Ricerca

...  ...

... "AXIS OF EVIL" AND "GREAT SATAN? " AN OPERATIONAL CODE ANALYSIS OF U.S.-IRAN MISPERCEPTIONS Zuzanna Maria King October 2020 1 The thesis titled "AXIS OF EVIL" AND "GREAT SATAN? " AN OPERATIONAL CODE ANALYSIS OF U.S.-IRAN MISPERCEPTIONS written by Zuzanna Maria King has been read and accepted on behalf of the faculty of the Department of History and Political Science, University of Indianapolis, in partial fulfillment of the requirements for the Master of Arts in International Relations degree Dr. Douglas Woodwell, Thesis Advisor Dr. Jyotika Saksena, First Reader Dr. Milind Thakar, Second Reader 2 ABSTRACT For the past forty years, the relationship between the United States and Iran has been defined in terms of hostility and mutual suspicion. Unfortunately, many scholars have subscribed to this view in the past, which was the product of often oversimplified and far too generalized approaches to studying foreign policy. While most traditional analysis methods focus on a macro-level examination of decision making and the strategic behavior of a nation as a whole, this study proposes an alternative viewpoint. Individual actors play a central role in the decision-making and shaping of foreign policy. Therefore, evaluating their beliefs and perceptions about the external world is crucial to understanding foreign policy consequences manifesting in the interactions between nations. This study aims to thoroughly investigate the belief systems of the U.S. and Iranian leaders between 1953 and 2020 to assess the true nature of their relationship and each one's strategic motivations. The leading research questions addressed in this thesis are: What are the belief systems of the U.S. and Iranian leaders that drive the interactions between the two countries? What are some shared patterns among these leaders' beliefs, and how does each compare to the average leader? Are these results consistent with theoretical assumptions about their relationship? To answer the research questions, this work employs the quantitative Operational Code Analysis, which uses individual leaders' verbal material to determine their belief systems. The outcomes are then analyzed by performing a series of difference of means tests. Following their statistical preferences are assessed in accordance with the principles of Brams' (1994) Theory of Moves (TOM) and Theory of Inferences about Preferences (TIP). The results show that the representation of the U.S. and Iranian leaders' nature and motivations across conventional wisdom is not entirely accurate. Hence, reinforcing the value and significance of individual-oriented approaches to studying foreign policy. Keywords: Iran, United States, foreign policy, operational code analysis, belief systems, strategic preferences, Theory of Moves (TOM), Theory of Inferences about Preferences (TIP) 3 ACKNOWLEDGMENTS First and foremost, I would like to thank those from whom I have received the greatest support, my dear parents, Justyna and Dariusz Michalak. Their relentless financial assistance and spiritual guidance made it possible for me not only to pursue my masters degree but to also effectively complete this thesis. I would like to extend my sincere gratitude to my supervisor, Dr. Douglas Woodwell, who has directly contributed to developing my passion for quantitative research. Your expertise and words of encouragement throughout this time were greatly appreciated. Special thanks to Dr. Michael Young, an internationally renowned expert in automated text analysis and the original creator of Profiler Plus, the coding software used in this study. I am incredibly grateful for the time and knowledge that you have selflessly shared with me during this process. I would also like to express my genuine appreciation to Dr. Jyotika Saksena and Dr. Milind Thakar, whose insightful feedback allowed me to strengthen the quality of my work. A heartfelt thank-you for the constructive criticism, valuable advice, and exceptional guidance that you have offered me on this path. In addition, I would like to acknowledge my brother Jan Michalak, as well as my friends, particularly my childhood friend Agnieszka Kutz for their help and encouragement. Thank you for keeping me in good spirits when faced with daunting tasks or when in doubt. Last but definitely not least, I would like to recognize my beloved husband, Andrew King, for his enormous support and patience that cannot be underestimated. Thank you for never doubting me and always being there for me. Your unshakeable faith in my abilities is what has truly empowered me to successfully complete my thesis. 4 TABLE OF CONTESTS TITLE PAGE.....................................................................................1 SIGNATURE PAGE..........................................................................................2 ABSTRACT...................................................................................3 ACKNOWLEDGMENTS.............................................................4 TABLE OF CONTENTS...............................................5 LIST OF TABLES.........................................................7 LIST OF FIGURES.......................................................8 CHAPTER I: INTRODUCTION...................................9 1.1 Overview and the Purpose of the Study..................9 1.2 Historical Background and Significance of the Topic......................11 1.3 The Structure of Research.............................13 CHAPTER II: LITERATURE REVIEW....................15 2.1 The Development of the Operational Code Analysis...........15 2.2 Quantitative Literature..............................20 2.3 Qualitative Literature....................................21 CHAPTER III: METHODOLOGY AND RESEARCH DESIGN..............25 3.1 The Operational Codes.............................25 3.2 Verbs in Context System (VICS) and Profiler Plus.26 3.3 Data Selection..........................................................26 3.4 Data Analysis...........................................................28 3.5 Theory of Inferences About Preferences (TIP)........................................29 5 3.6 Belief System Typology...........................................30 CHAPTER IV: DATA ANALYSIS AND DISCUSSION.........31 4.1 Philosophical and Instrumental Beliefs....................31 4.2 Phase I: 1953-1978...................................................32 4.3 Phase II: 1979-1997..................................................38 4.4 Phase III: 1998-2001.................................................41 4.5 Phase IV: 2002-2016.................................................43 4.6 Phase V: 2017-2020..................................................47 CHAPTER V: STRATEGIES AND LEADERSHIP PROPENSITIES..51 5.1 Phase I: A Special Friendship...........................................53 5.2 Phase II: Iranian Revolution and the Hostage Crisis........54 5.3 Phase III: Khatami and Dialogue Among Civilizations ...55 5.4 Phase IV: Axis of Evil and JCPOA...................................56 5.5 Phase V: President Trump and General Soleimani...........57 5.6 Testing the Hypotheses.....................................................58 CHAPTER VI: CONCLUSION..........................................................60 6.1 Efficiency of the Analysis Method...................................61 6.2 Making Sense of the Outcomes........................................61 6.3 Final Recommendations...................................................62 SELECT BIBLIOGRAPHY................................................................64 APPENDICES.....................................................................................70 6 LIST OF TABLES Table 1. George's ten questions about philosophical and instrumental beliefs.......16 Table 2. A Theory of Inferences About Preferences (TIP).........................................30 Table 3. The Operational Codes of the U.S. and Iranian Leaders in 1953-1978............34 Table 4. The Operational Codes of the U.S. and Iranian Leaders in 1979-1997............40 Table 5. The Operational Codes of the U.S. and Iranian Leaders in 1998-2001............42 Table 6. The Operational Codes of the U.S. and Iranian Leaders in 2002-2016............45 Table 7. The Operational Codes of the U.S. and Iranian Leaders in 2017-2020............48 Table 8. U.S.'s and Iranian Strategies 1953-1978...........................................................53 Table 9. U.S.'s and Iranian Strategies 1979-1997...........................................................54 Table 10. U.S.'s and Iranian Strategies 1998-2001.........................................................55 Table 11. U.S.'s and Iranian Strategies 2002-2016.........................................................56 Table 12. U.S.'s and Iranian Strategies 2017-2020.........................................................57 7 LIST OF FIGURES Figure 1. Revised Holsti Operational Code Typology............................................19 Figure 2. Cartesian Coordinate System...................................................................52 8 CHAPTER I: INTRODUCTION 1.1 Overview and the Purpose of the Study When approaching the study of foreign policy, it is critical to identify the underlying factors that play a crucial role in foreign policy decision-making, which consequently defines the nature of the relations between states. The more traditional among the international relations scholarship, concerning foreign policy, consider the nation-state to be the main subject of analysis, rather than directly studying the powerful state leaders-the individual actors who influence foreign policy first-hand. As the nation-states are essentially comprised of the decision-makers, it becomes necessary to examine their belief systems as a causal factor in policymaking (Richter, 2016). Thus, the scope of this thesis encompasses the examination of U.S.-Iran relations through the analysis of individual leaders' foreign policy beliefs. Over the years, the relationship between the United States and Iran has been debated among many scholars, country officials, and even the general public. Unfortunately, the common wisdom regarding U.S.Iran relations, frequently manifesting within the literature and the opinion of society, appears to resort to "uncritical and unproblematic treatment of our enemies as well as as an assumption that they are almost naturally destined to be our enemies". Furthermore, those enemies are usually perceived as irrational lunatics, leaders, hostile by nature, who use ideology to justify their aggressive foreign policy toward the U.S. and its allies (Malici and Walker, 2016, pp. 4-5). The thesis argues that each country's assumptions about another's beliefs are too often based on prejudice or long-standing and unchallenged misrepresentations, directly causing the tensions within the relationship. The general question that this study focuses on is: What are the actual foreign policy beliefs and motives of the U.S. and Iran, and are they consistent with one another's assumptions? In order words, 9 we examine whether this relationship's rhetoric is built on realities or if it is solely an outcome of distorted preconceptions on both ends. This work addresses the following research questions: 1. What are the philosophical (diagnostic) and instrumental (prescriptive) beliefs of the American and Iranian leaders? 2. How do the foreign policy beliefs of the examined decision-makers compare to the average world leader? How do they compare to one another? 3. What are the strategic orientations of the U.S. and Iran in pursuing their foreign policy objectives? Do they approach the issue of foreign policy in a matter that is predominantly rational/pragmatic or ideological/revolutionary? 4. How do these leaders view one another? How accurate are their assumptions about the other's strategies? This study attempts to close the gap between speculations and the reality of the U.S.-Iranian relations by testing their beliefs about one another against another's actual belief systems. The study uses qualitative literature to determine both countries' beliefs about the nature of their interactions. These theories are then examined in relation to quantitative data obtained from operational code analysis in order to reduce "analytical uncertainties" (Leverett, 2005, p. 18) about this relationship. Through the application of the operational code analysis method, this thesis demonstrates that the antagonistic nature of the relationship between the countries is not entirely the result of facts, but mainly the result of false assumptions about one another. The work also evaluates the United States and Iran's foreign policy 10 strategies to explain the specific character of their relationship by using the belief systems of the countries' leaders. Considering the historical aspect of the U.S.-Iranian relations, there is no doubt that their leaders hold a pessimistic view of each other as a hostile and uncooperative nation. However, what this thesis aspires to do is to check the validity of these assumptions. The research's underlying objective is to confront negative narratives dominating the way the countries see one another in hopes that a more thorough understanding of the driving forces behind the opponent's actions will allow for the relations between the U.S. and Iran to improve. 1.2 Historical Background and Significance of the Topic On November 4, 1979, a group of Iranian students occupied the U.S. Embassy in Teheran, taking hostage 66 people, including 52 U.S. citizens. The student-supporters of Ayatollah Ruhollah Khomeini, the Islamic Revolution leader, demanded that Iranian authoritarian ruler Mohammed Reza Pahlavi, a long-term American ally, is extradited to face a trial in Iran. That day was the beginning of a 444-day diplomatic standoff, which forever damaged the relations between the United States and Iran. Quite unfortunately, until the 1979 Hostage Crisis, many Americans knew nothing or very little about Iran. Even more unfortunate was the fact that on November 4, they finally learned about the Iranians-angry religious fanatics, the "devilish savages" (Campbell, 1993, p. 43) holding hostage innocent U.S. civilians. Since then, both the U.S. and Iran seem to be hanging onto certain false beliefs regarding one another, contributing to their growing mutual apprehension, further damaging their relations. While the events of 1979 originated a narrative among the American public dominated by intense opposition to Iran, the decline in the Iranian opinion of the U.S. dates back to 1953. On August 19, 1953, CIA and MI6-British intelligence forces carried out the "Operation Ajax", which led to regime change within Iran. The coup d'tat overthrew Mohammed Mossadegh- the Iranian Prime Minister and Chairman of Anglo-Iranian Oil Company, who was brought into the office in 1951 11 popular election. Unfortunately for Britain, shortly after the election, the Iranian Parliament held a vote resulting in the nationalization of the company. The oil nationalization was viewed by the U.S. and Britain as part of the Cold War power struggle, and therefore, likely to disturb the equilibrium across the region. The consequence of this belief was the "Operation Ajax", which replaced the popular Prime Minister with the autocratic Mohammad Reza Shah. As a result of the 1953 coup d'tat, there seems to be a notion among the Iranian scholars and clergy that the United States is after them, seeking regime change and Iran's global isolation. Similarly, it appears that some U.S. policymakers believe that Iran is extremely hostile toward America and is driven by the desire to sabotage U.S. efforts in the Persian Gulf (Sariolghalam, 2003). Historically, the fact that their interactions were marked with aggressive behaviors partially explains the emergence of anxiety that resulted in these presumptions. However, the mere existence of these beliefs is not sufficient evidence for their correspondence with the truth. Given the geostrategic importance of Iran and the US's global status, one would assume that the two sides would make efforts to mend their relationship. Besides the 615 miles of coastline that belong within Iranian territory, the country geographically dominates the Strait of Hormuz. Referred to as a "global chokepoint", the Strait of Hormuz serves as a point of access for oil tankers to global markets; therefore, it is highly relevant not only on a geopolitical but also on an economic level. As a result, in the face of relations with the U.S., Iran's geostrategic significance seems to carry a more substantial value than the ideological apprehensions (Ali, 2008). At the same time, Iran lies within the region that is very strongly affected by prolonged conflict and, therefore, directly linked to the matters of international security. The importance of analyzing the U.S.-Iranian relations also involves the American hegemonic position on the international stage and its continuous efforts of extending influence over the region through foreign policy toward the Middle East, essentially translating into multidimensional global security interests. In his writings, Kamrava (2011) expands on this theory, suggesting that that "Iran's regional security policy, in the meanwhile, is largely determined by the role and position of the United States in what Iran considers its rightful sphere of influence" (pp. 184-185). 12 1.3 The Structure of Research The countries' opinions of each other had been collected across various descriptive studies, while their actual belief systems were obtained from quantitative analysis. The method employs operational code analysis of the leaders' beliefs using verbal material of the U.S. and Iranian leadership within five distinct periods, which were determined influential to the two countries' relationship. For the purpose of examining beliefs and behaviors of the subjects, this method combines the operational code analysis with elements of a game-theoretic approach, which allows us to measure: how the leaders see themselves and the other; what are their foreign policy motivation; and whether these agents are oriented towards conflict or cooperation in their policy considerations. A further statistical comparison is conducted, serving as a tool for observing the differences and similarities between the two subjects' beliefs and those of the average leader. Additionally, this work examines how the beliefs of the American and Iranian leaders evolved over time. The analysis is divided into five phases across time, separated by significant historical events, which are believed to have generated critical shifts in the narrative of U.S.-Iranian relations. The first of the five episodes begins on August 19, 1953, with the U.S.-led coup d'tat, which resulted in Prime Minister Mossadegh's overthrow. The second one begins on April 1, 1979, the day Iran voted for the national referendum to become the Islamic Republic and create the new Constitution of the Theocratic Republic. The third episode starts in 1998 after the election of Mohammad Khatami as the new Iranian President. The next episode commences on January 29, 2002, when George W. Bush delivers his speech before the Congress in which he names Iran a part of the "axis of evil". This episode terminates with the end of 2016, the year of implementation of the JCPOA nuclear agreement. The last episode begins on January 20, 2017, with Donald Trump, who openly opposed the Iran nuclear deal, assuming the President of the United States' office. This episode concludes with Iranian general Soleimani's death on January 3, 2020. The organization of this thesis incorporates a total of six chapters, beginning with the introduction and ending with the conclusion. Following the introductory section is chapter 2, which takes a closer look 13 at the relevant literature. It begins by illustrating the essential aspect of the operational code analysis, followed by an overview of the quantitative approach within the relevant literature. The chapter ends by discussing the U.S. and Iranian beliefs that are the most prevalent across the qualitative studies on this subject. Chapter 3 presents the methodology and the research design, explaining the selection and preparation of data, as well as the process of analysis. Chapter 4 demonstrates the quantitatively measured belief systems of U.S. and Iranian leaders, resulting from the VICS analysis and difference of means test. These outcomes are then compared to the assumptions made by each country's leaders about the other's beliefs. Chapter 5 applies the VICS scores for master beliefs to construct the strategic preferences of the Iranian and U.S. leaders. The chapter also asks whether the leaders' predictions about their counterpart's strategies and tactics corresponded with reality. Chapter 6 discusses the research method's efficacy, outlines the finding, and incorporates general policy suggestions. 14 CHAPTER II: LITERATURE REVIEW The goal of chapter 2 is to evaluate the existing literature applicable to the focus of this study. Based on the methodological model, the literature has been organized into three groups. The first group examines the origin and evolution of the operational code analysis, as well as the fundamental principles of this approach. The following section evaluates the quantitative approach to studying the leaders' decision making and discusses how this method complements the research analyzing the interactions between states. Finally, the last section reviews previous qualitative works addressing the character of the U.S.-Iranian relations and their beliefs about one another. 2.1 Development of the Operational Code Analysis The concept of "operational code" was first defined by Merton (1940) as one that involves beliefs, world view, and vocabulary learned and exchanged by a person with other organization representatives. However, it was Nathan Leites, who first, in his studies: The Operational Code of Politburo (1951) and Study of Bolshevism (1953), applied this concept to the field of political psychology. Leites' 1951 analysis of the Politburo's operational codes was a combination of cognitive heuristics and characterological personality traits that driven Soviet decision-making (Walker, 1990). The study quickly gained popularity, as there was already a growing desire to understand what motivates the powerful elites. However, Leites' method was not without flaws. The first one to point them out was George (1969) who argued that, "While the complexity of the work adds to its richness and intellectual appeal, it has also made it unusually difficult for readers to grasp its structure or to describe its research mode" (p. 193). To resolve this issue, George decided to focus on the cognitive aspects of Leites' theory, emphasizing the analysis of beliefs as the "maxims of political strategy" George used Leites' division of "epistemological" and "instrumental" beliefs 15 and developed into a set of ten questions, which include five philosophical and five instrumental questions (see Table 1). The philosophical questions address the leader's beliefs about the fundamental nature of the political universe, nature of the conflict, and scope of the leader's role in the development of history. The remaining instrumental questions aim to uncover the leader's approach to political strategy. Along with the growth of the OCA (Operational Code Analysis) research program, scholars began to face additional difficulties. These mainly concerned the access to data from foreign leaders and ways to obtain their operational codes from sources that did not include material directly representing their philosophical and instrumental beliefs. These concerns eventually gave rise to the "at-a-distance" analysis method, which allows for identifying the leader's beliefs based on the patterns appearing in their verbal The Philosophical Beliefs in an Operational Code P-1. P-2. P-3. P-4. P-5. What is the "essential" nature of political life? Is the political universe essentially one of harmony or conflict? What is the fundamental character of one's political opponents? What are the prospects for the eventual realization of one's fundamental values and aspirations? Can one be optimistic, or must one be pessimistic on this score, and in what respects the one and/or the other? Is the political future predictable? In what sense and to what extent? How much "control" or "mastery" do (a) self and (b) other have over historical development? What is (a) self and (b) other's role in "moving" and "shaping" history in the desired direction? What is the role of "chance" in human affairs and historical development? The Instrumental Beliefs in an Operational Code I-1. I-2. I-3. I-4. What is the best approach for selecting goals or objectives for political action? How are the goals of action pursued most effectively? How are the risks of political action calculated, controlled, and accepted? What is the best "timing" of action to advance one's interests? I-5. What is the utility and role of different means for advancing one's interests? Table 1. George's ten questions about philosophical and instrumental beliefs. Source: George (1969, p. 200). materials. The method analyzes the content of individual's speeches using the Verbs in Context System (VICS), which enables the researcher to determine the answers to George's ten questions regarding philosophical and instrumental beliefs (Schafer and Walker, 2006). This work uses the software for an 16 automated version of VICS developed by Michael Young and Robert Woyach in the late 1990s, discussed in detail in the methodology section. As the operational code analysis continued to evolve, Holsti (1977) proposed a model of belief system typology composed of six types (A, B, C, D, E, F), each based on the source of conflict in the political universe (individual, nation, or, international), and its nature (temporary or permanent) (zdamar and Canbolat, 2018). This typology derives from the intersection of the two master beliefs about the nature of the political universe and the view about one's political opponent (Holsti, 1977). Holsti's typology was later revised by Walker (1983), where he combined and simplified the overlapping belief system into A, B, C, and DEF types. Many of the more recent OCA works combine the operational code analysis with Brams' (1994) game theory principles and the analysis of strategic motivations. Instead of looking at the game in a way proposed by structuralist or rationalists, this approach moves beyond a traditional concept of game-theory, attempting to define the game's nature, as seen or described by the subject. Schafer and Walker (2006) argue that the method defines the leader's beliefs about self's best approach and strategy and self's beliefs about the other's likely approach and strategy. These beliefs then form a subjective dyadic game that mirrors the interactions anticipated by the subject. The authors conclude that, It may be wrong, or off-target compared to reality, and external observers may be able to identify a game that more directly fits reality. In the end, however, it seems that the subjective beliefs held by a leader are the ones that are most likely to influence his/her choice of moves (Schafer and Walker, 2006, p. 11). While a general operational code analysis allows for determining the individual's beliefs, it does not reveal the impacts of behaving according to those beliefs. Conventional game-theory is a passive illustration of strategic interactions between two actors. What allows for dynamic analysis of these interactions is the combination of Brams' Theory of Moves (TOM) and the Theory of Interferences about Preferences (TIP). By identifying a set of beliefs from the possible combinations of subjective games described in TIP, this method models the strategies and respective equilibria formed for these subjective 17 games with strategies, moves, and tactics described in sequential game models of TOM. The combination model of TIP and TOM strategic preferences allows us to predict and understand the connections between beliefs, strategies, and outcomes concerning the relationship between two actors. It also helps determine how individual leaders are likely to exercise power and what types of political environment is likely to result from the exchange of power between them (Walker, 2004; Walker and Schafer, 2006). Because so many scholars had recognized the value and potential of the operational code analysis, the method had undergone a significant improvement over the years, causing it to become an incredibly useful tool for studying world leaders' belief systems. 