning and behavior of hatchling American alligators. Specifically, cognition of American alligators was evaluated by comparing the performance of exposed individuals to the performance of control individuals in the spatial learning task. Understanding the effects of organochlorine exposure on crocodilian cognition and behavior is important in a number of ways. Organochlorines, such as DDT, can bioaccumulate and are therefore particularly problematic for top predators. Furthermore, many crocodilians live in regions of the world where DDT continues to be used to combat malaria or in areas where the breakdown products of DDT, DDE and DDD (1,1-Bis(p-chlorophenyl)-2,2-dichloroethane), remain in the ecosystem. A common method used to assess the ability of an animal to learn is to train an animal to perform a task and then repeatedly change the task that is required of the animal. Such repeated changes allow one to observe how quickly the animal learns a new behavior. A habit-reversal experiment is one such assay as it relies on the same basic methodology (Bitterman, 1965). In a habit-reversal experiment animals are presented with two stimuli that are either visually or spatially distinguishable. One of the two stimuli is assigned to be the "correct" choice and consistently produces a reward when selected by the animal. Visual problems reward the correct stimulus regardless of position. Spatial problems 70 reward a correct location regardless of what stimuli are present at the location. Throughout the experiment animals are rewarded for choosing the pre-determined "correct" choice. Once the animal reaches a predetermined criterion of correct choices, the discriminanda are reversed and the negative stimulus is now given a positive sign and rewarded while the previously positive stimulus is now given a negative sign and if selected no longer produces a reward. Experimenters collect data on the total number of errors committed during each reversal. Animals of various taxa, including pigeons, some fish species, turtles, chickens and rats show progressive improvement and a decrease in the number of errors committed during each reversal (Gatling, 1951; Bitterman et al., 1958; Reid, 1958; Warren, 1960; Warren et al., 1960; Eskin and Bitterman, 1961; Gonzalez et al., 1964; Stearns and Bitterman, 1965). Very little is known about the ability of crocodilians to participate in these types of tasks. Previous experiments have explored learning abilities in juvenile American alligators as measured by performance in a series of visual and spatial learning experiments. A modified T-maze was used to determine the ability of the spectacled caiman (Caiman crocodiles) to reverse a habit formed in a spatial problem (Williams, 1967). Caimans require fewer trials to learn a new arm of a T-maze after each reversal. The spatial learning ability of Alligator mississippiensis compared to that of Crocodylus acutus has also been studied (Gossette and Hombach, 1969). Both species were asked to discriminate between two stimuli based on location and both species showed progressive improvement. American alligators are capable of performing both visual (Araneo and Farmer, unpubl., Fig 71 2.2) and spatial discrimination tasks (Gossette and Homback, 1969). Testing these hypotheses adds an ecologically important interdisciplinary dimension. Furthermore, these studies may provide insights into the cause of changes to population dynamics in areas of the world where crocodilians are exposed to organochlorine contaminants. This knowledge may prove important for making fully informed decisions about the risks to ecosystems versus the benefit to humans in the continued use of DDT. Materials and Methods Alligator eggs were collected from Central Florida in June of 2010 in collaboration with Dr. Guillette. Eggs were collected from Lake Apopka and the cleaner Lake Woodruff within 2 weeks of oviposition. All eggs were incubated in the Guillette laboratory at the University of Florida, Gainesville. Incubation took place in an environmentally controlled room were the temperature and humidity were monitored daily. Eggs were incubated at 100% humidity and 32°C, a temperature that produces both males and females (Milnes et al., 2005). On July 4th, 2010 eggs were transported from The University of Florida, Gainesville to the laboratory facilities at The University of Utah. Temperature was monitored throughout transportation and was never allowed to exceed 33 °C or drop below 26°C. Once the eggs arrived safely at the University of Utah, they were placed in an incubator where temperature and humidity were monitored daily. At the University of Utah eggs were incubated at a temperature of 32 °C and 100% humidity. 72 The experiment contained three treatment groups: (1) group 1, the treatment group made up of Lake Woodruff animals exposed to a topical application of the contaminate DDE in ovo; (2) group 2, containing control animals from Lake Woodruff; and (3) group 3, a group of animals from the contaminated Lake Apopka. On the first day of stage 23, a single topical pesticide treatment was administered to the eggshells of the in ovo treatment group (1) of Lake Woodruff animals (methodology of Crain et al., Milnes et al. and Spiteri et al.). A 0.1􀀀g DDE/g egg mass (ChemServ. West Chester, PA) treatment dosage was applied to the in ovo group. Such concentrations are shown to affect the differentiation of the gonad (Milnes et al., 2005; Matter et al., 1998; Crain 1997). A stock solution of 1 mg/ml DDE was made up by dissolving 5 mg powered DDE in 50 ml of 95% ethanol. This liquid treatment was then applied to each egg based on egg mass. This procedure ensured that all eggs were exposed to a standardized number of 􀀀g of DDE based on egg mass. On the day of hatching, experimental animals were weighed and web tags were attached to the left back foot of each hatchling for identification purposes. Between August 16, 2010 and September 29, 2010, a total of 71 alligators hatched, 21 individuals from embryonic group 1, 19 from embryonic group 2; and 32 from embryonic group 3. Hatchling alligators were trained to perform a spatial learning task utilizing a modified T-maze in which a free space is built into the T-maze (Figure 4.1). Alligators and Caimans can be trained to perform a spatial learning task reliably when return to a home cage is used as a reward (Davidson, 1966; Northcutt and 73 Heath, 1971). The modified T-maze was utilized for this learning task in order to prevent the alligators from using vocal and olfactory cues in order to navigate the T-maze. By constructing the T-maze with two arms, each originating from a start box and ending at the free space, no matter which arm the alligator chooses the olfactory and auditory cues will increase as the alligator moved towards the free space. However, these cues increased to the same degree in each arm, meaning that these cues may have encouraged the alligator to swim