18 Figure 1. Revised Holsti Operational Code Typology. Note: (Instrumental beliefs are in bold, and philosophical beliefs are not.) Source: Walker and Schafer (2006) 19 2.2 Quantitative Literature Review The more traditional international relations scholarship, concerning foreign policy, consider the nation-state to be the main subject of analysis, rather than directly studying the powerful state leaders-the individual actors who influence foreign policy first-hand. As the nation-states are essentially comprised of the decision-makers, it becomes necessary to examine their belief systems as a causal factor in policymaking. (Richter, 2016) Malici (2006) argues that "To consider actors as amorphous entities and to rob them of any consciousness is equivalent to denying the sociopsychological character of politics" (p.127). Rational Actor Models (RAM) suggests that the decision-makers act rationally concerning "longterm and persistent national interests of the country, and since national interests do not change, changes in leadership have little consequence" (Neack, 2008, p. 31). These models assume that "one size fits all". While they may accurately address the foreign policy strategies at the macro level, they fail to provide an exhaustive interpretation of the factors influencing individuals' decisions, which include religious, political, and moral beliefs, along with historical knowledge and experiences. These models disregard the autonomy of individuals, which acts as a direct cause of bias in their decision-making. (Richter, 2016) Quantitative "at a distance" leadership studies, such as the OCA, attempt to address those areas of foreign policy studies, which rational choice theory cannot. Feng (2006) claims that "decision-makers' beliefs are a critical factor in understanding, diagnosing, and prescribing the decision-making process of states, particularly under situations featuring high uncertainty and incomplete information" (p. 156). He continues by saying that individual beliefs strongly affect how a leader processes information by illustrating all presumptions and speculations that the subject considers to be true regarding the given situation. This method's advantages are demonstrated by Walker (et al., 2011), who compares the study of international relations to quantum physics. According to his theory, individual leaders resemble the characteristics of particles of energy, while the states compare to planet-size particles of matter. Walker (et al., 2011) claims that the field of international relations experiences the crisis of the observer effect, such as that described in quantum theory, where the principles applied to the behavior of large-scale particles of 20 matter (states) need to be thoroughly modified by the principles of quantum mechanics to justify the behavior of electrons and photons-the small particles of energy (individual leaders). zgr and Canbolat (2018) claim that "operational code analysis and its methodology is one of the strongest political psychology tools available to analyze agents' preferences and the implications of these preferences on foreign policythe approach reveals interesting and counterintuitive results with relevant policy implications" (p. 28). Moreover, even though it is beyond the scope of this work, operational code analysis illustrates the principles of the traditional International Relations theories. Walker and Schafer (2007) explain that there is a widespread agreement among political psychologists, constructivists, and institutionalists that "neorealist and neoliberal structural theories are underspecified without including agent-oriented models of beliefs to capture the micro-foundations of strategic interactions between states" (p. 771). They present a compelling argument that the most effective approach to studying aspects such as decision-making within international relations is the combination structural analysis of particular cases and an individual-level analysis of agents' belief systems. As this work strives to deliver the evaluation resembling a high level of comprehensiveness, it follows that exact methodological model. 2.3 Qualitative Literature Review The literature's common theme regarding the relationship between the United States and Iran seems to revolve around their mutual animosity. While this view may often reflect reality, it also appears as if the U.S. policymakers and academics lack a sufficient understanding of Iran's political leaders' beliefs (Malici and Buckner, 2008). However, the literature addressing the evolution of interactions between Iran and the United States successfully captures historical events that appear to have had significantly influenced the direction of their foreign policy towards one another. The assessment of that literature and common theories surrounding the relationship is presented below. According to Ali (2008) and Ayoob (2011), the events of August 19, 1953, and the "Operation Ajax" discussed in chapter 1, forever defined the narrative surrounding the U.S.-Iran relations, causing the 21 Iranian public to begin viewing America in a very negative light. Moreover, Ayoob (2011) suggests that by reinstalling the rule of the Shah, who eliminated any secular protestors of his regime, the United States created an environment for the Iranian clergy to become a leading group opposing the Iranian government. That is precisely why, Ayoob (2011) claims, some analysts define the 1979 Islamic Revolution as an inevitable if a delayed reaction to 1953. Ali (2008) argues that the period between 1953 and 1978 was characterized by seemingly positive relations between the two countries. However, they did not reach past the leadership. During that time, the United States policymakers were wrongfully convinced that Iran's government was in control and that there was no evidence of opposition's capacity, which caused them to miss the apparent signs of growing tensions within the country (Reisinezhad, 2011). Ironically, the two probably most notable incidents in the history of U.S.-Iranian relations occurred in 1979-not too long after Jimmy Carter called Iran the "island of stability" in 1977. The Islamic Revolution and the U.S. embassy hostage crisis of 1979 caused the hostilities underlying the relationship to surface and escalate. Maloney (2002) notes that the Iranian hostage crisis "fused the extremist dimensions of the divergent worldviews remaining within the revolutionary coalition: radical anti-Westernism and vehemently Islamist self-identication". She further indicates that the hostages represented "both symbol and stake in the internal struggle to dene the nature of post-revolutionary Iran" (p. 105). Hurd (2008) extends this view by arguing that the U.S. public's aversion to Iran, which became evident after the hostage crisis, is a product of American secularism ingrained in American identity. Hurd believes that American secularism not only "conditions American foreign policy" but also "constitutes what it means to be American" (pp. 103-104), therefore legitimizing the animosity toward Iran. According to this theory, since 1979, the U.S. foreign policy toward Iran is marked by secular bias, causing the American policymakers to view Iranians as irrational religious extremists, differing entirely from reasonable, secular Americans. As for the Iranian approach to foreign policy, while many see it as "revolutionary", or ideologydriven, Kamrava (2011) believes it is more practical and realistic, shaped by balance-of-power logic and security concerns. Nevertheless, post-revolutionary Iran's activities within the Gulf region during the 1980s 22 caused plenty of other actors to begin perceiving it as a threat. However, it is possible that in posing as a threat, Iran was motivated by security-related issues. When Mohammad Khatami was elected to be President in 1997, the tone of Iranian foreign policy became more friendly toward the region and the U.S. However, whatever dialogue Khatami's efforts managed to establish ended abruptly on January 29, 2002. That was when, in his infamous speech before the Congress, President George W. Bush labeled Iran, along with North Korea and Iraq, as the "axis of evil", accusing them of sponsoring terrorism and acquiring weapons of mass destruction. The situation did not improve when, in 2005, Mahmoud Ahmadinejad became the new President of Iran. Interestingly enough, when it comes to Ahmadinejad, the literature does not appear to be seeing eye to eye. Ali's (2008) view of President Ahmadinejad reflects that of many U.S. policymakers, who perceived the president as an irrational actor, guided by belligerent motives. Considering Malici and Buckner's (2008) analysis of Ahmadinejad's operational codes, as well as Kupchan and Clifford's (2005) study Iranian Beliefs and Realities, this work argues against Ali's (2008) claims. Malici and Buckner (2008) describe Ahmadinejad as a rational leader, who though sees the U.S.' sanctions as "highly hostile and threatening", favors cooperation strategy over conflict. Similarly, Kupchan and Clifford (2005) consider Iran under Ahmadinejad "stable". Additionally, they point out that, in this particular case, the policymakers fail to see the full picture by focusing on an individual leader. That prevents them from adequately assessing whether or not Iran poses a threat, as Iranian foreign policy decisions are made among the political elite and overseen by the Supreme Leader, rather than the President. Nevertheless, throughout the years, the relationship between the U.S. and Iran remained distant and continued to be defined by mutual suspicion. According to Abdullah (et al. 2020), it began to improve during the administration of Iranian President Rouhani, who successfully reached a nuclear deal with P5+1 countries. The signing of the JCPOA agreement pushed forward the lifting of sanctions against Iran, some of which dated back to 1979. However, any hopes for further progress in the relationship between the two countries were shattered when President Trump declared the U.S. withdrawal from the treaty (Abdullah et al., 2020). However, leaving the Iranian nuclear deal was not the only "bold" action taken toward Iran by 23 the Trump administration. On January 2, 2020, the U.S. drone strike successfully assassinated Qasem Soleimani- the head of the Iranian Quds Force. Suddenly some began predicting a "global war" resulting from Iran's retaliation in response to Soleimani's death. However, a good chunk of literature provides empirical evidence that nothing like a war between the U.S. and Iran will take place, at least not any time soon. While Iran's foreign policy is oriented more toward cooperation than conflict, Iran's acts of aggression are driven strictly by a security dilemma. As Asmar (2020) has observed, Iran responded to Soleimani's killing by firing missiles at two Iraqi military bases where the U.S. soldiers were stationed. The author argues that the missiles, rather than being a legitimate threat to the Americans in the region, served more as a psychological instrument, addressing the regime's need to display their power in order ensure security. The following chapters of this thesis are intended to extract similar assumptions about foreign policy and relations between the U.S. and Iran to those discussed here and re-evaluate them against the results of quantitative analysis. Overall, the qualitative literature agrees that throughout the years, the United States leaders had defined Iran as a rogue state, guided predominantly by extremist ideology. It further states that the Iranian decision-makers believed in the hostile nature of the United States, as well as their eagerness to isolate Iran internationally. The examined literature highlights the importance of combining qualitative and quantitative research models in order to provide a more meritorious explanation of the nature of state interactions. This study aims to contribute to foreign policy studies by emphasizing the value of the actor-oriented approaches while pointing out the misleading nature of the over-simplified theories regarding the states' behavior. 24 CHAPTER III: METHODOLOGY AND RESEARCH DESIGN This chapter offers a detailed overview of the research methods and describes the individual steps necessary to perform a thorough analysis of the data. Moreover, it discusses the process of data collection, including the criteria according to which the data has been selected, the size of the sample, as well as the sources and databases from which the texts were obtained. 3.1 The Operational Codes According to the qualitative research on the subject, both Iranian and American policymakers seem to have strong beliefs about the other's hostility, which drives their foreign policy decisions. The methodology described below has been chosen because it produces empirical data that relies entirely on the evidence derived directly from the collected texts of individual leaders. When tested against these assumptions, this data presents proof for either inaccuracy or correctness of the conventional beliefs about this relationship. The study of operational codes proves to be especially beneficial within circumstances featuring a high level of uncertainty, and the relationship between the United States and Iran is undoubtedly that. Unlike other research methods, quantitative analysis of leaders' belief systems allows for more objective conclusions and is likely the most direct way of studying foreign policy decision making. That is because the approach is designed to examine the leaders' actions and strategic motivation through "their own eyes", as opposed to third-person observation and subjective interpretation of their behaviors, usually featuring unavoidable bias. Furthermore, by presenting the subject's viewpoint and that of his opponent, the study generates a dyadic representation of the relationship. Additionally, the method allows for an efficient 25 yet, exhaustive comparison of the subjects' beliefs with those of the average leader from the norming group, the possibility which other approaches are not able to provide. 3.2 Verbs in Context System (VICS) and Profiler Plus For this study's purpose, the leaders' recorded speeches are used as fundamental units of analysis. The beliefs obtained from them serve as causal devices for foreign policy decisions across Iranian and American regimes (zdamar and Canbolat, 2018). One of the initial steps in the process is completed using Profiler Plus V. 7.3.17, automated software that applies VICS (Verbs in Context System) coding method to the speeches. The software analyzes the speeches' content, recognizes patterns, and assesses the values of the leader's philosophical and instrumental beliefs defined in George's (1967) ten questions. In order to reach valid results, the data selection must conform to the criteria specified by Walker, Schafer, and Young (1998): "(1) the subject and object are international in scope, (2) the focus of interaction is a political issue, and (3) the words and deeds are cooperative or conflictual" (p.182). Besides, the Profiler Plus software must be able to produce a minimum of fifteen coded verbs for the speech to be included in the analysis (Schafer and Walker, 2006). 3.3 Data Selection In an effort to calculate the American and Iranian leaders' operational codes between 1953 and 2020, a total of ninety-five speech acts were collected. Fifty-two of which belonged to the U.S. decisionmakers such as the Presidents, Vice Presidents, U.S. Ambassadors to Iran, Secretaries of State, etc. The majority of these transcripts were accessed through the website of The American Presidency Project hosted by the University of California, Santa Barbara (presidency.ucsb.edu), and from multiple volumes of the Foreign Relations of the United States, available at the official website of Office of the Historian of the Department of State in Washington, D.C. (history.state.gov). Other transcripts were obtained from the 26 presidential libraries' websites (fordlibrary.org, jimmycarterlibrary.gov, reaganlibrary.gov) and databases such as Nexis Uni (formerly Lexis Nexis). The remaining forty-three texts came from the Iranian officials; the Shah of Iran, the Supreme Leader, Presidents, Foreign Ministers, and the Iranian Ambassador to the United Nations. Most of the transcripts were collected from the Islamic Republic of Iran Ministry of Foreign Affairs website (mfa.ir), as well as the Office of the Supreme Leader website (leader.ir) and khamenei.ir. Plenty of texts were also acquired from databases such as irandataportal.syr.edu and Nexis Uni, while others came from the U.N. Library (digitallibrary.un.org) and the UNESCO (webarchive.unesco.org) archive websites. In addition, the sample includes a few statements from memri.org, npr.org, theguardian.com, pbs.org, and transcripts.cnn.com. In each phase 'n' represents the number of texts used for that particular episode. Because this work primarily examines foreign policy and the relations between the U.S. and Iran, all of the speeches focus on their foreign policy toward one another. Therefore, we indicate that the P-1 value, which tells us how an individual perceives others in the political universe, will always concern the specific actors they discuss in their speeches (Malici and Buckner, 2008). To obtain the results needed for this type of analysis, that is, the countries' specific beliefs about each other, some editing of the text is necessary prior to coding by the Profiler Plus software. Nevertheless, it must be noted that any editing of the original statements has been kept to a minimum for the sake of external comparability, as the sample is later compared with the norming group's data. Additionally, close attention has been paid to assure that the author's intention and the meaning of the text were not lost due to the editing. The steps taken as a part of the clean-up process were inspired by Lavikainen's 2016 study and are described below: 1. First, the references to very distant history, irrelevant to the events or foreign policy issues discussed in the speech, were edited out. 27 2. Direct quotes of people other than the speaker themselves were edited out, as they expressed the beliefs of the person being quoted, not those of the particular speaker, a distinction which the Profiler Plus software would not be capable of making. 3. Analogously to the previous step, all the quotes from books, poems, the Bible, the Quran, etc. were edited out. 4. Some of the content not relating to politics but a protocol, such as overly emphasized and repetitive pleasantries, were deleted. 5. The statements that exclusively discuss other countries' foreign policy, not involving or affecting the studied nations, were not analyzed. 6. Certain self-indicators and pronouns, which were likely to be coded by the software as other, were appropriately adjusted with minimum alteration of the original text. (e.g., George W. Bush saying: "American armed forces are at war against terrorism" would be changed to "Our armed forces are at war against terrorism," etc.) Thanks to the editing process, it can be assumed that the VICS's coding results successfully address how Iran and the United States perceive one another and what each believes is the other's preferred strategy to approach their political goals. 3.4 Data analysis After obtaining the leaders' scores for philosophical and instrumental beliefs in each phase, their average values and standard deviations were calculated. This information allowed for conducting a number of the difference of means tests. For every phase, there were three difference of means tests performed: (1) U.S. vs. Iran; (2) U.S. vs. Norming Group; (3) Iran vs. Norming Group. These tests act as a tool for understanding how the beliefs of each country's leadership compare to one another and to those of the average world leader. 28 The norming group mentioned throughout the study, and often referred to as the "average leader", are the mean scores recorder among the sample (n=168) of thirty-five world leaders from different geographical areas and historical time frames. These include leaders of both weak and strong nations and facing various economic circumstances. The data for the average leader has been kindly provided by Mark Schafer of the Department of Political Science at the Louisiana State University. The difference of means test shows which of the compared values present a statistical significance. It tells us whether the difference between the beliefs of the examined leaders and the average leader is 'large enough', or significant enough, to assume that the difference does, in fact, exist, or if instead, it was accidental (has occurred by chance). The norming group's values serve as a baseline for comparing and contextualizing the VICS scores reviewed in this study. The analysis and discussion reported in chapter 4 present a mechanism for translating the numbers into 'real-life' situations and provide a reader with a perspective for examining the meaning of the leadership belief systems. 3.5 Theory of Inferences about Preferences (TIP) The next step of the analysis utilizes a deductive Theory of Inferences about Preferences (TIP), which links the master beliefs with different preference orderings towards the strategic interaction outcomes of settlement, deadlock, domination, and submission between self and other (Walker, 2004; zdamar and Canbolat, 2018). That results in distinct general strategic orientations (one for self and one for other) for any provided interaction, which derive from each player belief system regarding the effects of settlement, deadlock, domination, or submission. Determining these will provide two significant elements of information regarding each leader's subjective game: the outcome expected by a particular actor and the anticipated response of his opponent (Schafer and Walker, 2006). 29 Self and other values and preference order in a 2 2 strategic game. Self Other Prop. 1. If (1-1, P-4a) or (P-l, P-4b) is (+, <), then Settle>Deadlock>Submit>Dominate= Appeasement Prop. 2. If (1-1, P-4a) or (P-l, P-4b) is (+, =), then Settle>Deadlock>Dominate>Submit= Assurance Prop. 3. If (1-1, P-4a) or (P-l, P-4b) is (+, >), then Settle>Dominate>Deadlock>Submit= Stag Hunt Prop. 4. If (1-1, P-4a) or (P-l, P-4b) is (-, <), then Dominate>Settle>Submit>Deadlock= Chicken Prop. 5. If (1-1, P-4a) or (P-l, P-4b) is (-, =), then Dominate>Settle>Deadlock>Submit= Prisoners' Dilemma Prop. 6. If (1-1, P-4a) or (P-l, P-4b) is (-, >), then Dominate>Deadlock>Settle>Submit= Bully "+" indicates above, and "-" indicates below the mean of the norming group. <, >, and = indicate below, above, and within one standard deviation of the norming average, which is P4a 1 SD. Norming group scores for N = 168 are P-1 = +.27, SD = 0.28; I-1 = +.35, SD = 0.34; and P-4 = 0.21, SD =0.11. Norming group data was obtained from Mark Schafer, Department of Political Science, Louisiana State University. Source: zdamar and Canbolat (2018) Table 2. A Theory of Inferences About Preferences (TIP) Source: Malici and Buckner (2008, p. 791) 3.6 Belief Systems Typology Finally, using the philosophical and instrumental indices, the following master beliefs are mapped on a Cartesian coordinate system (Figure 2) representing the four types of belief systems (A, B, C, DEF): P-1 (nature of the political universe); I-1 (strategic approach to goals); P-4a (belief about one's ability to control historical development); P-4b (belief about the other's ability to control historical development). The typology coordinate system is fixed along three motivational axes of power, affiliation, and achievement. Two main quadrants- upper and lower, represent two general foreign policy strategiesAssurance (upper) and Prisoner's Dilemma (lower). The upper half consists of two quadrants (Type A and Type C), divided into two smaller quadrants expressing the Assurance strategy's variations. That same pattern applies to the lower half. The coordinate system has been designed based on the mean values for the norming group of thirty-five world leaders, serving as a comparative reference (zdamar and Canbolat, 2018). 30 CHAPTER IV: DATA ANALYSIS AND DISCUSSION Chapter 4 presents an examination of the U.S. and Iranian beliefs about one another across the five episodes. The assumptions about each other earlier demonstrated in the qualitative literature are now assessed in relation to the VICS scores obtained through the operational code analysis. A short discussion about each country's view of its counterpart produces a distinct hypothesis, which is then either verified or rejected by the quantitative data. Next, a series of statistical tests are performed to compare each actor's beliefs to one another and the norming group. While this comparison addresses research question number two, a comprehensive evaluation of leaders' philosophical and instrumental beliefs addresses research question number one. 4.1 Philosophical and Instrumental Beliefs The VICS analysis results for each of the phases are displayed in the tables included in the following sections. The first five indices in each table reveal the leader's philosophical beliefs, providing the reader with an idea about the leaders' perception of external factors, such as the nature of the political universe and the other actors. The first value (P-1) represents the leaders' belief about the inherent nature of the political environment. Next is the P-2 index, which demonstrates whether the leader is optimistic or pessimistic about the prospects for realizing their political aspirations. P-3 score indicates the extent to which the individual is convinced about the predictability of the political future. Below is the P-4 value, expressing the balance between the level of control over historical development one attributes to self (P4a) and other (P-4b). Finally, the last of the philosophical questions (P-5) examines the individual's opinion of the role of "chance" in the development of political events and international relations. 31 Following the philosophical belief indices, the tables present the answers for the second group of George's (1969) questions, which deal with the leaders' instrumental beliefs. Contrary to the first group, instrumental beliefs represent one's internal view of themselves and what they consider to be the best approach to fulfilling their political goal. The instrumental questions, therefore, reveal the leaders' strategic and tactical preferences. Consequently, the I-1 value points out to the individual's direction of strategy (either cooperative or conflictual), while the I-2 unfolds one's opinion of the most effective tactics for pursuing the policy objectives. Risk orientation is measured by the I-3 score, which indicates how the leader perceives the risks associated with undesirable political outcomes. More specifically, this number shows how one deals with control, acceptance, and calculation of those risks. The I-4a and I-4b indices specify the best "timing" of action, showing the leader's tendency to shift their tactics in pursuit of advancing his or her interests. I-4a shows the propensity for switching between cooperative and conflictual moves, while I-4b shows shifting between words and deeds. The fifth instrumental question (I-5) determines the utility and role of different means in exercising one's power. The exercise of power is defined as the application of control in one's relationship with the other, achieved with positive or negative words and deeds. The positive or cooperative means include Reward, Promise, and Appeal, and the conflictual include Oppose, Threaten, and Punish. 4.2 Phase I: 1953-1978 The literature discussed in chapter 2 suggests that between 1953 and 1978, the leaders of Iran and the United States saw one another in a particularly positive light. Their foreign policy consequently emphasized their mutual friendship and reinforced the shared interests. That indicates that the American and Iranian beliefs about each other would show higher P-1 scores. Furthermore, their analogous view of themselves as cooperative nations predicts not only the collaborative direction of their strategies (I-1) and tactics (I-2) but also a sense of optimism about reaching their political goals (P-2). All these factors combined imply a higher likelihood of both countries using predominantly positive means in exercising 32 control in their mutual relationship (greater values for I-5 a, b, and c than for I-5 d, e, or f). Therefore, the first hypothesis upholds: Hypothesis 1: According to the U.S. and Iranian opinions of each other during 1953 and 1978, both nations' beliefs about the nature of the political universe (P-1), as well as their strategic (I-1) and tactical (I-2), orientations were more cooperative than those of the average leader. Both countries were also more optimistic about realizing their political values (P-2) and more prone to exercise their power via peaceful means (I-5a, b, c). The results in Table 3 demonstrate that during 1953-1978, the Iranian (P-1= 0.52) and American (P-1= 0.45) leaders believed the political universe to be "definitely friendly" (for reference, see Appendix 1). That essentially means that both counties' leaders saw the fundamental character of their political opponent (i.e., one another) as cooperative. Furthermore, both viewed the political world as more friendly than the average leader, as their P-1 values are significantly higher than those of the average leader (P-1= 0.27). When it comes to the prospects for the realization of one's political aspirations (P-2), the scores of both Iran (P-2= 0.29) and the United States (P-2= 0.25) suggest that they were somewhat optimistic. Even though these values could not be considered highly optimistic, they were still significantly more positive in anticipating the successful fulfillment of their political values than the average leader (P-2= 0.12). These scores are consistent with their high P-1 score because the governments who see the political universe as friendly also tend to believe in the likelihood of realizing their political aspirations and view conflict on the international stage as temporary. Another index that shows statistical significance is the predictability of the political future (P-3). The P-3 values for the U.S. (P-3= 0.17), Iran (P-3= 0.20), and the average leader (P-3= 0.12) are all low. However, it seems that, at least to the U.S.'s leaders, the future was significantly more predictable than to the average leader. The P-3 score also tells us whether one has a stronger tendency to attribute cooperative 33 Table 3. The Operational Codes of the U.S. and Iranian Leaders in 1953-1978 United States (n=15) Norming Group (n=168) Iran (n=11) Philosophical beliefs P-l Nature of political universe (conflict/cooperation) 0.454** 0.273 0.522** P-2 Realization of political values (pessimism/optimism) 0.254* 0.123 0.292* P-3 Political future (unpredictable/predictable) 0.169* 0.125 0.195 0.266 0.207 0.279* 0.950* 0.974 0.945* 0.731*** 0.346 0.766*** 0.390** 0.139 0.389*** 0.380 0.272 0.383** 0.268*** 0.599 0.233*** 0.488 0.492 0.530 a. Reward 0.256 0.145 0.231* b. Promise 0.098 0.065 0.035 c. Appeal/support 0.513 0.463 0.615*** d. Oppose/resist 0.047*** 0.166 0.063* e. Threaten 0.007*** 0.038 0.010* 0.082 0.123 0.044** Historical development (low control/high control) Role of chance (small role/large P-5 role) Instrumental beliefs P-4 I-1 I-2 I-3 I-4 Strategic approach to goals (conflict/cooperation) Intensity of tactics (conflict/cooperation) Risk orientation (averse/acceptant) Timing of action a. conflict/cooperation b. words/deed I-5 Utility of means f. Punish Significant differences are at the following levels (two-tailed test): *p 0.05, **p 0.01, ***p< 0.001. 34 or conflictual actions to the other. The leaders who view the political future as more unpredictable will likely link a variety of conflictual and cooperative actions to their opponents instead of designating them as one or the other. As for the control over historical development (P-4), Iran perceived their role in constructing the political environment (P-4a= 0.28) as relatively unimportant. Similarly, during that time, the United States did not believe that they possessed a mastery over foreign policy events (P-4a= 0.27). Still, Iran's view of their historical control was significantly higher than that of the average leader (P-4a= 0.21). These indices also demonstrate that since both countries did not see themselves as essential players in shaping history, they believed that the other was in control. Iran (P-4b= 0.72) and the United States (P-4b= 0.73) were convinced that between 1953-1978, the other held a relatively high locus of control over the historical development. Regarding the role of "chance" in political affairs (P-5), interestingly enough, while the average leader (P-5= 0.974) scored very high, the result values for Iran (P-5= 0.95) and the U.S. (P-5= 0.95) were significantly lower. Nevertheless, both of these countries saw the role of chance as relatively meaningful. These values are closely correlated with the P-3 and P-4 indices. Even though both leaders assigned the mix of cooperative and conflictual acts to the other and believed that their level of control over political consequences and historical developments was low, they did not view the political game as unusually unpredictable. As a result, the values for their role of chance index were expected to be lower than the obtained outcome. Turning the attention to the instrumental questions in the bottom half of Table 3 offers an insight into the leaders' prescriptive beliefs, specifying their strategic and tactical tendencies. According to the U.S. leaders of 1953-1978, the best approach to political objectives derived from a strongly cooperative strategy (I-1= 0.73). At the time, the Iranian leadership shared this viewpoint, as the direction of their strategy (I-1= 0.77) leaned towards extremely cooperative and harmonious. Both of these scores are significantly higher than those of the norming group, who favors a substantially more conflictual (I-1= 0.35) method of seeking 35 a political aim. That also indicates that the American and Iranian leaders would have attributed the cooperative actions to self, more often than not. Following the direction of strategy, the I-2 index illustrates the intensity of tactics each country deemed necessary for archiving its political aspirations. Here, the tactics of the United States (I-2= 0.39) and Iran (I-2= 0.39) proved to be significantly more focused on cooperation than those of the average leader (I-2= 0.14). As for the I-3-risk orientation (I-3), that is, the control, acceptance, and calculation of risks, the norming group is rather averse to taking risks (I-3= 0.27). The United States (I-3= 0.38) and Iranian (I-3= 0.383) governments appear to have been somewhat more accepting of taking risks, however in most cases, they would have still chosen to avoid them if they can. While the difference between the U.S. and the two other groups' scores does not demonstrate statistical significance, the Iranian government's risk acceptance was significantly higher than that of the average leader. The next score reveals the flexibility of tactics in shifting between words vs. deeds (I-4b) and between cooperative vs. conflictual moves (I-4a). Here, both Iran (I-4a= 0.23) and the United States (I-4b= 0.27) displayed significantly lower flexibility than the average leader (I-4a= 0.60) when it comes to shifting between cooperative and conflictual tactics. For the U.S. and Iran, this means that whatever their strategic approach to goals (I-1) was-cooperative or conflictual- it was likely to remain their preferred strategy, no matter the tactical approach of others. For the values of I-4b, there were no statistical differences observed between the three groups. The general tendency among leaders who are averse to taking political risks (low I-3 score) is the higher flexibility when it comes to words and deeds. That is reflected in the scores of the three groups, who, while they were relatively more risk-averse, achieved higher I-4b scores. Thus, the U.S. (I-4b= 0.49), Iran (I-4b= 0.53) and the average leader (I-4b= 0.49) were rather flexible in regard to words and deeds. Concerning the utility of means in exercising control, Table 3 shows that during the discussed period, the United States (I-5a= 0.26) and Iran (I-5a= 0.23) frequently utilized Reward as the means of control, with Iran using it significantly more often than the average leader (I-5a= 0.15). Each group also 36 seems to apply the means of Appeal often; however, again, Iran (I-5c= 0.62) was the only one displaying a significant difference from the average leader (I-5c= 0.46). All of the groups also rarely used Promise. Between the three of them, Iran used it the least (I-5b= 0.04), and the United States used it the most (I-5b= 0.10). As for the Oppose strategy, it was employed significantly less by Washington (I-5d= 0.05) and Tehran (I-5d= 0.06) than by the average leader (I-5d= 0.17). Similarly, Threat was used significantly less by Iran (I-5e= 0.01) and even more rarely by the U.S. (I-5e= 0.007) in comparison to the average leader (I5e= 0.04). Finally, the Punish means were at times used by the norming group (I-5f= 0.12). The United States, on the other hand, rarely made it their strategy of choice (I-5f= 0.08), while Iran has used it significantly less often (I-5f= 0.04) than average. The results of the statistical significance tests reported in Table 3 confirm that between the years of 1953 and 1978, the two countries displayed an exceptionally cooperative strategic and tactical orientation in pursuing their foreign policies. Additionally, they expressed a very optimistic approach to realizing their political values and a statistically more harmonious view of political life than the average leader. This evidence generally supports hypothesis 1, meaning that between 1953 and 1978, United States and Iran rightfully assumed the good intentions and cooperative attitude of their counterpart, which were confirmed by the counterpart's actual beliefs. The two outcomes that somewhat varied from what has been anticipated were the scores for P-5 and I-5. Although hypothesis 1 prescribes the regular utilization of all three cooperative means (Reward, Promise, and Appeal), the I-5 values suggest that Iran have had frequently resorted to the use of the two of them (Reward and Appeal) while rarely using Promise. On the other hand, the U.S. had not utilized those means any more commonly than the average leader. As for the negative means such as Oppose, Threaten, and Punish, both countries had employed them remarkably less than the average leader. The majority of the philosophical and instrumental indices see eye to eye, show definite proof for the considerable level of trust and a conviction about the friendly motives of one another. However, the P5 scores being relatively high indicate that both actors believed that "chance" plays a prominent role in the 37 political world. Chance influencing political events means that the future is quite uncertain, and the actions of others are unpredictable. However, that should not be the case if one is positive that the other's intentions are peaceful. That may indicate that despite the amicable relationship, both nations had their doubts about another's true aspirations. 4.3 Phase II: 1979-1997 There is no doubt that throughout 1979-1997, the relationship between the actors underwent a negative change, fundamentally due to the aftermath of the 1979 Iranian hostage crisis. During that time, many U.S. officials would have described Iranian leaders as highly hostile and irrational lunatics. If, in fact, the U.S. was right, the outcomes in Table 4 would reflect low Iranian scores for I-1 and I-2, suggesting strongly conflictual strategy and tactics. Following this notion, Iran is also expected to resort to negative means in advancing their interests, resulting in high values for I-5 d, e, and f. Iran's I-3 scores would further increase, assuming that the American decision-makers were accurately convinced about the increasing unpredictability of the Iranian "religious fanatics" whose political actions were driven strictly by ideology. These beliefs give a basis for the formulation of the second hypothesis: Hypothesis 2a: From 1979 to 1997, the Iranian approach to foreign policy had turned very aggressive and radical; therefore, their scores for I-5 d, e, and f had risen, while their I-1, I-2, I-3 values had considerably decreased. After 1979, a notable change had also occurred in Tehran's view of the American leaders. The qualitative literature indicates that during the second phase, the Iranian officials were convinced that the U.S. government was determined to isolate their country internationally, ultimately seeking the regime change within Iran (Ali 2008; Sariolghalam 2003). If correct, these assumptions would produce similar 38 outcomes to the assumptions made by the U.S. about Iran. If the U.S. did have malicious intentions toward Iran, their I-1, I-2, and I-5 d, e, and f scores would demonstrate a conflictual orientation. Despite what the U.S. believed about Iran, the Iranian leaders did not necessarily view the Americans as irrational actors; hence we should not expect a change in the U.S.'s I-3 scores. Consequently, another hypothesis derived from the Iranian view of the U.S. follows: Hypothesis 2b: The second period marks a striking drop of the U.S. I-1, I-2 scores due to their desire to pursue a regime change in Iran, as well as increased use of conflictual means (I-5 d, e, f) in their exercise of power. Table 4 shows that during the second phase, neither Iranian (I-1= 0.16) nor U.S.'s (I-1= 0.26) strategic approaches to foreign policy were more conflictual than that of the average leader (I-1= 0.35). Therefore, both nations' views about their opponent were incorrect. Iran failed to predict the intensity of the U.S.'s tactics between 1979 and 1997, which did not turn out to be remarkably hostile (I-2= 0.11). Even though Iran's preferred tactics seem to combine the cooperative and conflictual methods (I-2= 0.02), they do not exhibit significantly more aggression than those of the norming group (I-2= 0.14), proving the inaccuracy of the U.S.'s assumptions on this matter. Another evidence of the American misrepresentation of the Iranian nature is reflected in Iran's aversion to risky behavior (I-3= 0.14), showing that instead of being radical and blindly following a religious ideology as the U.S. had thought, Iranian leaders were rather pragmatic and rational. Looking at the I-5 scores further highlights the flaws in the countries' beliefs about each other, as the frequency of their use of negative means of control is comparable to the average. Iranian assumptions about the high likelihood of the U.S. resorting to the measure of threat are particularly inaccurate. The data indicates the U.S. use of Threat (I-5e= 0.02) to be significantly less common than that of the average leader (I-5e= 0.04). Based on VICS analysis results and the difference of means tests discussed above, we confidently disprove both hypotheses. 39 Table 4. The Operational Codes of the U.S. and Iranian Leaders in 1979-1997 United States (n=18) Norming Group (n=168) Iran (n=14) Philosophical beliefs P-l Nature of political universe (conflict/cooperation) 0.109* 0.273 (-)0.037*** P-2 Realization of political values (pessimism/optimism) 0.004* 0.123 (-)0.109*** P-3 Political future (unpredictable/predictable) 0.141 0.125 0.123 0.228 0.207 0.154* 0.968 0.974 0.982 0.260 0.346 0.162 0.108 0.139 0.022 0.242 0.272 0.210 0.588 0.599 0.730 0.647** 0.492 0.479 a. Reward 0.186 0.145 0.134 b. Promise 0.058 0.065 0.033* c. Appeal/support 0.386 0.463 0.414 d. Oppose/resist 0.181 0.166 0.181 e. Threaten 0.015* 0.038 0.078 f. Punish 0.174 0.123 0.161 Historical development (low control/high control) Role of chance (small role/large P-5 role) Instrumental beliefs P-4 I-1 I-2 I-3 I-4 Strategic approach to goals (conflict/cooperation) Intensity of tactics (conflict/cooperation) Risk orientation (averse/acceptant) Timing of action a. conflict/cooperation b. words/deed I-5 Utility of means Significant differences are at the following levels (two-tailed test): *p 0.05, **p 0.01, ***p< 0.001. Besides the discussed indices that directly relevant to the claims made in the hypotheses, statistically significant differences were also reported across other categories. Table 4 reveals a considerable shift in both countries' beliefs about the nature of the political universe compared to the first phase. While between 1953 and 1979, they viewed the political environment as very friendly, during the second phase, 40 the United States (P-1= 0.11) and Iran (P-1= -0.04) saw it as mixed and even hostile. Compared to the norming group, both countries perceived the nature of the political universe to be significantly more conflictual than the average leader (P-1= 0.27). The level of optimism about the realization of political goals has also dropped since the previous phase. During 1979-1997, the United States (P-2= 0.004) and Iran (P-2= -0.11) became significantly more pessimistic about achieving their political values than the average leader (P-2= 0.12). As for Iran specifically, they attributed a statistically higher locus of control over historical development to the other (P-4b= 0.85) than the remaining two groups. Furthermore, the relatively low value of I-5b suggests that Iran (I-5b= 0.03) used the means of Promise to exercise power notably less frequently than the average leader (I-5b= 0.07). Lastly, the general tendency of the U.S. leadership during the second phase was comparable to the average leader's belief system, with the exception of the I-4b index. According to the data outcomes, the U.S. (I-4b= 0.65) was considerably more likely to shift their tactics between words and deeds than both Iran (I-4b= 0.48) and the average leader (I-4b= 0.49). 4.4 Phase III: 1998-2001 According to qualitative research, the years between 1998 and 2001 represent another dramatic change in the Iranian approach. The new Iranian President, Khatami, introduced a much more peaceful narrative, seeking to re-establish the dialogue with the United States. The U.S. recognition of this new friendly approach would suggest that, in their eyes, Iran's foreign policy strategies began to alter to become more cooperative (Kamrava, 2001). Consequently, the hypothesis is formulated: Hypothesis 3: Between 1998 and 2001, the strategies of Iranian foreign policy toward the United States became progressively friendly, increasing their I-1 and I-2 scores. 41 Table 5. The Operational Codes of the U.S. and Iranian Leaders in 1998-2001 United States (n=3) Norming Group (n=168) Iran (n=3) Philosophical beliefs P-l Nature of political universe (conflict/cooperation) 0.221 0.273 0.371 P-2 Realization of political values (pessimism/optimism) 0.073 0.123 0.186 P-3 Political future (unpredictable/predictable) 0.123 0.125 0.180 0.247 0.207 0.253 0.970 0.974 0.957 Historical development (low control/high control) Role of chance (small role/large P-5 role) Instrumental beliefs P-4 I-1 Strategic approach to goals (conflict/cooperation) 0.429 0.346 0.515 I-2 Intensity of tactics (conflict/cooperation) 0.264 0.139 0.241 I-3 Risk orientation (averse/acceptant) 0.167 0.272 0.217 I-4 Timing of action a. conflict/cooperation 0.570 0.599 0.483 b. words/deed 0.747 0.492 0.667 a. Reward 0.280 0.145 0.217 b. Promise 0.047 0.065 0.040 c. Appeal/support 0.387 0.463 0.503 d. Oppose/resist 0.147 0.166 0.100 e. Threaten 0.023 0.038 0.027 f. Punish 0.113 0.123 0.120 I-5 Utility of means Significant differences are at the following levels (two-tailed test): *p 0.05, **p 0.01, ***p< 0.001. Looking at Table 4 and Table 5, we see the increase in Iran's I-1 and I-2 indices between phases II and II. Performing a difference of means test allows for a conclusion that the values for I-1 and I-2 reported for Iran in phase II have increased significantly between 1998 and 2001, proving that they did, in fact, become more friendly than they were before. Yet, Iranian I-1 and I-2 scores for phase II do not display any 42 statistical difference from the average (I-1= 0.35) (I-2= 0.14). Therefore, we cannot entirely prove hypothesis 3, even though there has been a positive change in Iran's strategic and tactical orientation. Table 5 also shows that Iran was using negative means of control less frequently during phase III than in the previous years, while their use of positive means did not change much from what it was in the second phase. The U.S.'s P-1 values show that they still saw Iran as rather unfriendly since there is no observed statistical difference between their P-1 scores from phases II and III. However, this does not indicate that their belief about Iran's hostility was significantly stronger than the average leader's belief about other nations' hostility in the political universe. The quantitative results for the I-5 index also demonstrate that in the third phase, the U.S. was not applying the means of Oppose and Threaten more often than the average leader would, which generally displays a more friendly approach. Overall, between the years of 1998 and 2001, U.S.'s and Iranian scores for all the VICS indices proved to be neither statistically different from one another nor the norming group. 4.5 Phase IV: 2002-2016 This phase is defined by the aftermath of the 2002 George W. Bush's State of the Union speech, in which he stigmatizes a group of nations, including Iran, as the "axis of evil". The President's statement has only added to the already deep-seated belief about Iranian hostility among the American decision-makers. Additionally, as the new Iranian President- Ahmadinejad-took office in 2005, his strong claims rejecting the legitimacy of the Holocaust and proposing the removal of Israel from the map only caused a further decrease of Iranian popularity among the Americans. Soon enough, the U.S. government and the majority of the U.S. public formed the opinion of Ahmadinejad as a lunatic and an irrational madman. At that time, Iranian support for terrorism and their possession of weapons of mass destruction became a widely-spread 43 conviction among Americans; therefore, to them, Iran under Ahmadinejad was a rogue state, posing a serious threat to the United States and their allies. During that time then Iranian leaders had become increasingly frustrated as a result of U.S. arrogance. The U.S. attempts to deny the Iranian nation the right to autonomy in the form of the independent civilian nuclear program were yet another manifestation of the American hostility toward Iran. Qualitative studies also suggest that both countries viewed each other as far more conflict-oriented than the average leader. Some would argue that if a group of independent observers was to decide which country's strategic direction is more antagonistic than another's, many would have pointed to Iran. That would not have come as a surprise considering the intensity of Ahmadinejad's hostile narrative. From this overview, the following hypotheses are derived: Hypothesis 4a: According to the U.S. beliefs, phase IV shows a rise in Iranian hostility and aggression (low I-1 and I-2 scores, and the increased I-5 d, e, and f values). Their antagonistic position eliminates any tendencies for shifting between conflict and cooperation, as their strategic orientation stays confrontational at all times (low I-4a score). Hypothesis 4b: In the eyes of Iranians, the United States pursues their adversary foreign policy while frequently applying the means of threat and Punishment against Tehran (high I-5e and f scores). Furthermore, the U.S. arrogant nature manifests in their attribution of a high level of political control to self (larger P-4a value). Starting with an overall direction of strategy, Table 6 shows that the U.S. leaders (I-1= 0.04) held a notably more conflictual strategic approach than the leaders of Iran (I-1= 0.35) or the average (I-1= 0.35). Contrary to the other two groups, the U.S. (I-2= -0.07) also favored significantly more conflictual tactics than Iran (I-2= 0.13) or the norming group (I-2= 0.14). We can therefore conclude that the Iranian assumptions about the U.S.'s strategic and tactical orientations were correct. On the other hand, the 44 Table 6. The Operational Codes of the U.S. and Iranian Leaders in 2002-2016 United States (n=9) Norming Group (n=168) Iran (n=9) Philosophical beliefs P-l Nature of political universe (conflict/cooperation) 0.073* 0.273 (-)0.003* P-2 Realization of political values (pessimism/optimism) 0.016* 0.123 (-)0.114** P-3 Political future (unpredictable/predictable) 0.094 0.125 0.144 0.236 0.207 0.183 0.979 0.974 0.971 0.037* 0.346 0.351 (-)0.068* 0.139 0.128 0.206 0.272 0.282 a. conflict/cooperation 0.771* 0.599 0.650 b. words/deed 0.513 0.492 0.414 a. Reward 0.097 0.145 0.119 b. Promise 0.073 0.065 0.040 c. Appeal/support 0.344* 0.463 0.514 d. Oppose/resist 0.193 0.166 0.184 e. Threaten 0.062 0.038 0.031 f. Punish 0.226 0.123 0.110 Historical development (low control/high control) Role of chance (small role/large P-5 role) Instrumental beliefs P-4 I-1 I-2 I-3 I-4 I-5 Strategic approach to goals (conflict/cooperation) Intensity of tactics (conflict/cooperation) Risk orientation (averse/acceptant) Timing of action Utility of means Significant differences are at the following levels (two-tailed test): *p 0.05, **p 0.01, ***p< 0.001. American leaders' beliefs about Iranian hostility and aggression were contradicted by Iran's cooperative values for I-1 and I-2 indices. The U.S. has also falsely assumed that Iran's dominant strategy would remain confrontational, as the I-4a score for Iran suggests the average propensity for shifting between conflict and cooperation. 