towards the free space but did not aid the animal at the choice point. The single choice point T-maze consisted of a free space, a start box, a choice point and two arms. The start box led to a single choice point, where an individual had to choose between the left or right arm of the maze. Each arm led to a single entrance to the free space. At the end of each arm a ramp was installed. This ramp ensured that the arms could only be one-way streets. As trials progressed, more and more alligators occupied the free space. It was necessary to design a means of preventing the individuals who had already completed their trials from entering the arms of the T-maze. The ramps accomplished this goal. All areas were filled with water 10.2cm deep. Four submersible heaters and two aquarium pumps maintained the water temperature to ensured reduced variability throughout the maze. The circulation of water by the aquarium pump reduced any chemical cues left behind by the previous alligator. The modified T-maze was maintained at a temperature of 30°C (Figure 4.1). Animals were trained to swim in the T-maze using a progressive protocol that culminated in data collection. The goal of the first step of the training protocol 74 was to train the alligators to swim down the arms of the T-maze as well as train them to expect both doors to provide entry back into the free space. Animals were trained five days a week, Monday through Friday. Training began by removing the alligators from the free space of the T-maze and placing them in a smaller holding tank. The holding tank contained 10.2cm of water maintained at a temperature of 30°C by a submersible aquarium heater. At the beginning of each trial, a single animal would be removed from the holding container and placed in the start box. During training session both arms were open and led to the free space. A single trial consisted of the experimenter removing a single alligator from the holding tank, placing it in the dry start box and allowing it to walk out off the start box platform and into the T-maze filled with water. At the choice point, the alligator would then have the option of swimming down the left or the right arm of the T-maze. The alligator was considered to have made a choice once all four limbs passed a designated line within the arms of the T-maze, and the experimenter recorded the selection Initially alligators did not automatically swim away from the experimenter and down the arms of the T-maze when placed in the start box. Most individuals required some encouragement to swim the length of the T-maze arm and walk up the ramp in order to return to the living space. Alligators were motivated by several methods. Experimenter would clap 5.1cm diameter PVC end caps together or shake a piece of plastic tarp over the gators head in order to motivate them to swim. Finally, motivating the alligators to walk up the ramps at the end of the alleyways proved most difficult. In this scenario the experimenter had to 75 touch the base of the alligators tail with an aquarium net in order to motivate the animal to walk up the ramp initially. Once all alligators from all groups were swimming the arms of the T-maze without motivation techniques, the second step of the training protocol was begun on the next scheduled day of training. The goal of the second step of the training protocol was to train the alligators to swim through the one-way doors. Each door was constructed from a plastic cafeteria tray and an entry way was drilled in the center of the tray. A piece of black opaque plastic covered the entryway and extended into the water. Velcro was either attached to the top and bottom of the entryway, when the door was "closed", or simply just to the top when the door was to be "open". A single one-way door was installed at the end of each arm. By passing through the entryway of these one-way doors the alligator was led onto the ramp and back to the free space of the T-maze. Initially, after the introduction of the one-way door the alligators again need some motivation to swim through the door and walk up the ramp. Once all the individuals were swimming through the one-way doors on a consistent basis the third and final step of the training protocol was begun. In the third and final step of the training protocol, one by one, an individual would be removed from the holding tank and placed in the start box. The alligator would exit the start box, swim down an arm of the T-maze (thereby choosing between the left and right alleyways), swim through the one-way door, walk up the ramp and be returned to the free space. The alligator's choice would then be recorded. The one-way doors are not visible to the animal until it enters one of 76 the arms of the T-maze. During these trials both the left and the right one-way doors were open so that both doors led the animal back to the free space. The goal of the final step of the training protocol was to ensure that there was not a group positional preference. In other words, we wanted to ensure that as a group the alligators did not prefer one arm of the T-maze over another. To test for a group positional preference, all the individual alligators were given a single trial for 4 days in a row. No positional preference was observed. Specifically, the results were as follows in run I 35 alligators selected the left alley way, 37 alligators selected the right alley way, in run II 32 alligators selected the left alley way, 40 animals selected the right alley way, in run III 33 individuals selected the left alley way, 39 alligators selected the right alley way and in run IV 32 animals selected the left alley way, 40 alligators selected the right alley way. After establishing that a group positional preference did not exist, data collection was begun on the next scheduled day of training. During each data collection trial, one arm of the T-maze, the "incorrect" arm led to a closed one-way door and the "correct" arm lead to an open one-way door. Only by selecting the correct arm will the alligator be able to return to the free space. All individuals were given two to four trials a day; with a 2 hour interval between trials, and the number of incorrect trials were recorded. Each correct trial was rewarded with entrance into the free space. However, if an animal selected the incorrect arm of the T-maze, once it reached the one-way door, the animal was picked up and moved to a small non-home cage container were the animal would be housed individually for the 2 hour interval trial. After 77 this 2 hour period all the animals were again placed in the holding container and the next scheduled trial was begun. The total number of correct and incorrect trials was recorded for every individual in each embryonic group and the group average was calculated. The "correct" choice was reversed after all three embryonic groups achieved a group average of at least 70% correct trials in a single day for 4 days