45 While among the three groups, the United Stated achieved the highest score for self's control over historical development (P-4a= 0.24), it was not enough to demonstrate a statistically significant difference from the scores of Iran (P-4a= 0.18) and the norming group (P-4a= 0.21). Hence, the Iranian claim proved to be a misrepresentation of the actual U.S. beliefs. As for the utility of negative means in the exercise of power, the anticipations of both countries fell short, as neither U.S. nor Iranian leaders had resorted to the use of the Oppose, Threaten, or Punishment means. Nevertheless, what strengthens the indication that the United States was predominately conflictoriented, is their infrequent use of the means of Appeal. The U.S.'s score for the I-5c index is remarkably lower (0.34) than that of Iran (I-5c= 0.51) or the average leader (I-5c= 0.46). The above analysis shows that the U.S. view of the Iranian leadership's beliefs between 2002 and 2016 had been entirely contrary to reality. More so, this view appears to have been guided by prejudice and the unquestioned generalizations of Iranian beliefs and behavior. Consequently, hypothesis 4a must be disputed. Looking at hypothesis 4b, we can determine that the Iranian leaders were wrongfully convinced about the U.S. attribution of a high level of political control to self, as well as their preference of means of thereat and Punishment in exercising power. However, hypothesis 4b cannot be altogether rejected, as the Iranian assumptions about the conflictual nature of the U.S.'s strategy and tactics have corresponded with the truth. A few of the remaining VICS scores are also worth noting. Within the course of the fourth phase, Iran (P-1= -0.003) and the U.S. (P-1= 0.07) have again drifted away from the beliefs of the average leader about the nature of the political universe, beginning to see it as significantly more hostile than the average leader would (P-1= 0.27). The considerably lower U.S.'s P-1 score may be the evidence for their belief 46 about Iran's hostility increased under Ahmadinejad. Similarly, Iran's decreasing P-1 value explains why they continued to view the U.S. as the "Great Satan". The norming group (P-2= 0.12) is also significantly more optimistic about achieving their political aims than either the United States (P-2= 0.02) or Iran (P-2= -0.11) within this period. Also, the U.S.'s score for the I-4a index was considerably higher than that of two other groups, suggesting more tactical flexibility. That means that even though their overall approach was conflictual, the United States leaders were more prone (I-4a= 0.77) than others to switch between the conflictual and cooperative tactics; therefore, the state of conflict was not necessarily permanent. 4.6 Phase V: 2017-2020 Over the years of 2017-2020, the United States continued to perceive Iran as hostile and aggressive. Based on President Trump's stance regarding Iran and Muslims in general, it is plausible to assume that Iranian leaders viewed the United States as their enemy. Additionally, the events of the 2020 assassination of Iranian general Soleimani had confirmed the extent of the U.S.'s belligerent intentions in the eyes of many Iranians. Therefore, the final hypothesis follows: Hypothesis 5: Between the years 2017 and 2020, both the U.S. and Iranian leaders expected the opponent's behavior to be notably conflictual. Moreover, both were convinced that the other would be a lot more likely to implement the negative rather than positive means in exercising control. Looking at the I-1 and I-2 indices of both countries, we observe that here again, their beliefs about each other prove to be a misrepresentation. These scores argue against the claims made in hypothesis 5, as neither the tactics (I-2= -0.15) nor strategies (I-1= -0.14) of the Iranian leaders are highly confrontational. Even though Iran's I-1 and I-2 scores lean toward the direction of conflict, they were not significantly less 47 Table 7. The Operational Codes of the U.S. and Iranian Leaders in 2017-2020 United States (n=7) Norming Group (n=168) Iran (n=6) Philosophical beliefs P-l Nature of political universe (conflict/cooperation) (-)0.123** 0.273 0.043* P-2 Realization of political values (pessimism/optimism) (-)0.160* 0.123 (-) 0.060*** P-3 Political future (unpredictable/predictable) 0.114 0.125 0.133 0.241* 0.207 0.150* 0.974 0.974 0.980 0.074 0.346 (-) 0.138 (-) 0.001 0.139 (-) 0.147 0.110*** 0.272 0.265 a. conflict/cooperation 0.710 0.599 0.450 b. words/deed 0.624 0.492 0.548 a. Reward 0.133 0.145 0.080 b. Promise 0.103 0.065 0.000*** 0.304** 0.463 0.352 d. Oppose/resist 0.200 0.166 0.302 e. Threaten 0.083 0.038 0.072 f. Punish 0.179 0.123 0.193 Historical development (low control/high control) Role of chance (small role/large P-5 role) Instrumental beliefs P-4 I-1 I-2 I-3 I-4 I-5 Strategic approach to goals (conflict/cooperation) Intensity of tactics (conflict/cooperation) Risk orientation (averse/acceptant) Timing of action Utility of means c. Appeal/support Significant differences are at the following levels (two-tailed test): *p 0.05, **p 0.01, ***p< 0.001. cooperative than in the previous years. Equally, the U.S.s strategic (I-1= 0.07) and tactical (I-2= -0.001) orientations, do not demonstrate a noticeably higher hostility than the average (I-1= 0.35) (I-2= 0.14). As for the exercise of power, Table 7 shows that compared to the norming group (I-5c= 0.46), the U.S. resorted to the use of Appeal significantly less frequently (I-5c= 0.46). The VICS scores for Tehran 48 also suggest a relatively rare use of the means of Threat in exercising power (I-5e= 0.07), which was comparable to the frequency the average leader would apply the means of Threat (I-5e= 0.07). Hence, the beliefs of Iran and the United States about each other stated in hypothesis 5 do not match the reality. The remaining VICS values tell us that during the fifth phase, Iran and the United States were a lot more pessimistic about the overall essence of the political stage and the realization of their political values. The average leader's belief about the political universe (P-1= 0.27) is significantly more friendly than those of the U.S. (P-1= -0.12) and Iranian (P-1= 0.04) leaders. In a political world that resembles hostility, the decision-makers usually do not anticipate a substantial likelihood of realizing their political goals. Table 7 supports this claim as the United States (P-2= -0.16) and Iran (P-2= -0.06) were highly more pessimistic about fulfilling their political values than the average (P-2= 0.12). When it comes to the level of control over historical development attributed by the actors to themselves and one another, neither the U.S. nor Iran resembled a significant dissimilarity to the norming group (P-4a= 0.21). However, their views opposed each other, as the U.S. leaders believed in having a significantly higher impact (P-4a= 0.24) on political events than the Iranians (P-4a= 0.15). The latter saw themselves as less in control over the political outcome. Nevertheless, both of their scores attributed a lower locus of control to self, meaning that they both saw the opponent as having more influence over historical development. Iran's scores for other indices are close to the average and even to some of the U.S. scores. The only significant discrepancy manifests in Iran's I-5b score, which is 0.00, revealing that in contrast to other groups, Iran would have virtually never applied the means of promise to gain control in their relationships with others. The matter on which the United States disagreed with the other groups was their risk orientation. They were significantly less accepting of risks associated with the undesired political outcome (I-3= 0.11) than both Iran (I-3= 0.27) and the average leader (I-3= 0.27). The U.S.'s I-3 value reflects the conviction about Iran's highly aggressive nature, showing a strong U.S. aversion to undesirable outcomes. Since their 49 priority was risk-avoidance, they were more likely to shift between cooperative and conflictual tactics than Iran is. That means that the U.S. would have risked submission to Iran if they behaved cooperatively to achieve a settlement. If they attempted to reach the scenario in which they dominate Iran, they would have been potentially risking deadlock as the end result. 50 CHAPTER V: STRATEGIES AND LEADERSHIP PROPENSITIES This chapter focuses on assessing the United States and Iran's foreign policy strategic orientations using the leaders' VICS values for master beliefs from the previous chapter. Accordingly, chapter 5 addresses the research questions regarding the countries' strategic preferences and the accuracy of their assumptions about the strategies of the other. The combination of I-1 and P-4a scores for each country allows for determining their preferred strategy for self. In contrast, the combination of their P-1 and P-4b scores reveals the kind of strategy they attribute to the other. The placement of these values on a Cartesian coordinate system (Figure 2) indicates each country's leadership type, derived from Holsti's revised operational code typology (Walker, 2004). Letters in parenthesis in Figure 2 stand for self (S) and other (O), while lower index numbers reflect the consecutive periods. Combined with the recommendations from the Theory of Inferences about the Preferences (TIP) (Table 2), the master belief scores reveal the desired political outcomes according to each country's beliefs. These analytical tools have allowed for the formulation of the hypotheses 6 and 7, which will be further evaluated at the end of this chapter. Hypothesis 6: Throughout the course of their relations, the U.S. policymakers predominantly believed Iran to be a rogue state, headed by irrational leaders, whose foreign policy motives were inherently belligerent and ideologically-driven. Hypothesis 7: Over the years, the Iranian officials widely viewed the United States as an aggressive superpower, pursing domination through the means of their hostile (specifically toward Iran) foreign policy. However, contrary to the U.S. leaders' beliefs about the Iranian regime, Iran considered the United States a relatively rational actor. 51 Figure 2. Cartesian Coordinate System. Letters in parenthesis stand for self (S) and other (O). Lower index numbers reflect the consecutive time periods. 1. 1953-1978: US: (S) (0.27, 0.73); (O) (0.73, 0.45); Iran: (S) (0.28, 0.77); Other (O) (0.72, 0.52. 2. 1979-1997: US: (S) (0.23, 0.26); (O) (0.73, 0.11); Iran: (S) (0.15, 0.16); (O) (0.85, -0.04). 3. 1998-2001: US: (S) (0.25, 0.43); (O) (0.75, 0.22); Iran: (S) (0.25, 0.52); (O) (0.75, 0.37). 4. 2002-2016: US: (S) (0.24, 0.04); (O) (0.76, 0.07); Iran: (S) (0.18, 0.35); (O) (0.82, -0.003). 5. 2017-2020: US: (S) (0.24, 0.07); (O) (0.76, -0.12); Iran: (S) (0.15, -0.14); (O) (0.85, 0.04). The origin of the coordinate system is the norming group mean values (P-1= 0.27, SD= 0.28; I-1= 0.35, SD= 0.34; P-4= 0.21, SD= 0.11). DEE = De-escalation, Escalation, Escalation (Deter); DDE = De-escalation, De-escalation, Escalation (Exploit); EDE = Escalation, De-escalation, Escalation (Bully); EEE = Escalation, Escalation, Escalation (Punish); EDD = Escalation, Deescalation, De-escalation (Compel); EED = Escalation, Escalation, De-escalation (Bluff); DED = De-escalation, Escalation, Deescalation (Appease); DDD = De-escalation, De-escalation, De-escalation (Reward). 52 5.1. Phase I: A Special Friendship. Actor United States Iran TYPE C TYPE C Prop.2. (+, =) Settle>Deadlock>Dominate>Submit= Assurance Prop.2. (+, =) Settle>Deadlock>Dominate>Submit= Assurance Other Type TYPE A TYPE A Other Strategic Preference Prop.2. (+, =) Settle>Deadlock>Dominate>Submit= Assurance Prop.2. (+, =) Settle>Deadlock>Dominate>Submit= Assurance Self Type Self Strategic Preference Table 8. U.S.'s and Iranian Strategies 1953-1978 The U.S. master belief scores for self are I-1= 0.73 (direction of strategy), and P-4a= 0.27 (control over historical developments for self). Respectively, these values are above (+) the average of the norming group (0.35), and within (=) one standard deviation (1 SD= 0.11) from the group's mean (0.21). These signs are consistent with TIP Proposition 2 (see Table 2), revealing that out the possible political outcomes, the U.S. preferred settlement (4), over deadlock (3), over domination (2), over submission (1). The U.S. master belief scores for other are P-1= 0.45 (nature of the political universe/image of others), and P-4b= 0.73 (locus of control for the other). P-1 value is above (+) the mean of the P-1 score of the norming group (0.27), and P-4b is within (=) one standard deviation (1 SD= 0.11) of the norming group's average (0.79) for the P-4b index. These values also meet the premises of TIP Proposition 2.; therefore, what the United States has presumed to be the preference ordering of the other, matches that of self. Furthermore, this indicates that during phase I, the United States' government can be placed in the Type C leadership category, suggesting a generally cooperative strategy and higher than average sense of historical control. Following the same logic, Washington's master beliefs about Iran during that time reveal that they saw the Iranian leadership as the Type A, with the same preference order of political outcomes as the U.S. Between 1953 and 1978, the master beliefs of the government in Tehran about self and others corresponded with those of the U.S. That means that they saw themselves as generally more cooperative- 53 oriented, determined to delay escalatory actions whenever possible. Similarly, they viewed each other as rather friendly nations, willing to compromise and negotiate before utilizing the escalatory means. As scores of both countries for Ego and Alter fit the conditions of TIP Proposition 2, the subjective games of the United States and Iran for this period would have been identical. 5.2 Phase II: Iranian Revolution and the Hostage Crisis. Actor United States Iran TYPE B TYPE DEF Prop.5. (-, =) Dominate>Settle>Deadlock>Submit= Prisoners' Dilemma Prop.5. (-, =) Dominate>Settle>Deadlock>Submit= Prisoners' Dilemma Other Type TYPE DEF TYPE B Other Strategic Preference Prop.5. (-, =) Dominate>Settle>Deadlock>Submit= Prisoners' Dilemma Prop.5. (-, =) Dominate>Settle>Deadlock>Submit= Prisoners' Dilemma Self Type Self Strategic Preference Table 9. U.S.'s and Iranian Strategies 1979-1997 We can observe that while the two actors' beliefs about self and other during this phase were the same, Ego and Altar's identities were switched around. That means the U.S. saw themselves as Type B, which corresponds to how Iran perceived them. Iranians considered themselves as the Type DEF, the same way the U.S viewed them. Hence, we observe that both countries have correctly estimated each other's strategic preferences. Type B, which both countries assigned to the U.S. leadership, is characterized by a somewhat more aggressive approach, open to the use of force when faced with the prospect of a considerable payback that is achievable with minimum risk. Type B leaders also generally look at their opponents as rational actors who can be deterred (Walker, 1986). During this phase, Iran was believed by both countries to fit the Type DEF category, suggesting that they saw the conflict as permanent, resulting from human nature, nationalism, and international anarchy 54 (Walker, 1986). According to Type DEF, it is the imbalance of power on the international stage, which is the most likely cause of war. Type DEF is generally flexible in pursuing its goals and not likely to use any extreme measures to achieve them; therefore, they would not use force unless the opponent gives them no other choice. Both actors' beliefs about self and other correspond to TIP Proposition 5. Hence, each country assumed that its rival preferred the same order of the possible political outcomes. This order is prescribed by TIP Proposition 5, meaning that both actors and their counterparts would have chosen domination (4) over settlement (3) over deadlock (2) over submission (1). 5.3 Phase III: Khatami and Dialogue Among Civilizations. Actor United States Iran TYPE C TYPE C Prop.2. (+, =) Settle>Deadlock>Dominate>Submit= Assurance Prop.2. (+, =) Settle>Deadlock>Dominate>Submit= Assurance Other Type TYPE DEF TYPE A Other Strategic Preference Prop.5. (-, =) Dominate>Settle>Deadlock>Submit= Prisoners' Dilemma Prop.2. (+, =) Settle>Deadlock>Dominate>Submit= Assurance Self Type Self Strategic Preference Table 10. U.S.'s and Iranian Strategies 1998-2001 In the third phase, Iran and the United States saw themselves as Type C, just like they did during the first phase. Iran believed that the U.S. resembled the qualities of Type A, which again corresponds to Iran's belief about the U.S. in the first phase. The U.S., on the other hand, viewed Iran as Type DEF. Between 1998 and 2001, while both counties believed in their own cooperative nature, Iran saw their opponent as comparatively friendly and open to negotiations. The U.S. held a contrasting opinion of Iran as a hostile nation, overall dissatisfied with the international balance of power. The U.S. also preferred the order of outcomes that fits TIP Proposition, while in their mind, Iran favored the order described in 55 Proposition 5. As for Iranian leadership, they were of the opinion that the U.S. would have chosen the same order of outcomes as they did (Proposition 2.). The popular at the time containment strategy would have suggested that the U.S. preferred to dominate Iran over the mutual settlement. Nevertheless, U.S. master beliefs indicate the settlement and not dominance as the U.S.' most desired outcome. 5.4 Phase IV: Axis of Evil and JCPOA. Actor United States Iran TYPE B TYPE A Prop.5. (-, =) Dominate>Settle>Deadlock>Submit= Prisoners' Dilemma Prop.2. (+, =) Settle>Deadlock>Dominate>Submit= Assurance Other Type TYPE DEF TYPE B Other Strategic Preference Prop.5. (-, =) Dominate>Settle>Deadlock>Submit= Prisoners' Dilemma Prop.5. (-, =) Dominate>Settle>Deadlock>Submit= Prisoners' Dilemma Self Type Self Strategic Preference Table 11. U.S.'s and Iranian Strategies 2002-2016 The U.S. master beliefs about self and other in the fourth phase were precisely the same as in the second phase. They viewed both themselves and their opponents as unfriendly and believed that rivals should be deterred. The U.S. also predicted the same order of preferred outcomes for Self and Other as they did in the second phase. Iran had a more positive opinion about themselves, as they placed themselves in the Type A category. Type A leaders fear the "conflict spiral" as the main danger of war, resulting from players misunderstanding each other intentions, which pushes them to take impulsive actions. Type A is, therefore, convinced that their opponents are likely to act under irrational convictions. Interestingly, Iran saw the other as Type B, which matches the U.S. opinion of themselves. Iran's preferred order of political outcomes 56 favored settlement (Proposition 2.), while they believed the U.S. would have more likely chosen to dominate Iran, rather than settle (Proposition 5.). Iran's choice of Type A for self and Type B for other shows that the U.S. assumptions about Iran were not always accurate. Iran's association with Type A demonstrates that rather than being driven by evil motives and highly aggressive, they strived for compromise and political settlement. 5.5 Phase V: President Trump and General Soleimani. Actor United States Iran TYPE B TYPE DEF Prop.5. (-, =) Dominate>Settle>Deadlock>Submit= Prisoners' Dilemma Prop.5. (-, =) Dominate>Settle>Deadlock>Submit= Prisoners' Dilemma Other Type TYPE DEF TYPE B Other Strategic Preference Prop.5. (-, =) Dominate>Settle>Deadlock>Submit= Prisoners' Dilemma Prop.5. (-, =) Dominate>Settle>Deadlock>Submit= Prisoners' Dilemma Self Type Self Strategic Preference Table 12. U.S.'s and Iranian Strategies 2017-2020 During the last phase, both nations' master beliefs were again the same as between 1979 and 1997 (second phase). Here, the actors attributed hostile characteristics to both themselves and the other. Both agreed that the U.S. saw their opponents as rational actors who could be deterred, yet, believed the conflict to be temporary. On the other hand, they maintained that Iran considered conflict as permanent, resulting from the unfair distribution of power in the international system. Comparably to the second phase, both nations favored dominance over the settlement, believing that their opponent would have chosen the same outcome. Interestingly enough, while Iran's typology predictions differed substantially during the previous years and were more accurate than those of the U.S., the final phase has brought about a significant shift, possibly indicating that Iran has adopted the U.S. point of view. Between 2017 and 2020, Iran has classified 57 themselves as Type DEF- the same type that the U.S. has assigned to them. Moreover, Iran placed the Other in the Type B category, matching the U.S. view of self. That suggests that both countries viewed the U.S. as relatively friendly. After the assassination of General Soleimani, many have anticipated an aggressive response on the part of Iran. However, based on their strategic preferences, both nations would have argued against it as they classified Iran as Type DEF. That demonstrates a relatively small chance for a forceful retaliation on the part of Iran since Type DEF strives to achieve its goals with moderate means and would use military force only in the absence of an alternative. 5.6 Testing the Hypotheses Given the above evaluation and the data discussed in chapter 4, it is now possible to determine whether these results successfully proved the hypotheses 6 and 7. To check the validity of these hypotheses, we must focus specifically on both countries' VICS scores for P-1, P-2, P-4, I-1, I-2, and their overall ranking of preferred political outcomes. Based on hypothesis 6, Iran's inherent hostility would have resulted in a strong inclination toward domination as the ending state of political interaction. A statistically significant variation between Iran's I1 and I-2 scores accumulated across all five time periods, and those of the average leader would have provided evidence for their ideologically-driven foreign policy. Further proof that Iran was a rogue state would have been displayed in an overwhelming preference for non-cooperative and escalatory strategies (Canbolat, 2014). Looking at Iran's approach to goals (I-1= 0.34) and intensity of tactics (I-2= 0.13) across the five phases, it can be concluded that they do not differ significantly from those of the norming group. Hence, Iran has not been any more belligerent in pursuing foreign policy than the average leader. During the years of 1953-2020, the Iranian regime has classified itself twice as Type C and Type DEF, and once as Type A. Types A and C subscribe to the TIP Proposition 2, which preferers the outcome of the settlement. In 58 contrast, Type DEF meets the conditions of TIP Proposition 5, making domination a most desired outcome. That suggests a combination of preference for settlement and domination, yet with a stronger tendency toward settlement and cooperation. As Type DEF generally strives to avoid the aggressive strategy, unless there is no other choice, it appears that Iran has been relatively peaceful and rational, showing misjudgment on account of the hypothesis. Furthermore, the hypothesis has been contradicted by the U.S. opinion of themselves as Type B leadership, which indicates that they believed that opponents behave rationally. Hence, we can disprove the hypothesis 6. Across the five phases, the U.S. strategic (I-1= 0.34) and tactical directions (I-2= 0.15) were alike those of the average leader; therefore, the U.S. has not been overly hostile, as perceived by the Iranian regime. The U.S. leaders' attitude toward the realization of political values (P-2= 0.06) also seemed moderate. However, Washington believed that they possessed significantly more control over the political events (P-4=0. 24) than the average leader, which justifies the Iranian conviction about the American arrogance and their illegitimate interference in the world politics. Throughout the years, on two occasions, the U.S. leaders have classified themselves as Type C in favor of settlement as the final state. However, they have associated themselves with Type B on three other occasions, suggesting domination as their generally preferred political outcome. Over the years, Iran saw the political universe as highly conflictual (P-1= 0.16). They have upheld a significantly more pessimistic viewpoint about reaching their political goals (P-2= 0.02) than average (P2= 0.12), which was the result of the U.S. influence on the global political stage. The fact that Iranian leadership viewed themselves as Type DEF is another evidence that Iran has been troubled by the lack of international equilibrium. Their belief about the United States being the Type B confirms that they saw them as rational and predictable. Even though the U.S. foreign policy did not seem to be driven by conflictual strategy and tactics, as it has been argued, hypothesis 7 has accurately predicted that the U.S. had a stronger desire to dominate Iran than to reach a state of settlement. Moreover, the Iranian opinion about the U.S. leaders as rational actors proved accurate, therefore providing enough evidence to support hypothesis 7. 59 CHAPTER VI: CONCLUSION For the past sixty-seven years, the relationship between the United States and Iran and the strategic motivations behind their leaders' foreign policy agenda have been, more often than not, unclear, and therefore continuously misread. As the examined studies observe, the quality of historical interactions between the two nations generated a perceived crisis state. Based on the theoretical assumptions, while the American policymakers view themselves as democratic and peace-seeking leadership that provides security to different regions across the world, the Iranians have quite a different opinion. Similarly, while the Iranian government believes that they are a rational and cooperative nation, yet a victim of the hostile American attempts to further their international isolation, Washington begs to differ. History has shown that misinterpretation of one's political opponent's behavior and a flawed comprehension of their strategic motivations can have disastrous consequences. The severity of these consequences has been perhaps best explained by Knorr (1976): The greatest dangers to realistic threat perception do not inhere in the intellectual difficulties resulting from poor evidence and future uncertainty. The greater danger lies in rigid preconceptions and attitudes of which the perceiver is unaware, or not aware enough. Such preconceptions make him desire to see certain things happen, and to make what he wants to do seem justified. And the push and pull of emotions that are attached also to foreign actors in such forms as hatred and contempt can lead the perceiver astray. These intervening preconceptions and attitudes produce selectivity in the receipt and use of information; they therefore contribute to a distorted image of reality and to false expectations (p. 85). 60 Thus, this study has undertaken the task of providing the methodological means "to counter a dangerous and devastating cognitive bias, namely the fundamental attribution error" (Malici and Walker 2008, p. 786) in foreign policy analysis. 6.1 Efficiency of the Analysis Method This thesis supplements foreign policy analysis research by applying the operational code methodology to explain state behavior and expand the understanding of how one regime's political beliefs and strategies refer to the average leader and other actors on the international stage. More importantly, it counters the over-simplified assumptions, which Hart (2005) argues, "significantly affect policymaking and unnecessarily close off policy options" (p. 9). Though beyond this work scope, this approach further allows scholars to predict leaders' potential moves and their responses to specified future circumstances. By utilizing the individual-oriented psychological approach, this study helps to develop an understanding of policymakers' beliefs and, therefore, a more accurate interpretation of the effects of their foreign policy decisions. Additionally, by comparing the quantitative data and testing the conventional wisdom and often overly-generalized claims about the U.S.-Iranian relations, this paper contributes to the descriptive literature on this subject. 6.2 Making Sense of the Outcomes The results of this analysis exhibit quite a few unexpected findings, suggesting the applicability of this methodology. One of the more surprising results has been the United States' attribution of leadership types, classifying themselves as Type B, and assigning the Iranian leadership to the Type DEF category. The associating of Iran with Type DEF leadership qualities suggests that, according to the U.S., Iran believes that conflict is permanent. That follows the notion that the U.S. sees Iran in a very negative light. However, the U.S. choice of Type B for self goes against the common belief that the U.S. views the 61 regime in Tehran as irrational religious fanatics because the leaders of Type B presume their opponents to behave rationally. Tehran makes a correct prediction about the overall strategic orientation of Washington, as their assumption about the American leadership matches the U.S. vision of self as Type B. Iran also successfully anticipates that domination is the U.S.'s most sought-after political outcome. Washington's predictions about Iran displaying the characteristics of Type DEF is somewhat accurate since Tehran classifies themselves as both Type DEF and Type C. Nevertheless, the United States fails to accurately assess the Iranian order of preference for political outcomes, as they are convinced that Iran is seeking dominance. On the contrary, the empirical data shows that Iran would instead be more inclined to choose settlement or deadlock as the final outcome. That explicitly suggests that Iran is not a potential threat to American national security, as it is often commonly believed. The study's overall results prove that widespread assumptions about the Iranian regime as irrational and hostile and the U.S. as the beacon of global harmony and peace do not correspond with the reality of their relations. This thesis demonstrates that while the Iranian leaders acknowledge that a good part of their national identity has been shaped by ideology and religion, they see themselves as progressively pragmatic in their interaction with the outside world. On the other hand, the operational code analysis suggests that American foreign policy is strongly influenced by the desire to inflict their domination on the Iranian government. 6.3 Final Recommendations Within the broader context, it is vital for policymakers to continuously strive for a more comprehensive foreign policy analysis method, which grants a higher degree of accuracy in assessing the nature of foreign relations. The unaddressed or ignored misperceptions of the relations between states, over time, result in effort and resources wasted for pursuing irrelevant foreign policy paths, which instead could have been used toward the policy areas that need improvement. The misperceptions embedded in a country's 62 international rhetoric not only become an unwanted distraction from actual issues but also are likely to lead to an unnecessary escalation of the conflict. Consequently, this work argues that we, as human beings, tend to fear what we do not understand, for the very reason that we do not understand it, and not because it represents a legitimate threat (at least not a threat that has been rationally assessed as such). That is precisely the case surrounding the foreign policy of both the United States and Iran. Their relationship is characterized by a sense of insecurity and mutual suspicion resulting from the lack of in-depth knowledge about the motives and beliefs of one another. Therefore, if each leader develops an understanding of how the opponent perceives them and how they see themselves, the historically antagonistic relationship will potentially evolve into a more informed and constructive interaction. 63 SELECTED BIBLIOGRAPHY Abdullah, M. H., Ani, F., Othman, Z. & Rahman, R. (2020). Nuclear Issues in the Iran - United States Relations: From Friends to Foes. Tamkang Journal of International Affairs. doi:10.6185/TJIA.V.202001_23(3).0001 Ali, L. A. (2008). Historic U.S. Iran Relations: Revisiting Ideology and Geostrategy. Pakistan Journal of American Studies, 26. Asmar, M. (2020). US-Iranian Relations After the Assassination of Suliemani. Middle Eastern Studies Journal, (91), 173178. Ayoob, M. (2011). American Policy Toward the Persian Gulf: Strategies, Effectiveness, and Consequences. Middle East Policy, 18(2), 110119. doi: 10.1111/j.1475-4967.2011. 00489.x Brams, S. J. (1994). Game Theory and Literature. Games and Economic Behavior, 6(1), 3254. doi: 10.1006/game.1994.1002 CAMPBELL, D. (1993) Cold Wars: Securing Identity, Identifying Danger. Rhetorical Republic: Governing Representations in American Politics, edited by F. M. Dolan and T. L. Dumm, 39-60. Amherst: University of Massachusetts Press. Crichlow, S. (2006). The Eyes of Kesteven: How the Worldviews of Margaret Thatcher and Her Cabinet Influenced British Foreign Policy. Beliefs and Leadership in World Politics, 77-99. doi:10.1057/9781403983497_4 64 Feng, H. (2006). Crisis Deferred: An Operational Code Analysis of Chinese Leaders Across the Strait. Beliefs and Leadership in World Politics,151-170. doi:10.1057/9781403983497_7 George, A. (1969). The "Operational Code" A Neglected Approach to the Study of Political Leaders and Decision-Making. International Studies Quarterly, 13(2), 190-222. doi:10.2307/3013944 Grove, A. K. (2015). Framing identity in Iran: Culture, conflict and the West. International Politics, 52(5), 549-566. doi:10.1057/ip.2015.25 Hart, J.-A. (2005). Perceptions and Courses of Actions toward Iran. Retrieved from https://www.armyupress.army.mil/Portals/7/Hot%20Spots/Documents/SA-Iran/Hart-Iran2005.pdf Holsti, O. (1970). The "Operational Code" Approach to the Study of Political Leaders: John Foster Dulles' Philosophical and Instrumental Beliefs. Canadian Journal of Political Science / Revue Canadienne De Science Politique, 3(1), 123-157. Retrieved from http://www.jstor.org/stable/3231421 Holsti, O. R. (1977). Foreign Policy Decision-Makers Viewed Psychologically: Cognitive Processes Approaches. Thought and Action in Foreign Policy, 1074. doi: 10.1007/978-3-0348-5872-4_2 Hurd, E. S. (2008). U.S. and Iran. Contesting Secularism. doi: 10.4324/9781315574035-9 Kamrava, M. (2011). International politics of the Persian Gulf. Syracuse, NY: Syracuse University Press. 65 Knorr, K. (1976). Threat Perceptions. In Historical dimensions of national security problems (p. 85). Lawrence, KS: Allen Press. Krauthammer, C. (2005, December 16). In Iran, Arming for Armageddon by Charles Krauthammer. Retrieved October 03, 2020, from https://www.realclearpolitics.com/Commentary/com12_16_05_CK.html Kupchan & Clifford. (2005). Iranian Beliefs and Realities. National Interest, 81, 106110. Lavikainen, J. J. (2016). The Operational Code of Vladimir Putin in the Ukrainian. Retrieved October 14, 2020, from https://helda.helsinki.fi/bitstream/handle/10138/163809/Lavikainen_poliittinen%20historia.pdf Leites, N. (1951). The Operational Code of the Politburo. New York: McGraw-Hill. Leites, N. (1953). A Study of Bolshevism. Free Press. New York. Leverett, F. L. (2005). Inheriting Syria: Bashar's trial by fire. Washington, D.C.: Brookings Institution. Levy, J. S. (1994). Learning and foreign policy: Sweeping a conceptual minefield. International Organization, 48(2), 279-312. doi:10.1017/s0020818300028198 Malici, A., & Malici, J. (2005). The Operational Codes of Fidel Castro and Kim Il Sung: The Last Cold Warriors? Political Psychology, 26(3), 387-412. doi:10.1111/j.1467-9221.2005. 00423.x Malici, A. (2006). Reagan and Gorbachev: Altercasting at the End of the Cold War. Beliefs and 66 Leadership in World Politics, 127149. doi: 10.1057/9781403983497_6 Malici, A., & Buckner, A. L. (2008). Empathizing with Rogue Leaders: Mahmoud Ahmadinejad and Bashar al-Asad. Journal of Peace Research, 45(6), 783800. doi: 10.1177/0022343308096156 Malici, A., & Walker, S. G. (2017). Role theory and role conflict in U.S.-Iran relations: enemies of our own making. New York, NY: Routledge, Taylor & Francis Group. Maloney, S. (2002). Identity and Change in Iran's Foreign Policy. Identity and foreign policy in the Middle East. Ithaca, NY: Cornell University Press. Marfleet, B. G., & Walker, S. G. (2006). A World of Beliefs: Modeling Interactions Among Agents with Different Operational Codes. Beliefs and Leadership in World Politics, 53-73. doi:10.1057/9781403983497_3 Merton, T. R. (1940). Camouflage in Modern Warfare. Nature, 146(3700), 429429. doi: 10.1038/146429a0 Neack, L. (n.d.). The New Foreign Policy: Power Seeking in a Globalized Era. Retrieved from https://books.google.com/books/about/The_New_Foreign_Policy.html?id=7N9O-igbK_gC Nhandara, S. (2015). The Operational Code of Tony Blair - DiVA portal. Retrieved October 3, 2020, from http://www.diva-portal.org/smash/get/diva2:943474/FULLTEXT01.pdf Nick Levine and Michael D Young (2014) Leadership Trait Analysis and Threat Assessment 67 with Profiler Plus. Proceedings of ILC 2014 on 8th International Lisp Conference, Montreal, QC, Canada August 14 - 17, 2014. Association for Computing Machinery. Reisinezhad, A. (2011). U.S. Policy in Iran 1977-79. doi: 10.1007/978-3-319-89947-3_8 Richter, & Elizabeth. (2016, August 30). Operational code analysis of Iran's Supreme LeadershipAyatollahs Khomeini and Khamenei. Retrieved from http://repository.bilkent.edu.tr/handle/11693/32382 Sariolghalam, M. (2003). Understanding Iran: Getting Past Stereotypes and Mythology. The Washington Quarterly, 26(4), 6982. doi: 10.1162/016366003322387118 Schafer, M., & Walker, S. G. (2006). Beliefs and leadership in world politics methods and applications of operational code analysis. New York, NY: Palgrave Macmillan US. Walker, S. G., Malici, A., & Schafer, M. (2011). Rethinking Foreign Policy Analysis. doi: 10.4324/9780203834862 Walker, S., Schafer, M., & Young, M. (1998). Systematic Procedures for Operational Code Analysis: Measuring and Modeling Jimmy Carter's Operational Code. International Studies Quarterly, 42(1), 175-189. Retrieved from http://www.jstor.org/stable/2600822 Walker, S. G., & Schafer, M. (2007). Theodore Roosevelt and Woodrow Wilson as Cultural Icons of U.S. Foreign Policy. Political Psychology, 28(6), 747776. doi: 10.1111/j.1467-9221.2007. 00602.x Walker, S. G. (1983). The Motivational Foundations of Political Belief Systems: A Re-Analysis of the 68 Operational Code Construct. International Studies Quarterly, 27(2), 179. doi: 10.2307/2600545 Walker, S. (1990). The Evolution of Operational Code Analysis. Political Psychology, 11(2), 403-418. doi:10.2307/3791696 Walker, S. G. (2004). Role Identities and the Operational Codes of Political Leaders. In Margaret Hermann, ed. Advances in Political Psychology (pp. 71106). Amsterdam: Elsevier. zdamar, ., & Canbolat, S. (2018). Understanding New Middle Eastern Leadership: An Operational Code Approach. Political Research Quarterly, 71(1), 1931. doi: 10.1177/1065912917721744 69 APPENDICES Appendix 1. Interpreting VICS Indices (Walker et al., 2008, pp. 227-231). Source: Richter (2016) 70 Appendix 2. Calculating VICS Indices. Source: Malici and Buckner (2008) 71  ...

... Dihydroflavonol 4-reductase gene expression and color change in the protandrous flowers of Saponaria officinalis By Taylor Renee York Honors Project Thesis Submitted in Partial Fulfillment Of the Requirements for a Baccalaureate Degree With Distinction In the Ron and Laura Strain Honors College of THE UNIVERSITY OF INDIANAPOLIS May 2020 Faculty Advisor: Sandra L. Davis, Ph.D. Executive Director of Honors: James B. Williams, Ph.D. @ Taylor York, 2020. Copyright in this work rests solely with the author. Please ensure that any reproduction or re-use is done in accordance with relevant national copyright protection. i Abstract Flowers of Saponaria officinalis exhibit protandry and floral color change: in early stages of anthesis, flowers bloom white and are male, then the petals turn pink as they transition to female. The depth of pinkness is determined by genetic and environmental factors. The pink color is due to the accumulation of anthocyanidin pigments, which are created through a biochemical pathway, each step catalyzed by enzymes, whose production is controlled by genes. Variation among plants in expression of these genes may lead to differences in final color or the rate that color change occurs. If these morphological differences result in fitness variation among individuals, the differences in gene expression may be subject to natural selection. Therefore, the pattern and extent of gene expression may represent an adaptive trait. This experiment quantified Dihydroflavonol 4-reductase gene (DFR) expression using quantitative polymerase chain reaction. First, gene expression was measured in samples obtained from previously collected late-stage female flowers, all which had their color quantified. Next, DFR expression was measured in flowers collected at five different life stages from ten different plants. We found DFR expression has no significant correlation with the level of pinkness reached in the female stage flower but does significantly increase as plants change from male to female. ii Table of Contents Abstract i Table of Contents ii Acknowledgements iii List of Tables iv List of Figures v Statement of Purpose 1 Introduction 2 Methods 8 Results 11 Discussion 19 Reflection 24 References 26 Appendix 29 iii Acknowledgments I would like to thank my parents for pushing me into accepting the invitation to join the Honors College at the University of Indianapolis and encouraging me to continue each year. I would like to thank Dr. Davis for working with me continually for the past two and half years. Finally, I would like to thank the Honors College for supporting me through my undergraduate work and funding my research. iv List of Tables Table 1: Plate Layout for all qPCR Runs 10 Table 2: A Tukey HSD 17 Appendix II: An ANOVA Statistical Analysis: Location vs Pinkness 30 Appendix III: Pearson Correlation Coefficient of Pinkness 31 Appendix IV: An ANOVA Statistical Analysis: Life Stage vs DFR Expression 32 Appendix V: Tukey HSD and Tamhane Analysis 33 Appendix VI: An ANOVA Statistical Analysis: Life Stage vs DFR Expression. Does not include the Bud Stage 36 Appendix VII: Tukey HSD and Tamhane Analysis. Does not include the Bud Stage 37 Appendix VIII: A Tukey HSD. Does not include the Bud Stage 40 v List of Figures Figure 1: The Anthocyanin Pathway 5 Figure 2: The Pinkness Index Equation and Absorbance Spectrum 6 Figure 3: The Correlation Between ANS Expression and Pinkness 7 Figure 4: Pinkness Index versus location 12 Figure 5: DFR expression versus Pinkness Index 13 Figure 6: Pinkness Index of Floral Stage 15 Figure 7: DFR Expression Across Floral Stages 16 Figure 8: The average DFR expression across the floral stages 19 1 Statement of Purpose The purpose of this study was to contribute to the understanding of where in the anthocyanidin production pathway regulation of pink flower pigment production occurs. It has already been shown that the gene for anthocyanidin synthase (ANS), which acts later in the biochemical pathway, does not have a high level of expression until the flowers begin to change color, so I wanted to see if DFR, which acts earlier in the biochemical pathway, shows the same expression pattern. The significance of this experiment is that it is part of a larger research program that seeks to answer the question: is the color change seen in the flowers of S. officinalis an adaptive trait? Adaptation is the result of natural selection. If natural selection is at work, individuals that have gene variants that confer a phenotype that increase their fitness, pass those genes to the next generation at a higher rate than those that lack those variants. Over time, this theoretically weeds out most of the bad genes and leaves the good genes in the population. Color change may be an adaptive response to many things, one of which is pollinators. By changing colors, these flowers could be signaling to pollinators which flowers contain pollen (the white male flowers), and which flowers do not (the female pink flowers). This may increase visits to the male-phase flowers which would benefit due to increased pollen dispersal over visits to female-phase flowers which would not benefit as greatly since s single pollinator visit could provide a large enough pollen deposit to fertilize all the ovules (Willson, 1994). Color change may also be an adaptive response to environmental conditions. Flowers of S. officinalis turn pinker when exposed to the sun (Jabbari et al., 2012), and anthocyanins may serve as protection against damage due to UV light as they have antioxidant properties (Glover & Martin, 2012). When in the biosynthetic pathway anthocyanin production is regulated will help to determine which genes 2 may be involved in governing differences between individuals in their color and how quickly plants can respond to environmental stimuli. Using quantitative polymerase chain reaction (qPCR), I measured the expression of the gene that codes for the enzyme, Dihydroflavonol 4-reductase (DFR), an enzyme that is active in the anthocyanidin metabolic pathway. This pathway is a series of biochemical reactions that produce anthocyanins, the pigments that produce the pink color in flowers of Saponaria officinalis. To examine whether differences in how pink the flowers of individual plants become are associated with differences in DFR expression, I measured DFR expression by qPCR, then compared DFR expression to pinkness index in these flowers. In a second experiment, to see if the increase in color seen during the development of flowers was associated with an increase in DFR expression as they age, I again used qPCR to compare DFR expression across maturation stages in the Saponaria officinalis flowers. Introduction: Background Information The angiosperm Saponaria officinalis shows two distinct traits that are of interest to plant reproductive biologists: protandry and floral color change. Protandry is a reproductive trait whereby the flowers male reproductive organs, the stamens, mature and senesce before the female reproductive structures, the pistils, become receptive. This reduces the amount of selffertilization that can occur, which may be beneficial because self-fertilization decreases the genetic diversity of the offspring produced, which carries negative evolutionary consequences. Therefore, protandry is considered an adaptation that increases outcrossing (Lloyd & Webb, 1986). However, it is pointed out by Lloyd and Webb (1986) that protandry may not have evolved solely for this reason. 3 The second feature of interest in Saponaria officinalis is the floral color change. Individual flowers first bloom as white, and over a few days, the petals turn pink. This color change is due to an accumulation of anthocyanin pigments in the petals (Jabbari et al., 2012). There are some environmental factors that affect how pink the flower turns. For example, plants mostly in the sun have flowers that turn pinker than those growing in the shade (Jabbari et al., 2012). Furthermore, Davis et al. (2014) found heritability contributes to color change, as clonal lines differed significantly in the final pinkness level that flowers reach. Combined with the protandrous nature of the flowers, this means that male-phase flowers are still clearly white in all plants, but they transition to pink as they switch to female-phase. Therefore, the degree of pinkness a female-phase flower reaches depends on the genetic make-up of the parent plant which produce it and the environment in which that plant is growing. The evolutionary significance and molecular mechanisms underlying floral color change in Saponaria officinalis are still largely unknown. One possibility is that the increased anthocyanin levels are due simply to the side effects of aging (Davis & Jones, 2008; Jabbari et al., 2012), as many flowers follow this pattern of color change during senescence (Delph and Lively, 1989). Another possibility is that Saponaria growing in the shade do not turn as pink as those in the light since the Saponaria officinalis growing in the sun may be light stressed and produce anthocyanins in response (Nawrocki, 2013). Anthocyanins are also used as antioxidants to protect tissues from oxidative stress that is caused by light under stressful conditions (Glover & Martin, 2012). Yet another possibility is that floral color change has been naturally selected to attract pollinators to individual plants but provide the insects with a signal to avoid older, nonreproductive flowers within the plant. Makino and Ohashi (2016) did not work with the Saponaria officinalis but did work with other types of color-changing flowers. They 4 demonstrated that full-color changes give honest signals to pollinators with spatial memory to help with plant-level avoidance, as plants that give pollinators rewards are visited more than plants that do not give rewards. They also suggest if the plant changes color, the pollinators do not remember that the plant does not give a reward. A unique feature of Saponaria officinalis is that because the changes in color occur during its transition from male and female phases, flowers are still sexually active after the color change. This suggests that the color change is not simply due to a side effect of senescence. Researchers demonstrated that there is a difference between male- and female-phase flowers when it comes to how many pollinators visit the flowers. Pollinators clearly favored white malephase flowers over pink female-phase flowers (Nawrocki, 2013; Davis et al., 2014). In this sense, the color change may be favored by selection as a mechanism to increase the visitation of male-phase flowers over female-phase flowers because female-phase flowers do not need as many insect visits. They receive enough pollen grains to fertilize all their eggs, whereas malephase flowers will increase pollen dispersal with increased pollinator visits (Willson, 1994). Problem Statement The color change in the petals of S. officianalis is caused by the production of pigments called anthocyanins, which have a complex, multistep pathway through which they are produced (Fig 1). Each step in the biochemical pathway that produces anthocyanins is catalyzed by a different enzyme. In turn, each of these enzymes is coded for by its own gene. For example, the enzyme anthocyanidin synthase (ANS) is coded for by the anthocyanidin synthase gene (ANS). Previous work by students under Dr. Daviss supervision have looked to see if the production of this enzyme correlated to the color of the Saponaria officinalis (unpublished data). 5 Figure 1: The Anthocyanin Pathway. From: Kang et al., 2014 In one study, 53 Saponaria officinalis female-phase flowers of different plants were collected by Dr. Davis in the summer of 2015, and it was recorded where the flower ranked on a pinkness index (PI). The pinkness index was measured by using a reflectometer to record the amount of light reflected from an individual petal across the visual light spectrum. This reflectance spectrum was then used to calculate a single number to quantify how pink a flower is (Fig 2). Once the phenotype had been recorded, the RNA was extracted from the petals of each flower and reverse transcription was used to create cDNA (complementary DNA) samples. RealTime or Quantitative Polymerase Chain Reaction (qPCR) was then performed on each of the samples using primers specific for the ANS gene. This measures the concentration of messenger 6 Figure 2: The equation to calculate the pinkness index and an absorbance spectrum (Davis et. al, 2014) RNA (mRNA) for the ANS gene that was present in the original sample. Since the production mRNA is the first step in the expression of a gene, the amount of mRNA is used to quantify the level of expression of the gene. In this case, there was a significant correlation between PI and ANS expression, so that the higher the expression level of ANS in the flower, the higher they pink index for the flower (Eckert, unpublished data; Fig 3). This indicates that the differential expression of the ANS gene contributes to the final color of female-phase flowers. 7 Figure 3: The correlation between ANS expression and pinkness. Eckert, unpublished Data. In a second study, another student researcher determined if there was a significant difference between the sexual phases of the Saponaria officinalis flowers and the expression of ANS. She found that there was a significant difference in the expression of ANS in male-phase compared to female-phase flowers (Foy, unpublished data). Therefore, increases in ANS expression are also involved in the transition from white to pink as the flower ages. For my project, I explored the connection between flower color and expression of a second gene, Dihydroflavonol 4-reductase (DFR), which codes for an earlier enzyme in the anthocyanin pathway. Theoretically, DFR and ANS should both show a similar relationship to color since they are found in the same pathway. However, Farzad et al. (2003) showed another story as they explain the color-changing in the Viola. They examined three genes in the Viola cornuta, chalcone synthase (CHS), DFR, and ANS, and examined how their expression changed over the lifetime of the Viola cornuta and in plants growing in different light conditions. They found ANS showed a dramatic increase in expression as the flower aged, but CHS and DFR were 8 differentially expressed in plants growing under differing light conditions. Therefore, the color change may be regulated by different gene combinations in the same biochemical pathway due to different stimuli. For my Honors projects, I compared expression of the DFR gene to floral color change in the Saponaria officinalis. To carry out my experiment, I used the same 53 samples used in the previously described study, with the phenotypes known, and used qPCR to determine if, like ANS, DFR expression is correlated with how pink female-phase flowers become. I hypothesized that the higher the flower of the Saponaria officinalis was on the pink index, the higher the DFR gene expression should be. In the second part of my project, I used a different set of samples collected in 2016, also with known phenotypes, and again used qPCR to determine whether DFR expression differed between male- and female-phase flowers. I hypothesized that the later the phase the Saponaria officinalis was in, the higher the expression of DFR should be. My findings showed that DFR does not have a statistically significant correlation with pinkness in the late female-phase of the Saponaria officinalis. This suggests that for the S. officinalis, ANS expression contributes more to the flower color than DFR. However, while DFR does not correlate with pinkness, it does significantly increase over the lifespan of the Saponaria officinalis. Methods: Correlation of DFR expression vs color of female-phase flowers In 2015, Dr. Davis and colleagues collected 53 samples of late female-phase flowers of Saponaria officinalis from plants in three separate locations at the DePauw University Nature Park: an experimental sun-exposed garden plot, an experimental shade-exposed garden plot, and 9 natural populations found in the park (partially wooded). The reflectance spectrum of one petal was taken when the flowers were collected. The remaining petals were removed, and RNA extraction was done using the TRIzol protocol. Reverse transcription was conducted on each TM sample by a previous student to generate cDNA from each flower. The start of my experiment was to create one to ten dilutions of the 53 cDNA samples by taking five microliters of the cDNA and mixing it with 45 microliters of nuclease-free water. Of the 53 samples, five had been completely exhausted in other experiments, leaving me with a total of 48 samples. For each run of qPCR, I chose five of the 1:10 dilutions at random, until I ran all 48 samples. Then, I chose one of the undiluted samples that corresponded to one of the five dilutions to use to create a standard curve. For each round of qPCR, I created two separate master mixes, one for DFR and one for ACT. Actin, or ACT, is a highly conserved, essential protein found in plant cells, produced on a continual cycle due to its many uses in the cell (Dominquez & Holmes, 2011). Due to its constant production, for this experiment, ACT is being used as a control, to correct for pipetting errors or loss of sample due to handling such small volumes. In each master mix, I added 16 microliters of SYBR Green and 16 microliters of the corresponding eight molar ACT/DFR TM primer mix. For the standard curve dilutions, I preformed six serial dilutions, starting with two microliters of the undiluted sample and 18 microliters of water. Reagents were then loaded into a 48-well plate to run in the real-time thermocycler. The plate layout shown in Table 1 was used for all qPCR runs, both in the first part of the experiment and the second part. Each well was filled with six microliters of either ACT or DFR primer mix and six microliters of a chosen sample as detailed below, for a total of 12 microliters in each well. 10 Table 1: Plate layout for all qPCR runs. Blue wells were those filled with ACT primer mix. Pink wells were those filled with DFR primer mix. The description in each well was what sample was placed in each well. Sample # are the randomly picked 1:10 dilutions for each session. 1 2 3 4 5 6 7 8 A Water 1*10-1 1*10-4 Sample 1 Sample 2 Sample 3 Sample 4 Sample 5 B Water 1*10-2 1*10-5 Sample 1 Sample 2 Sample 3 Sample 4 Sample 5 C Water 1*10-3 1*10-6 Sample 1 Sample 2 Sample 3 Sample 4 Sample 5 D Water 1*10-1 1*10-4 Sample 1 Sample 2 Sample 3 Sample 4 Sample 5 E Water 1*10-2 1*10-5 Sample 1 Sample 2 Sample 3 Sample 4 Sample 5 F Water 1*10-3 1*10-6 Sample 1 Sample 2 Sample 3 Sample 4 Sample 5 The plate was then placed in the thermocycler which was set to run the following protocol: 40 cycles of 1-minute denaturation at 95C, 1-minute of annealing at 54C, and 2 minutes of extension at 72 C. Changes in DFR expression as the flower transitions from male to female For the second part of my project, I used samples collected by Davis and Dudle from July 19-22, 2016. They collected samples from 12 different Saponaria officinalis at the bud, early staminate, late staminate, early pistillate, mid pistillate, and late pistillate. Again, it was recorded where the flower ranked on the pink index. Once the phenotype had been recorded, the samples had their RNA extracted and quantified how much RNA was in the sample. Five mature buds were marked on the same day, so they would all open on the same day. One bud was taken, the reflection spectrum was recorded, and the petals were snap frozen in liquid nitrogen, then ground to a powder. The RNA was extracted using the TRIzol procedure. The flowers were tracked TM through their five stages, and at each stage, the reflectance was taken, as was a petal for RNA extraction. The TRIzol procedure was once again used to extract RNA. For this part of my TM 11 experiment, I started with these RNA samples instead of cDNA samples and therefore had to use reverse transcription to convert the samples to cDNA. Each reaction consisted of four microliters of RT Enzyme Master Mix (which was made with 52l of RNase free water, 52l RT buffer, 26l of AMV reverse transcriptase), 2l of oligo dT, 4l of dNTP, and a volume of the RNA solution to contain 2g of RNA and enough water to bring the total volume to 20l. The samples were placed in a thermocycler program for 40 cycles of 1-minute denaturation at 95C, 1-minute of annealing at 54C, and 2 minutes of extension at 72 C. Reverse transcription was conducted on 25 samples at a time, for 50 total samples. The resulting cDNA was then purified using a Qiagen PCR Purification Kit . I then ran qPCR on each sample using the same procedure for TM part one. To quantify gene expression, I took the average C (the cycle number that the sample T started its exponential phase) value for each sample using the DFR and ACT primers, and found the difference between them by subtracting the C value of ACT from the C value for DFR. Then, T T I calculated the delta delta value by taking the greatest difference found between the two (positive or negative), then taking the positive value of that difference and subtracting all differences by that value. Finally, I squared the final difference and that became the ratio that was used to measure DFR expression. Results Correlation of DFR expression vs color of female-phase flowers It has already been shown that the environment does have an impact on how pink the late female-phase flowers get, particularly if they have been grown in shade or grown in the sun (Jabbari et al, 2012). However, the environment the flowers were growing in was taken into 12 consideration because we wanted a wide variety of pinkness values so that we could establish whether there is a correlation between DFR expression and how pink the late pistillate flowers become. The pinkness indexes from the flowers in this experiment ranged from 0.036472605 to 0.323499179. Figure 4: Pinkness Index of female-phase S. officinalis flowers collected from plants growing in different locations. This graph shows the relationship between how deep of a pink the late pistillate phase flowers get depending on their location from forty-eight different flowers. There were 21 samples from the sun, 8 samples from the shade, and 24 samples from the nature park. While there might not be any significant difference between the groups, there does seem to be the right trend for the sun to have the highest amount of pink, the shade to have the least amount of pink, and the nature park to have an intermediate level (Fig. 4). Unlike previous studies, there was not a significant correlation between the environment the Saponaria officinalis was grown in and how pink the late phase pistillate flowers became (P=0.072, R2=0.034; 13 Appendix II). I hypothesize that the lack of correlation is due to the small sample size from the shade samples. There was no significant correlation between the expression of DFR and how pink the late pistillate-phase flowers became (P =0.373; Appendix III). There again does seem to be a small trend towards my hypothesis, shown in Figure 5, but there are many DFR expression values near 0 across all pinkness levels. This may indicate either our samples were of low quality, or our procedure was not sensitive enough to reliably make accurate measurements of DFR expression. Figure 5: The lack of correlation between DFR expression and pinkness of late phase pistillate flowers. The relationship between DFR expression and the color of the pistillate phase flowers. As shown, the pinkness of the pistillate-phase flowers does not seem to be dependent on DFR expression in the Saponaria officinalis. 14 For both environmental factors and DFR expression, P values are above 0.05, meaning that neither impact how pink the late pistillate flowers of the Saponaria officinalis get. When looking at differences among the environments, we performed a one-way ANOVA, but when looking for the relationship, or lack of, between PI and DFR expression, we did a PearsonProduct Correlation. This statistical test tells us how much of a linear relationship the two variables have, which would be the relationship we are looking for if DFR expression impacted pinkness in the Saponaria officinalis. Changes in DFR expression as the flower transitions from male to female Davis and Dudle (2016) compared pinkness of S. officinalis to flower life stage. An ANOVA test determined there was a significant difference in how pink the Saponaria officinalis flower was and its life stage it was in. Since there was a significant difference, a Tukey Post hoc test was used to determine homogeneous subsets and establish which stages had significant differences in their pinkness index. The homogenous subsets broke down the stages into two groups. Group a consisted of Bud, Early Staminate, Late Staminate, and Early Pistillate. Group b consisted of Early Pistillate, Mid Pistillate, and Late Pistillate phases (Fig 6). 15 Figure 6: The average Pinkness Index of each floral stage for the Saponaria officinalis. This includes sixty-seven total samples: nine from bud stage, twelve from early staminate, twelve from late staminate, twelve from early pistillate, ten from mid pistillate, and twelve from late pistillate. The pinkness index of Group a was significantly different than that of Group b. (Davis & Dudle, 2016) These two groups are broken down into the male phase of the Saponaria officinalis, and the female phase, with the very early female phase being the transition. While not statistically different, each stage progressing from early male to late female does seem to become pinker than the previous (Fig. 6). For my research, I quantified DFR expression using the same sample sets Davis and Dudle (2016) used in their research. To determine if there was significant difference in DFR expression across the stages, again, an ANOVA with a post-hoc test was run. However, one outlier was found by the statistical program for the Early Staminate phase that we excluded, which gave us N = 8, 8, 8, 8, 16 8, 7 per stage. The ANOVA test showed there was a significant difference in DFR expression across the different stages (P=0.03; Appendix IV). Next, the Tukey Post Hoc test was used to compare each stage to one another to show which stages were significantly different from one another. Stage A (bud) was significantly different from Stage B (early staminate) for DFR expression (P=0.010; Appendix V). Stage B, along with being different from Stage A, was also significantly different from Stage E (mid pistillate) (P=0.007; Appendix V). Figure 7: The average DFR expression found across the floral stages of the Saponaria officinalis. This analyzes forty-eight samples which includes eight bud (A), nine early staminate (B), eight late staminate (C), eight early pistillate (D), eight mid pistillate (E), and seven late pistillate (F). Group a has a significant difference in DFR expression than Group b. Again, homogeneous subsets were determined and broke down the stages into two groups. Group a contained Early Staminate, Late Staminate, Early Pistillate, and Late Pistillate (Table 2, Fig 7). Group b contained Bud, Late Staminate, Early Pistillate, Mid Pistillate, and Late 17 Pistillate (Table 2, Fig 7). As Figure 7 shows, Early Staminate seems to have the least amount of DFR expression. Some stages are in both group a and b because they are not statistically different from any of the stages in group a, nor are they statistically different than any of the stages in group b. Table 2: A Tukey HSD. This table was used to group the stages into subgroups. This is done by grouping together the stages that did not have significant differences from one another. With this data, there are only two subsets. Subset one has Early Staminate, Late Staminate, Early Pistillate, and Late Pistillate, all which do not have a significant difference between their expression of DFR. Subset two has Bud, Late Staminate, Early Pistillate, Mid Pistillate, and Late Pistillate, all which do not have a significant difference between their expression of DFR. A=bud, B=Early Staminate, C=Late Staminate, D=Early Pistillate, E=Mid Pistillate, F=Late Pistillate. Based on flower structure, A-C are male, and D-F are female. Subset Stage N 1 2 Tukey HSDa,b,c B 9 D 8 12.5125 12.5125 F 7 12.7760 12.7760 C 8 18.8026 18.8026 A 8 25.6177 E 8 26.4251 Sig. 5.0664 .202 .191 Means for groups in homogeneous subsets are displayed. The error term is Mean Square (Error) = 137.263. a. Uses Harmonic Mean Sample Size = 7.958. 18 b. The group sizes are unequal. The harmonic mean of the group sizes is used. c. Alpha = 0.05. An ANOVA test was used to determine if there was significant difference in DFR expression across the stages without including the bud phase. It was determined that without including the bud phase, there was still a significant difference across the stages (P= 0.002; Appendix VII). Since there was significant difference among the stages, a Tukey Post Hoc test was used to establish differences between individual stages. These statistical tools (Appendix VIII) showed that Stage B (early staminate) was significantly different from both Stage C (late staminate) (P=0.05) and Stage E (mid pistillate) (P=0.001). The five different stages were broken up into two groups by a homogenous subset test. Group a consisted of Early Staminate, Late Staminate, Early Pistillate, and Late Pistillate (Fig. 8, Appendix VIII). Group b consisted of Late Staminate, Early Pistillate, Mid Pistillate, and Late Pistillate (Fig. 8, Appendix VIII). These groups are comparable to the groups determined earlier (Fig. 7), apart from the bud phase no longer in Group b. 19 Figure 8: The average DFR expression found across the floral stages of the Saponaria officinalis. This analyzes forty samples which includes nine early staminate, eight late staminate, eight early pistillate, eight mid pistillate, and seven late pistillate. Group a has a significant difference in DFR expression than Group b. Data analyzed without Bud Phase. I had hypothesized that the Bud stage would have the least amount of DFR expression, so I also analyzed the data without the bud phase, since the stage deviated from my original prediction. From my results, DFR expression does not influence how pink the late female stage flowers get. However, there is a significant change in DFR expression across the Saponaria officinalis lifetime. Discussion Correlation of DFR expression vs color of female-phase flowers Although there was not a statistically significant difference in pinkness levels of the flowers collected from the various locations, previous studies have shown that sun exposure increases the color and concentration of anthocyanins in petals of S. officinalis (Jabbari et al, 20 2012). The purpose of collecting sun and shade flowers in this study was to provide a sample of flowers displaying a range of pinkness levels to look at gene expression in pistillate-phase flowers of S. officinalis. To that effect, in the 53 flowers collected, PI index ranged almost tenfold, from 0.036472605 to 0.323499179. However, as we were interested in how DFR expression affects final color, we preferentially collected flowers across the pink color and collected only a few (8) samples that were from plants growing in the shade and had paler flowers. It is likely that the small sample size of shade flowers contributed to the lack of significant differences in PI across sun exposure levels, but the trend of shade plants having paler female-phase flowers than sun plants matched what we would predict based on previous studies. Expression of the gene responsible for producing anthocyanidin synthase (ANS), an enzyme later in the anthocyanin pathway, has been shown to have a significant positive correlation with the pinkness level reached in a pistillate-phase flower (unpublished data). Theoretically, DFR should have shown the same correlation that ANS showed, since DFR is an earlier enzyme in the pathway, and therefore for there to be reactions that need ANS enzymes, there first needs to be products of reactions that DFR enzymes catalyze. However, DFR does not show significant correlation with pinkness (Fig 5). This does not support my hypothesis that the pinker the late pistillate-phase flowers were, the higher the DFR expression. Therefore, I revised my original hypothesis. I hypothesized that ANS expression is solely responsible for the differences among the Saponaria officinalis in their female-phase flower color. This phenomena of ANS being the determining step for differences in color change among individuals is not just found in the Saponaria officinalis. Farzad et al (2003) showed that when comparing light and dark treatments, ANS had a large increase in expression in light treated flowers, while DFR production only had a slight increase in expression. This is like my 21 experiment as we chose flowers with a wide range of pinkness to see if DFR expression varied, including those grown in shade and those grown in sun, and we concluded there is no correlation between pinkness and DFR expression. Farzad et al (2003) similarly concluded that most of the floral color change from white to purple is dependent on changes in ANS expression, while DFR expression changes very little. Changes in DFR expression as the flower transitions from male to female In a previous study, Jabbari et al. (2012) found that as flowers of S. officinalis transition from white to pink, the petals show a significant increase in the concentration of anthocyanidin pigments, and this is also correlated with the gender change in the individual flowers. The purpose of this study was to further analyze the genetic mechanism that underlies this increase in pigment production. We confirmed that the flowers used in this study also showed this increase in pinkness (Fig. 6), as we found that the staminate-phase flowers had significantly paler flowers than the pistillate-phase flowers. Since there was no correlation established between DFR expression and how pink flowers of the pistillate-phase of the Saponaria officinalis became, I did not expect to observe any difference between DFR expressions in the different stages of the Saponaria officinalis, even though I hypothesized earlier that there would be. While not every single stage had significant differences of DFR expression, there were a few stages that were different from one another. The bud stage did have a significantly higher level of DFR expression than the early staminate stage (Fig. 7). This is interesting, since we believed that anthocyanins were there mainly to change the color of the flower. However, it is not unprecedented, as Kawbata et al. (2009) found that there was a preferential accumulation of flavonoids (Fig 1 shows how flavonoids relate to 22 anthocyanins) at the bud stage and at anthesis in the Eustoma grandiflorum. However, this increase in DFR expression may not be related to color production. Flavonoids are a large group of organic molecules found in plants that are responsible for secondary metabolism, and one subgroup of these chemicals are anthocyanins (Davies et al., 2002). Flavonoids protect plants against various biotic and antibiotic stresses, including absorbing harmful UV radiation that can induce cellular damage (Samanta et al., 2011). Samanta et al. (2011) concluded that some other functions include signaling molecules that activate pathways for root nodule growth and help promote germination and early plant growth. Similarly, flavonoids may be important to the bud stage to promote growth, provide signaling molecules for flower tissue formation, and possible light stress from the sun. Since the higher expression of DFR in the bud stage was interesting, but may be unrelated to floral color, we decided to analyze the data without the bud stage. Not including the bud phase, the DFR expression of early staminate-phase showed a significant difference from the late staminate-phase and the mid pistillate-phase. The early staminate-phase and the mid pistillate-phase had consistent significant differences in DFR expression each set of data. Figure 8 clearly shows that the early staminate phase has the lowest amount of DFR production, whereas the mid pistillate phase has the highest amount of DFR production. Farzad et al. (2003) found in Violas that transition from white to purple, DFR expression is not only present in all stages of the Violas life but seems to be expressed at consistent levels across the floral lifespan. This is contrary to what I found, where DFR expression seems to peak as the color change begins. However, they also did find that DFR expression does increase over Viola's life by about 56% (Farzad et al., 2003), a similar finding to the second part of my 23 experiment. This means that the pattern seen in S. officinalis in which DFR expression increases during floral ontogeny may be a widespread mechanism controlling floral color change. Overall, I found that in the Saponaria officinalis, DFR expression does not seem to correlate with pinkness in late phase females, but DFR expression does seem to increase over the Saponaria officinalis lifetime. However, before researchers move forward, they should revisit this experiment and make a few improvements before making any final conclusions. For the first part of this experiment, larger and equal sample sizes should be collected from the Saponaria officinalis in sun, shade, and wild locations. This may give researchers a better idea on whether DFR expression has any impact on pinkness for female phase flowers. For the second part of this experiment, again, doing the experiment with a larger sample size may help solidify whether DFR expression shows a similar pattern to what was found in this study. QPCR also has many steps with micro amounts of liquids being transferred, meaning that there is plenty of room for error. There is also the possibility of lost or damaged DNA and RNA, so repeating this study all around would be a good idea. If a general increase in DFR as flowers develop is observed again, more research should be done to reveal why only two phases of six are significantly different from one another. One observation, made by reviewers, is stages A, C, and E had higher DFR production than B, D, and F. Stage A could be considered a pre-B, Stage C could be considered a pre-D, and Stage E could be considered pre-F. This could correlate with the flower first blooming as a male (A to B), the flower losing its male reproductive organs and gaining its female reproductive organs (C to D), and the maturation of the female flower (E to F). Further research could investigate whether these changes can be linked to DFR production, or perhaps they will find other reasons why these three stages seem to have higher production of DFR than the other three. 24 Reflection The original goals of this project were to learn the technique of qPCR, to present my findings in an oral presentation, and to achieve significant results. I fully met my goal with learning how to do qPCR, and I also achieved significant results in both portions of my project, as both parts of my experiment contributed to the larger picture of understanding the color change of S. officinalis. However, I was not able to orally present my project due to the coronavirus pandemic preventing in person meetings. However, I will still be turning in a poster to the Biology department and a manuscript to the Honors College, so my findings are not completely unpresented, just in a different format than originally intended. There were two main difficulties that I had to overcome during this project. First, very early on, the DFR primers did not seem to be working, even though they were indeed amplifying the correct gene. However, this was probably due to my lack of experience of running qPCRs, as just a drop of extra or less solution will mess up findings, and by the time I was ready to run actual samples, all the kinks seemed to have worked out. Secondly, the -80 freezer in the Biology department malfunctioned and the samples that I was using for the life stages portion of my research were ruined. Luckily, DePauw University had some RNA versions of the samples I was to use, so I was able to carry out the second portion of my research. This project was personally important to me as I am now one step closer to graduation, something to which I have been looking forward. I did not find much interest in my high school graduation, but as my senior year has come and gone, I find myself looking ever so forward to my college graduation, something that is sure to be ten times more special, especially since I will now graduate with distinction because of this project and my work in the Honors College. This 25 project was professionally important to me for two different reasons. Firstly, coming into college, I had no idea what I wanted to do after college, let alone high school. I knew that I enjoyed biology, and that my genetics units were always my favorites in my biology classes. But beyond that, I had no clue what to do with my career. However, through this project, I have learned that I actually like genetic research, and while Im not hugely thrilled with presentations and papers, I also recognize that if I want to do genetics research, it includes the presentations and papers. Secondly, if I want to continue my education, which I do, or even get a job related to any kind of research, I need to have experience. This project has taken place over two years and has been a tedious process. But it is also wonderful practice for any upcoming research I do in my future. I had to learn a new technique, work around a few obstacles, run and analyze data, make conclusions, and share my data. 26 References Davies, K. M., Schwinn, K. E., Deroles, S. C., Manson, D. G., Lewis, D. H., Bloor, S. J., & Bradley, J. M. (2003). Enhancing anthocyanin production by altering competition for substrate between flavonol synthase and dihydroflavonol 4-reductase. Euphytica, 131, 259268. Davis, S. L., & Dudle, D. A. (2016). Time Course Pinkness Index Data Davis, S. L., Dudle, D. A., Nawrocki, J. R., Freestone, L. M., Konieczny, P., Tobin, M. B., & Britton, M. M. (2014). Sexual Dimorphism of Staminate- and Pistillate-Phase Flowers of Saponaria officinalis (Bouncing Bet) Affects Pollinator Behavior and Seed Set. PLoS ONE, 9(4). doi: 10.1371/journal.pone.0093615 Davis, S. L., & Turner-Jones, L. (2008). Potential for mixed mating in the protandrous perennial Saponaria officinalis (Caryophyllaceae). Plant Species Biology,23(3), 183-191. doi:10.1111/j.1442-1984.2008.00219.x Delph, L. F., & Lively, C. M. (1989). The Evolution of Floral Color Change: Pollinator Attraction Versus Physiological Constraints in Fuchsia excorticata. Evolution, 43(6), 1252-1262. doi: 10.2307/2409360 Dominguez, R., & Holmes, K. C. (2011). Actin structure and function. Ann.Rev. Biophya. 40, 169-186. Retrieved from https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3130349/ Freestone, L. M., Konieczny, P., Tobin, M. B., & Britton, M. M. (2014). Sexual Dimorphism of Staminate- and Pistillate-Phase Flowers of Saponaria officinalis (Bouncing Bet) Affects Pollinator Behavior and Seed Set. PLoS ONE,9(4), 1-14. doi:10.1371/journal.pone.0102865 Farzad, M., Griesbach, R., Hammond, J., Weiss, M. R., & Elmendorf, H. G. (2003). Differential expression of three key anthocyanin biosynthetic genes in a color-changing flower, Viola 27 cornuta cv. Yesterday, Today and Tomorrow. Plant Science,165(6), 1333-1342. doi:10.1016/j.plantsci.2003.08.001 Glover, B. J., & Martin, C. (2012, March 5). Anthocyanins. Current Biology,22, R147-R150. Retrieved from https://www.sciencedirect.com/science/article/pii/S096098221200022X Jabbari, S. G., Davis, S. L., & Carter, E. J. (2012). Interaction between floral color change and gender transition in the protandrous weed Saponaria officinalis. Plant Species Biology,28(1), 21-30. doi:10.1111/j.1442-1984.2011.00352.x Kang, J.-H., McRoberts, J., Shi, F., Moreno, J. E., Jones, A. D., & Howe, G. A. (2014, March 1). The Flavonoid Biosynthetic Enzyme Chalcone Isomerase Modulates Terpenoid Production in Glandular Trichomes of Tomato. Retrieved from http://www.plantphysiol.org/content/164/3/1161 Kawabata, S., Li, Y., Saito, T., & Zhou, B. (2009). Identification of differentially expressed genes during flower opening by suppression subtractive hybridization and cDNA microarray analysis in Eustoma grandiflorum. Scientia Horticulturae, 122(1), 129133. doi: 10.1016/j.scienta.2009.03.011 Lloyd, D. G., & Webb, C. J. (1986). The avoidance of interference between the presentation of pollen and stigmas in angiosperms I. Dichogamy. New Zealand Journal of Botany,24(1), 135-162. doi:10.1080/0028825x.1986.10409725 13 Makino, T. T., & Ohashi, K. (2016). Honest signals to maintain a long-lasting relationship: Floral colour change prevents plant-level avoidance by experienced pollinators. Functional Ecology,31(4), 831-837. doi:10.1111/1365-2435.12802 Nawrocki, J. (2013). An analysis of the relationship between floral color, pollinator behavior, and fitness in the protandrous weed Saponaria officinalis (University of Indianapolis, Unpublished Honors Thesis) Samanta, A., Das, G., & Das, S. K. (2011). Roles of Flavonoids In Plants. Int J Pharm Sci Tech, 28 6(1), 1235. Weiss, M. R. (1995). Floral Color Change: A Widespread Functional Convergence. American Journal of Botany,82(2), 167. doi:10.2307/2445525 Willson, M. F. (1994). Sexual Selection in Plants: Perspective and Overview. The American Naturalist, 144, S13-S39. doi: 10.1086/285651 29 Appendix 30 Appendix II An ANOVA statistical analysis describing the effect of location (grown in sun, grown in shade, or grown in the nature park) on pinkness of late pistillate phase flowers. Dependent Variable: Pinkness Source Type III Sum of Squares df Mean Square .019a 2 .010 1.876 .164 Intercept 1.321 1 1.321 258.806 .000 Exposure .019 2 .010 1.876 .164 Error .245 48 .005 Total 2.187 51 .264 50 Corrected Model Corrected Total a. R Squared = .072 (Adjusted R Squared = .034) F Sig. 31 Appendix III Pearson Correlation Coefficient of Pinkness describing that DFR expression is not correlated at all with how pink late phase pistillate flowers get. PI PI Pearson Correlation expression 1 Sig. (1-tailed) N expression Pearson Correlation Sig. (1-tailed) N -.050 .373 51 45 -.050 1 .373 45 47 32 Appendix IV ANOVA describing the effect of the life stage of the Saponaria officinalis on the amount of DFR expression. Source Type III Sum of Squares Df Mean Square F Sig. Corrected Model 2894.332a 5 578.866 4.217 .003 Intercept 13583.46 1 1 13583.461 98.959 .000 Stage 2894.332 5 578.866 4.217 .003 Error 5765.060 42 137.263 Total 22055.88 9 48 Corrected Total 8659.391 47 33 Appendix V Tukey HSD and Tamhane analysis. Each stage is broken down and compared to each of the other four individual stages to look for significant differences between the two stages. Significant differences are those whose Sig. column is less than 0.05. 95% Confidence Interval (I) Stage (J) Stage Mean Difference (I-J) Tukey HSD A B C D Std. Error Sig. Lower Bound Upper Bound B 20.5513* 5.69293 .010 3.5565 37.5461 C 6.8151 5.85797 .851 -10.6724 24.3026 D 13.1052 5.85797 .243 -4.3823 30.5927 E -.8074 5.85797 1.000 -18.2949 16.6801 F 12.8417 6.06358 .298 -5.2596 30.9430 A -20.5513* 5.69293 .010 -37.5461 -3.5565 C -13.7362 5.69293 .175 -30.7310 3.2586 D -7.4461 5.69293 .779 -24.4409 9.5487 E -21.3587* 5.69293 .007 -38.3535 -4.3639 F -7.7096 5.90428 .780 -25.3354 9.9161 A -6.8151 5.85797 .851 -24.3026 10.6724 B 13.7362 5.69293 .175 -3.2586 30.7310 D 6.2901 5.85797 .889 -11.1974 23.7776 E -7.6225 5.85797 .783 -25.1100 9.8650 F 6.0266 6.06358 .917 -12.0747 24.1279 A -13.1052 5.85797 .243 -30.5927 4.3823 B 7.4461 5.69293 .779 -9.5487 24.4409 C -6.2901 5.85797 .889 -23.7776 11.1974 E -13.9126 5.85797 .188 -31.4001 3.5749 34 E F Tamhane A B C F -.2635 6.06358 1.000 -18.3648 17.8378 A .8074 5.85797 1.000 -16.6801 18.2949 B 21.3587* 5.69293 .007 4.3639 38.3535 C 7.6225 5.85797 .783 -9.8650 25.1100 D 13.9126 5.85797 .188 -3.5749 31.4001 F 13.6491 6.06358 .237 -4.4522 31.7504 A -12.8417 6.06358 .298 -30.9430 5.2596 B 7.7096 5.90428 .780 -9.9161 25.3354 C -6.0266 6.06358 .917 -24.1279 12.0747 D .2635 6.06358 1.000 -17.8378 18.3648 E -13.6491 6.06358 .237 -31.7504 4.4522 B 20.5513 6.16505 .148 -4.9681 46.0708 C 6.8151 6.56452 .997 -18.4378 32.0680 D 13.1052 6.14990 .645 -12.4478 38.6582 E -.8074 9.34745 1.000 -33.8948 32.2800 F 12.8417 6.38982 .694 -12.4509 38.1344 A -20.5513 6.16505 .148 -46.0708 4.9681 C -13.7362* 3.05387 .014 -25.1071 -2.3654 D -7.4461* 2.01357 .032 -14.4379 -.4543 E -21.3587 7.32174 .268 -52.0580 9.3406 F -7.7096 2.65762 .199 -17.6372 2.2180 A -6.8151 6.56452 .997 -32.0680 18.4378 B 13.7362* 3.05387 .014 2.3654 25.1071 D 6.2901 3.02316 .622 -5.0780 17.6583 E -7.6225 7.66114 .998 -37.8829 22.6379 35 D E F F 6.0266 3.48540 .819 -6.4532 18.5064 A -13.1052 6.14990 .645 -38.6582 12.4478 B 7.4461* 2.01357 .032 .4543 14.4379 C -6.2901 3.02316 .622 -17.6583 5.0780 E E -13.9126 7.30899 .779 -44.6441 16.8189 F -.2635 2.62228 1.000 -10.1869 9.6600 A .8074 9.34745 1.000 -32.2800 33.8948 B 21.3587 7.32174 .268 -9.3406 52.0580 C 7.6225 7.66114 .998 -22.6379 37.8829 D 13.9126 7.30899 .779 -16.8189 44.6441 F 13.6491 7.51198 .812 -16.7472 44.0454 A -12.8417 6.38982 .694 -38.1344 12.4509 B 7.7096 2.65762 .199 -2.2180 17.6372 C -6.0266 3.48540 .819 -18.5064 6.4532 D .2635 2.62228 1.000 -9.6600 10.1869 E -13.6491 -44.0454 16.7472 The error term is Mean Square (Error) = 137.263. *. The mean difference is significant at the 0.05 level. 7.51198 .812 36 Appendix VI ANOVA describing the effect of the life stage of the Saponaria officinalis on the amount of DFR expression. However, this test does not include the Bud Stage Source Type III Sum of Squares df Mean Square Corrected Model 2234.543a 4 558.636 Intercept 8607.829 1 Stage 2234.543 4 558.636 Error 3652.660 34 107.431 Total 14582.275 39 Corrected Total 5887.203 38 a. R Squared = .380 (Adjusted R Squared = .307) F Sig. 5.200 .002 8607.829 80.124 .000 5.200 .002 37 Appendix VII Tukey HSD and Tamhane analysis. Each stage is broken down and compared to each of the other four individual stages to look for significant differences between the two stages. Significant differences are those whose Sig. column is less than 0.05. This data analysis is without the Bud Stage. 95% Confidence Interval (I) Stage (J) Stage Mean Difference (I-J) Tukey HSD B C D E Std. Error Sig. Lower Bound Upper Bound C -14.9306* 5.18245 .050 -29.8537 -.0075 D -8.6405 5.18245 .467 -23.5636 6.2826 E -22.5531* 5.18245 .001 -37.4762 -7.6300 F -8.9040 5.36435 .471 -24.3509 6.5429 B 14.9306* 5.18245 .050 .0075 29.8537 D 6.2901 5.18245 .744 -8.6330 21.2132 E -7.6225 5.18245 .588 -22.5456 7.3006 F 6.0266 5.36435 .793 -9.4203 21.4735 B 8.6405 5.18245 .467 -6.2826 23.5636 C -6.2901 5.18245 .744 -21.2132 8.6330 E -13.9126 5.18245 .077 -28.8357 1.0105 F -.2635 5.36435 1.000 -15.7104 15.1834 B 22.5531* 5.18245 .001 7.6300 37.4762 C 7.6225 5.18245 .588 -7.3006 22.5456 D 13.9126 5.18245 .077 -1.0105 28.8357 F 13.6491 5.36435 .104 -1.7978 29.0960 38 F Tamhane B C D E F B 8.9040 5.36435 .471 -6.5429 24.3509 C -6.0266 5.36435 .793 -21.4735 9.4203 D .2635 5.36435 1.000 -15.1834 15.7104 E -13.6491 5.36435 .104 -29.0960 1.7978 C -14.9306* 2.84569 .006 -25.4939 -4.3673 D -8.6405* 1.68115 .002 -14.3509 -2.9300 E -22.5531 7.23738 .151 -51.1953 6.0891 F -8.9040 2.41552 .058 -18.0702 .2622 B 14.9306* 2.84569 .006 4.3673 25.4939 D 6.2901 3.02316 .477 -4.3597 16.9400 E -7.6225 7.66114 .986 -35.8501 20.6051 F 6.0266 3.48540 .680 -5.7137 17.7670 B 8.6405* 1.68115 .002 2.9300 14.3509 C -6.2901 3.02316 .477 -16.9400 4.3597 E -13.9126 7.30899 .635 -42.4220 14.5967 F -.2635 2.62228 1.000 -9.5549 9.0279 B 22.5531 7.23738 .151 -6.0891 51.1953 C 7.6225 7.66114 .986 -20.6051 35.8501 D 13.9126 7.30899 .635 -14.5967 42.4220 F 13.6491 7.51198 .671 -14.6452 41.9434 B 8.9040 2.41552 .058 -.2622 18.0702 C -6.0266 3.48540 .680 -17.7670 5.7137 39 D .2635 E -13.6491 Based on observed means. The error term is Mean Square (Error) = 107.431. *. The mean difference is significant at the 0.05 level. 2.62228 1.000 7.51198 .671 -9.0279 9.5549 -41.9434 14.6452 40 Appendix VIII A Tukey HSD. This table is used to group the stages into subgroups. This is done by grouping together the stages that did not have significant differences from one another. With this data, there are only two subsets. Subset one has Early Staminate, Late Staminate, Early Pistillate, and Late Pistillate, all which do not have a significant difference between their expression of DFR. Subset two has Late Staminate, Early Pistillate, Mid Pistillate, and Late Pistillate, all which do not have a significant difference between their expression of DFR. This data was analyzed without the Bud Stage. Subset Stage N 1 2 Tukey HSDa,b,c B 8 D 8 12.5125 12.5125 F 7 12.7760 12.7760 C 8 18.8026 18.8026 E 8 26.4251 Sig. 3.8720 .055 .084 The error term is Mean Square (Error) = 107.431. a. Uses Harmonic Mean Sample Size = 7.778. b. The group sizes are unequal. The harmonic mean of the group sizes is used. Type I error levels are not guaranteed. c. Alpha = 0.05.  ...

... Comics and Curricula: Multimodal Literature in the Modern Classroom By Rebeka Wilder Honors Project Thesis Submitted in Partial Fulfillment Of the Requirements for a Baccalaureate Degree With Distinction In the Ron and Laura Strain Honors College of THE UNIVERSITY OF INDIANAPOLIS May 2020 Faculty Advisor: Jessica Bannon, B.A., M.A., Ph.D. Executive Director of Honors: James B. Williams, Ph.D. Rebeka Wilder 2020. Copyright in this work rests solely with the author. Please ensure that any reproduction or re-use is done in accordance with relevant national copyright protection. i Abstract For my Honors Project, I created a course curriculum for an eleventh- and twelfth-grade high school English Language Arts class that uses comic books to teach students how to read and analyze multimodal texts through a variety of theoretical lenses, including historical, postcolonial, and feminist approaches. Multimodal texts utilize multiple modalitiesforms of expression such as dialogue, images, text, etc.to communicate messages to readers. Recently, educators have begun expanding the traditional literary canon used in English Language Arts classrooms to be more inclusive. My project explores the potential of expanding that canon to include multimodal texts, specifically comic books. The main texts in my curriculum are a variety of Marvel Comics starring the heroes Captain America, Black Panther, and Captain Marvel. These comic books serve as important subjects of literary analysis because they provide significant examples of multimodal literary devices and thematic development that students would benefit from studying. In addition, the popularity of these comics and the Marvel Cinematic Universe can ensure increased student engagement in course concepts. ii Table of Contents Abstract.............................................................................................................................................i Table of Contents.............................................................................................................................ii Introduction....................................................................................................................................iii Literature Review.............................................................................................................................v Analysis........................................................................................................................................xiv Conclusion.....................................................................................................................................xx Afterword......................................................................................................................................xxi Works Cited................................................................................................................................xxiii iii Introduction In English Language Arts classrooms across America, many educators will teach texts from the traditional canon of accepted high school literature. These texts are often called classics and include novels such as The Great Gatsby by F. Scott Fitzgerald and To Kill a Mockingbird by Harper Lee, as well as Shakespeare plays like Hamlet or Romeo and Juliet. Those texts have significant literary merit, and they contain engaging themes, motifs, and characters that educators have relied on for many years to teach various skills in English Language Arts classrooms. However, by only teaching texts within the traditional literary canon, educators overlook a whole host of literature that employ various modes of communicating meaning. My project addresses this issue of a too-narrow canon of literature by integrating multimodal texts into a high school English Language Arts curriculum. Educators already use diverse learning strategies to support and engage students with a variety of learning needs; integrating a diverse selection of literature can further support student engagement by introducing critical reading and writing through multimodal texts similar to those students encounter in their daily lives. Although multimodal texts are being used more often as educators realize they must help prepare students for comprehending texts in todays technology-driven culture, there are doubts surrounding the complexity of multimodal texts. Dr. Michelle Falter, a former middle and high school ELA teacher and current Professor of English Education, addresses this concern, stating, There is a stereotype that graphic literature is a lesser form of writing. This is absolutely untrue. The tasks and thinking skills required to read a multimodal text are actually higher level than if reading a print-based text alone. You have to see images and words work together, and when and why authors chose to put them together in a frame. I also think people think graphic novels are for ELL or elementary and middle grades students iv only. Although they are definitely very helpful for these students, this misconception comes from the fact that people see this genre as easyand it is not. (qtd. in Gonzalez) While my project uses texts from the traditional literary canon, those texts are used as supplementary material rather than as the main texts in the curriculum. The main texts consist of comic books published by Marvel Comics. Comic books are not only excellent examples of multimodal literatureas they contain an interesting overlap between dialogue, text, images, and colorbut they have also evolved over the years to better reflect and challenge their readers cultures, political systems, and social relationships. Although the main characters have remained the same over the years, their purpose changes as the authors, illustrators, and society changes. For example, Steve Rogers has always been the original Captain America but over the different iterations of his comics, his character evolves as new plotlines, villains, and fellow heroes emerge. I chose to integrate Marvel Comics in my curriculum since characters from Marvel Comics have become popular in recent years due to the popularity of the movie franchise. Students familiarity with Marvel Comics characters as a result of the movie franchise will increase their interest in the course reading and their motivation to engage in literary analysis of these multimodal texts. The Marvel Cinematic Universe is the highest grossing movie franchise of all time, having made over $22 billion in the Worldwide Box Office over the past fourteen years alone with several movies currently in production (Movie Franchises). Due to the incredible popularity of Marvels movie franchise select movies are used as supplementary materials in my curriculum. The movies included in the curriculum serve the same purpose as the music videos and artwork listed on the supplementary materials listto expose students to a diverse selection of multimodal work with the intent to further develop their ability to analyze multiple modes of communication. v Literature Review Scholarly Criticism of Marvel Comics Within the last two decades, numerous scholars have published analyses of various comic books published by Marvel Comics. This scholarly work serves two functions in the context of my Honors Capstone Project: first, it provide evidence that comic books can be analyzed using different literary criticism theories to find complex meanings; and second, I will integrate them in my curriculum as examples of how to successfully analyze multimodal texts using close reading skills and applying literary criticism theories to comic books. I selected the following articles due to their readability and more recent publication dates. Since comic books are typically viewed as more simple forms of literature because they contain illustrations and often cater to a younger audience, scholars have recently begun to seriously consider the value of comic books. In fact, some educators who advocate for the inclusion of graphic novels in ELA curricula dismiss comic books as being silly, refusing to consider their merit as multimodal texts (Gonzalez). It is important for students to have reliable examples and sources to refer to when learning about how to analyze multimodal texts and integrate textual evidence into their essays. Being able to break down and study how scholars created their arguments is an instructional strategy to teach writing skills to writers of all skill levels. One of the Marvel superheroes whose comic books I plan to teach in my curriculum is Captain America, one of the heroes with the longest history. Since the original publication in 1941, as different authors and illustrators created new versions of the character or expanded Captain Americas story, the meaning of the comics changed as well. Scholars Mike Milford and Mike Dubose both wrote articles applying historical criticism to two different Captain America comics, and their conclusions were vastly different. This shows that the same lens can be applied vi to the same comic book character, but the application of historical criticism shows how the character and themes have changed over time to reflect societal and historical concerns. In Veiled Intervention: Anti-Semitism, Allegory, and Captain America, Milfords historical criticism of the original Captain America comics released in 1941 clearly leads readers through the isolationist and anti-Semitic political and social climate of the early 1940s, explains the visual and thematic impact of Steve Rogers heroic crusade against the Nazis, and concludes by detailing the impact of allegorical devices on audiences. Milfords break-down of such broad subjects as isolationist American policies and the events leading up to World War II is incredibly insightful and allows him to form an impactful and convincing argument that comic books were a powerful rhetorical tool for marginalized groups within America. He also provides a detailed and clear analysis of Captain America as an allegorical tool for interventionist rhetoric supported by specific panels from the comic book. Demonstrating another analytical approach, Duboses article Holding Out for a Hero: Reaganism, Comic Book Vigilantes, and Captain America uses three main points to support his claim that in the 1980s, comic book superheroes such as Captain America defined what popular culture viewed as heroism in the face of President Reagans failure to fulfill that role politically. When he analyzes Captain Americas evolving role within Marvel Comics, he notes that the character adapts to reflect the morals and values at the time of publication. Dubose sets up his argument by detailing the progression of Captain America from the 1940s where he was a loyal soldier and American icon to the 1980s where he shed the role of Captain America and took on the name Nomad to fight the corruption within the American government. This depiction of the American government as the villains of such an iconic American comic book hero reflects the disenfranchised feelings of popular culture in America at the time. Duboses analysis of the vii evolution of Captain American comics is an excellent example for student reference. He references a wide range of quotes and panels from the comics and connects them to historical events and figures to create a clear image of the state of American society and politics in the 1980s, and his use of historical criticism to construct his argument is compelling. The strong examples of historical criticism in Milford and Duboses articles prove that comic books contain compelling themes and messages which can be analyzed by scholars. Furthermore, he visual nature of comic books means that readers have a very clear picture of what all the characters look like. In fact, the illustrations are sometimes the most compelling modality for readers. However, it does mean that issues of race are prominent in many comic books even if not directly addressed by the authors. Some comic books or heroes do address race issues directly, such as Black Panther and Lucas Cage. Although I plan to use Black Panther comic books in my curriculum, Tracy Bealers analysis of a 1970s Lucas Cage comic book is a great example of the application of critical race studies to a comic book. In The Man Called Lucas: Luke Cage, Mass Incarceration, and the Stigma of Black Criminality, Bealer uses examples and specific panels from the comics to describe and analyze how Cages incarceration and escape from prison present a sympathetic view of his subsequent social and economic marginalization and systemic racism within the government. The combination of various literary theoriessuch as Marxism, historical criticism, critical race theory, and readerresponse criticismto address multiple aspects of the text and show the impact the comics had on readers and its relevancy to the socio-political issues of both 1970s and modern times makes Bealers article engaging and informative. This article will serve as an example of several different types of literary criticism theories, and even though it does not analyze a Black Panther comic book, students will still get a better idea of how critical race studies works. Critical race viii studies do have similarities to other literary theories that apply to other social issues, such as feminist criticism, due to the similar ways they break down societal constructs like race and gender. Therefore, Bealers article will be of some use during the unit on Feminism and Captain Marvel. In my curriculum, the last unit is designed to give the students more freedom and the opportunity to choose the text and literary criticism theory to apply. I also plan to have students analyze at least one movie or television adaptation of a comic book hero, but film analysis does differ from multimodal analysis, since there are different modalities to consider. Ashley Robinsons application of Marxist theory to the first Iron Man movie analyzes how Tony Stark and his superhero identity Iron Man represent modern American culture through the portrayal of Tony Starks transformation. In her article titled We Are Iron Man: Tony Stark, Iron Man, and American Identity in the Marvel Cinematic Universes Phase One Films, Robinson examines each of the three movies and relates Stark to modern anxieties about the American identity, economic and political turmoil, and social responsibility. Her use of quotes and descriptive language paints a clear picture for readers who may not have seen the movies to easily visualize the scenes used as evidence to support her claims. The use of evidence and connection to modern anxieties held by many Americans creates an incredibly in-depth analysis that reveals the way Iron Man represents the reconceptualized American Dream by combining what she terms frontier values such as individualism and freedom with more modern values such as moral relativism and technological advancement. Robinsons article is an astounding example of both historical criticism and Marxism applied to a modernized comic book figure. Her argument is strengthened by her referral to the original Iron Man comics, by explaining how the overarching purpose of Iron Manto represent the restructuring of American valuesstays the same from ix the 1960s to the release of the first Iron Man movie in 2008, and the subsequent Marvel Studios movies. Since students will be completing an independent study in the last unit of the curriculum, this means they will also have the freedom to choose the literary theory they apply to their chosen comic book. This means that students are not limited to only choosing new historicism, feminism, and race studies. Instead I plan to teach several short lessons about various literary theories the week before students must make their final decisions regarding their independent study. Since I want students to choose a literary theory they feel comfortable working with, but that they are also personally interested in exploring, I plan for the short lessons to cover literary theories such as queer theory, ecocriticism, Marxism, and critical disabilities studies. Once students have made final choices about their comic book and literary theory, I can then begin further instruction on the literary theories students picked that were not taught in prior units. Although this instruction will not be as lengthy as the instruction students received covering new historicism, feminism, and race studies, it will be enough to ensure they can complete their independent studies to similar levels of proficiency. In order to support their work, I compiled a list titled Optional Scholarly Sources in the curriculum that will be used during this period of instruction. This list consists of scholarly sources that apply a variety of literary theories to several different comic books published by Marvel Comics such as Daniel Martins dual use of Marxism and race studies in The Americanization of the Hong Kong Kung Fu Hero: Orientalism and Social Class in Marvel Comics Iron Fist. In addition, it includes Anthony DAgostinos use of queer theory and feminism in Flesh-to-Flesh Contact: Marvel Comics Rogue and the Queer Feminist Imagination to analyze the young hero Rogues sexual identity across several X-Men comics. Also listed is P. Millers comprehensive study of X-Men comics x using critical disabilities studies, queer theory, and Marxism to explore the various marginalized groups represented in X-Men comics titled Mutants, Metaphor, and Marginalism: What XActly Do the X-Men Stand For? As students complete the last unit in the curriculum, they will conduct individual research to find sources that support, expand, or inform their independent study, but the articles on the Optional Scholarly Sources will provide them with a starting point. Curriculum Building, Instructional Strategies, and Justification A major challenge of creating my curriculum was balancing different learning objectives and strategies to best teach students the various analytical skills I want them to learn. One of the most important skills for students to learn when attempting to apply a literary criticism theory, or to critically think about a text, is close reading. Obtaining close reading skills will allow the students to better craft their own sustained interpretations of various texts with enough evidentiary support to defend their interpretations. Jennifer Gonzalez, an educational blogger and former teacher, wrote an article titled Graphic Novels in the Classroom: A Teacher Roundtable, which is a great resource for teachers who are interested in using multimodal texts in their classroom. She includes advice from four teachers who have experience using multimodal texts in their classrooms as reading materials, whether as main texts or supplementary materials. The four teachers answer five questions about why they chose to use multimodal texts, the benefits of using them, as well as different pedagogical techniques that worked best to help their students understand how to read and analyze multimodal texts. This article was very helpful in developing my curriculum, because most methods used to teach textual analysis and close reading skills were constructed around the analysis of traditional texts which only utilize the written word as a mode of communication. xi Another important learning objective of my curriculum is to teach students how to critically analyze a text using specific literary theories. Like the close reading instructional strategies I plan to implement in my curriculum, other teachers have already successfully integrated those strategies. Beth Wilson shares her struggle to teach students literary theory and how to apply it, in addition to several instructional strategies in her article titled Teach the How: Critical Lenses and Critical Literacy. To introduce literary theory and criticism to students she suggests an activity where the teacher generates two separate word lists, each with a different theme. For example, the theme of list one might be nature and contain words such as wind, flower, tornado, forest, and sun. The teacher would then color-code each list so that when they are projected or written on the board at random, the words in list one are all blue and the words in list two are red. Doing this ensures that when students first view the words there is no obvious pattern, so that when the teacher asks them to identify a pattern it is difficult to do so. After this, the teacher then hands out red and blue tinted lenses, much like those on 3D glasses at movie theaters. Students will then be instructed to view the words through each lens and identify patterns now. The red colored lens will cause the words written in blue to stand out, but the words written in red will be difficult to see. After students identify the two themes, the teacher will explain that the colored lenses are literary theories and the words are textsliterary theories are lenses scholars use to analyze texts and identify meaningful patterns that provide structured insight into a text. This activity is meant to visually illustrate the importance of literary criticism theories to students and give them a concrete way to relate such abstract concepts to their own future application of literary theories. Wilson more broadly suggests lessening the amount of modeling done by the teacher and allowing students to find their own textual interpretations, no matter how tempting it is to lead them in a certain direction. She also provides two examples of xii lessons teaching different literary theories and scaffolding instruction, so students have support but can draw their own conclusions. Wilsons article is a thorough, insightful view into the obstacles and tricks to successfully teaching literary theory in a high school classroom. However, she also strongly advocates for the benefits students gain from learning to view texts through various theoretical lenses and how it can help them think more critically about social issues they are already exposed to. There are those who doubt the practicality and usefulness of multimodal texts in classrooms, raising concerns such as the vocabulary and themes presented in graphic novels did not challenge students and that reluctant readers or struggling students would always choose graphic novels as the easy choice and slow down their progress (Do Graphic Novels). Some critics also argue that Since much of the story in a graphic novel is expressed in pictures, rather than wordsreplacing too many standard books with graphic novels will result in a decline in overall literacy (Are Graphic Novels). Contrary to these beliefs, there are several benefits of students reading multimodal texts in a classroom setting, such as literacy skills, engagement, and the prevalence of multimodality in modern society. In recent years, several scholars have begun advocating for the increased use of multimodal texts in classrooms in order to benefit students. One such scholar, Sean Connors, details the importance of multimodal texts to students literacy experiences outside of the classroom and uses research from other scholars to establish that those experiences should be reflected in classrooms. In Expanding Students Analytical Frameworks Through the Study of Graphic Novels, Connors articulates the uses of multimodal texts by applying it to the development of students reading and writing practices. Like Connors, teacher Rachel Williams advocates for the use of multimodal texts for three reasons in her article Image, Text, and Story: Comics and Graphic Novels in the Classroom: students are typically xiii more engaged in the genre, they are mostly inexpensive, and they usually contain easy-to-read vocabulary while still addressing a variety of serious issues. Both she and Connor support their arguments by providing examples of past students work, revealing the importance and efficiency of integrating multimodal texts in ELA curricula. Including multimodal texts in an ELA curriculum benefits not just students, but teachers as well. The inclusion of multimodal texts opens up new possibilities for instruction, curriculum design, and expanding the variety of literature to which they expose students. James Paul Gee, one of the first scholars to gain notoriety for his evaluation of multimodal texts value in classrooms, uses video games to make a case for why teachers should apply learning principles found in video games to their classrooms. Many of the thirty-six learning principles that he defines in his book What Video Games Have to Teach Us About Learning and Literacy can be found in multimodal texts such as comic books. Although the focus of the book is on how these learning principles found in video games apply to the development of literacy skills in classrooms, Gee creates examples for nearly all content areas, showcasing the value of his work for all teachers. His theories and principles are so widely applicable and formed a basis for my own understanding of multimodal texts and their practical applications within classrooms. Another scholar who argues that multimodal texts benefit both students and teachers is Rob Simon, who focuses on how diversifying reading lists can improve relations between teachers and their class by orienting teachers as connoisseurs of their students literacy interests. Simon supports an approach to teaching literacy that is focused on developing a relationship with students by understanding their needs, goals, current literacy practices, and complex identities. Building this relationship allows teachers to better comprehend what will engage their students interests, and to plan accordingly. In the article, Without Comic Books, xiv There Would Be No Me: Teachers as Connoisseurs of Adolescents Literate Lives, Simon uses two student teachers experiences teaching literacy with unconventional and multimodal texts to show that the current typical cannon of literature taught in school needs to be reconsidered and expandedthis helps illustrate Simons arguments and ground them in practical applications. His proposed connoisseur or inquiry approach to teaching combined with support for using multimodal texts in classrooms is a balance between education and literary research, a balance my own Honors Capstone Project attempts to achieve. Analysis All students will have a notebook or journal that they bring to class each day in which they write their responses to the Journal Entry prompts. These prompts will be used to activate prior knowledge about a topic discussed previously in class, to engage students in anticipating what the days lesson will be about, or to reflect on an activity or important concept at the end of the lesson. The Journal Entries in each unit will, generally, follow the theme specified in the Assignments section at the bottom of each unit breakdown on the curriculum. The five themes of the Journal Entries are: reflection on literary theories, patriotism, gender constructs, racial experiences, and project progress. Although the Journal Entries make up only ten percent of the students final grade, due to the quick and relaxed nature of the assignment, they will prompt students to engage in metacognition. Metacognition means having an awareness of ones own thought processes, and when students engage in metacognitive activities it means that they are reflecting on their own knowledge, performance, or feelings. This helps students identify gaps in their knowledge, see areas of success or improvement, and understand how they best learn. Students will complete between two to three Journal Entries per week, and all Journals will be checked at the end of each unit in accordance with the grading rubric in the course curriculum. I xv chose to make the Journal Entries a low percentage of students final grade because I want students to participate in low-stakes, personal writing. The grading rubric makes no mention of grammar conventions or word count but focuses instead on addressing the prompt fully and including relevant details. This, as well as my in-person reminders of the goal of Journal Entries, will ensure students can focus more on the metacognition and writing processes than on their writing conforming to Standard English conventions. The Vocabulary Quizzes serve as a check to confirm whether students can recall contentspecific vocabulary common among scholars who engage in literary criticism and analysis. The expectation that students will use their course vocabulary in class discussions, Journal Entries, and Analysis Essays are the primary ways I support each students ability to perform well on the Vocabulary Quizzes. There will also be short vocabulary review activities to begin or end lessons throughout each unit, which will give me an idea of which words students are struggling with or whether certain students need more direct support to succeed. Additionally, the Vocabulary Quizzes guarantee that students meet any state standards and essential questions related to vocabulary usage in each unit. The Comprehensive Vocabulary Test serves as the Final Exam of the course and the final way for students to show their general knowledge about course concepts, since it requires students to recall all key vocabulary terms used throughout the school year. During each unit, students will complete an Article Summary and Evaluation after they learn about key vocabulary and theories but before they write an Analysis Essay. I purposefully situated the Article Summary and Evaluation assignments in this way to give students an exemplary model of theory application and use of close reading strategies to analyze comic books before they are asked to do so themselves. For each Article Summary and Evaluation xvi students will complete a graphic organizer, all of which will be graded following the rubric in the course curriculum. Below is an example of a successfully completed Article Summary and Evaluation graphic organizer that would earn full credit, or thirty points. Name: Jane Doe Article Title: The Man Called Lucas: Luke Cage, Mass Incarceration, and the Stigma of Black Criminality Author: Tracy Bealer Thesis Statement: in YOUR words 3 Supporting Arguments (Identify 1 piece of evidence that supports each supporting point) Luke Cages heroic origin story, Hero for Hire, reveals the extreme degree to which institutional racism in politics, economics, and society in the 1970s contributed to the mass incarceration of black men in America. 1. The first issueis interrupted by an extended flashback sequence, narrated by Lucas himself. This account both implicates him in the sort of criminal underworld tough-on-crime politicians partially invoked, partially created, in justifying over-policing and mass incarceration, and takes great pains to pardon him from the social fiction of pathological black criminality (Bealer 172). a. The author supports this point both by analyzing dominant images on the comics cover and by describing panels and dialogue depicting Lucas willing association with young criminals yet his resistance to participating in criminal activity. 2. Because our hero is currently incarcerated, the writers must quickly invert the readers expectation that the guards are the good guys, and they do so through the officers increasingly explicit expressions of racism (Bealer 175). a. On page 176 in Figure 2, Bealer uses close reading strategies to analyze how the prison setting visually reveals institutional racism and how the dialogue between guards shows the normality of racist attitudes towards all prisoners in the four panels. 3. The iconic trappings of Luke Cages superhero identityhis costume, name, and the powers themselvesare elegant encapsulations of the way prison marked and continues to mark his psychosocial biography (Bealer 179). xvii a. Bealer uses Luke Cages superhero costume and moniker as evidence to support this point. The imagery of his costume is representative of how the stigma of criminality haunts even his heroic actions and choices, while the moniker Cage is chosen specifically as a reference to how influential prison is on his decision to be a hero in Harlem. Bealer references a specific instance of dialogue in the comic book where Lucas explains his rationale to a friend. In my opinion Bealers argument is well-constructed and Evaluation well-supported by textual and secondary sources; therefore In my opinion the article it is a very strong argument. I believe Bealers argument is very strong because Bealer uses examples and specific panels from the comics to describe and analyze how Cages incarceration and escape Justification from prison present a sympathetic view of his subsequent I believe the argument is social and economic marginalization and systemic racism ___ because within the government. Its relevancy to the socio-political issues of both 1970s and modern times makes Bealers article engaging and informative. To support students successful completion of the Article Summary and Evaluation assignments, I will model the processes of identification, summarization, and evaluation in daily lessons since those are three key skills students need to complete the assignments satisfactorily. Even though not all the strategies will be modeled using examples of scholarly application of literary theories to comic books, watching the teacher model these processes are the first step to students internalizing and developing the skill themselves. My goal is to get students comfortable and familiar with the three processes to complete the first Article Summary and Evaluation in Unit One in small groups. This will provide students who are still developing the skills with peer scaffoldingor instructional supportand students who have developed the skills with time to teach those skills to the developing students. Then, by the time the second Article Summary and Evaluation in Unit Two is assigned, students should be able to complete the assignment xviii individually with time to complete it in class; this allows developing students to receive scaffolding as needed and developed students to practice the skills independently. In Unit Three, when students are assigned the third Article Summary and Evaluation, my goal is to have all students familiar and comfortable with identification, summary, and evaluation to complete the assignment wholly independently with minimal peer or teacher scaffolding. The Article Summaries and Evaluations are also important for developing students writing abilities. Reading and evaluating good examples of writing are two common strategies used to incorporate active learning in daily lesson plans. Active learning allows students to actively participate in the learning process instead of passively receiving information. Providing examples of well-written arguments and analyses also encourages the development of writing abilities because it gives students reputable examples of literary analysis to reference when they write their own analyses later in the curriculum. Students will write three Analysis Essays during the course which make up thirty percent of their final grade. The Analysis Essay grades are weighted more heavily than other assignments because it allows me to evaluate students understanding of course concepts and vocabulary, their ability to organize their ideas logically, and their integration of sources to support their claims. The importance of these skills is reflected in the grading rubric, so students will know the success criteria they need to meet in order to get the best grade possible. Each Analysis Essay is also a chance for students to engage in a long process of active learning, and since the demand for independent work and use of several skills is higher than prior assignments they have completed, the point value is correspondingly higher. There are several instructional strategies I will use to support students in their effort to complete the Analysis Essays. Other course assignments (Journal Entries, Vocabulary Quizzes, xix Article Summaries and Evaluations) are intended to build students skills associated with understanding and correctly using course concepts and vocabulary, evaluating and finding useful quotes in secondary sources, and developing writing skills including organization. The completion of these assignments and the way they support students learning and skill acquisition has already been discussed, but there is another, vital instructional strategy I plan to use to support students essay writing: Essay Workshops. On the course calendar, the week before each Analysis Essay is due, I have days specified as Essay Workshop days. These lessons will focus on developing specific writing skills, peer and teacher review, and personal rereading and editing time. Before the Essay Workshop days begin, I will either do a brief check of student progress or have students complete a survey to gauge and plan the lessons for workshop week so that they address actual student needs and not random topics. Since writing is a personal process and students cannot be expected to work at the same pace the Essay Workshop week is a time for me to provide more individualized scaffolding or independent work time depending on each students needs. The Multimodal Project is a chance for students to creatively showcase the skills they acquired throughout prior units. Like the Analysis Essays, the Multimodal Project is thirty percent of students final grade. This last unit has nine weeks dedicated to it, simply because of how demanding and detailed the final product is expected to be. The Multimodal Project essentially gives students the chance to further explore topics that interest them personally and then complete a Vocabulary Quiz, Article Summaries and Evaluations for their independently researched sources, and write an extended Analysis Essay applying their chosen theory to their chosen text. Although the project is largely independent, since students will be reading different texts and using different theories, the routine established in prior units will continue. Students xx will still complete a Vocabulary Quiz specific to their literary theory, read and discuss their focus text with classmates, and reflect on their knowledge and progress with Journal Entries. So even though students are given quite a bit of choice in Unit Five, they will all be receiving instructional support depending on their need. However, it is expected in Unit Five that the majority of students will be able to receive minimal scaffolding because of their progress and accomplishments in prior units. This cumulative project requires students to draw on all previously learned skills and present the results using multiple modes of communication. The design element of the project will have one week dedicated to it, since students have not yet been graded on designby this week students should have their materials either fully written, reviewed, and edited or be close to that stage. This allows students to do final checks, practice their presentation in front of select peers, and dedicate time to polishing the visuals, audio, and text elements of their project. Conclusion In the education field my project expands the traditional literary canon to include a genre of literature often overlooked and dismissed in academic settings. Although my project incorporates comic books in an ELA curriculum due to my personal interest in Marvel Comics, the same implication applies to other genres of commonly dismissed literature such as science fiction, fairytales (original or retold versions), LGBT+ literature, or multicultural literature, to name a few. However, teachers could also integrate other multimodal texts or mediapodcasts, music, television episodes, social media posts, etc.in their curricula as either supplemental material or as the focus of a unit. Expanding the traditional literary canon does not, of course, imply that the traditional literary canon has no place in ELA classrooms. Rather, the expansion my project suggests will enrich and broaden learning opportunities students have in ELA xxi classrooms. Teachers who wish to integrate comic books or other alternative forms of literature or media in their classroom can do so to whatever degree they think is best for their students whether it is a whole unit or companion material to a more traditional novel, the inclusion of nontraditional literature or media in ELA curricula is not only possible, it is necessary for the benefit of students. Afterword Based on the outcome of my project, future areas of investigation for myself or another researcher are practical application of the curriculum and the inclusion of other genres of literature or media. Sadly, I have not yet gotten the chance to apply my curriculum in a classroom setting which would have given me the opportunity to study direct results and learning outcomes from the various readings and assignments. Another researcher or I could also create a curriculum focused on other types of multimodal texts such as graphic novels, other types of multimodal media such as podcasts, or other genres of literature excluded from the traditional literary canon taught in classrooms such as science fiction. Future research could also explore the potential of teaching nontraditional texts alongside traditional texts as companion texts or material, or even creating units that each focus on a different type of literature not yet part of the traditional literary canon. My personal inspiration for this project was my love of comic books and desire to explore their potential in an academic setting. The project gave me the chance to reflect on and begin articulating my personal teaching philosophy as an English Language Arts teacher. Doing so helped me realize the importance of exposing students to several modes of communication and various types of literature. This is especially important because of the heavy presence of multimodal literature and media in students lives ranging from fun apps and games like TikTok xxii and Fortnite to serious government announcements broadcast on news networks. It is essential for students to have the skills necessary for interpreting messages communicated through the combination of visuals, sounds, and text they ingest daily so they can make informed decisions and responses. Although I chose to integrate comic books as the curriculums main texts, I made sure to incorporate other forms of multimodal literature as well so that I could maximize the modes of communication students would analyze. This is why I changed the title from Supplemental Texts to Supplemental Material in the curriculum, since I originally had compiled a list of text-only materials but upon reflection added music videos, paintings, and a mix of fiction and nonfiction texts. This project has made me extremely conscious of how my choices as an educator can lessen or heighten the learning opportunities students receive in my classroom. xxiii Works Cited Are Graphic Novels Appropriate for the Classroom? Resilient Educator, 11 Sept. 2012, resilienteducator.com/classroom-resources/can-graphic-novels-be-appropriate-forclassroom-use/. Bealer, Tracy L. The Man Called Lucas: Luke Cage, Mass Incarceration, and the Stigma of Black Criminality. Inks: The Journal of the Comics Studies Society, vol. 1, no. 2, 2017, pp. 165185. EBSCOhost, ezproxy.uindy.edu/login?url=http://search.ebscohost.com/login.aspx?direct=true&db=mz h&AN=2018302000&site=ehost-live. Connors, Sean P. Expanding Students Analytical Frameworks Through the Study of Graphic Novels. Journal of Childrens Literature, vol. 41, no. 2, Fall 2015, pp. 5 15. EBSCOhost, ezproxy.uindy.edu/login?url=http://search.ebscohost.com/login.aspx?direct=true&db=eh h&AN=110321484&site=ehost-live. DAgostino, Anthony Michael. Flesh-to-Flesh Contact: Marvel Comics Rogue and the Queer Feminist Imagination. American Literature: A Journal of Literary History, Criticism, and Bibliography, vol. 90, no. 2, June 2018, pp. 251281. EBSCOhost, doi:10.1215/00029831-4564298. Do Graphic Novels Have A Place in The Classroom? Continental, 13 Feb. 2020, www.continentalpress.com/blog/do-graphic-novels-have-a-place-in-the-classroom/. Dubose, Mike S. Holding Out for a Hero: Reaganism, Comic Book Vigilantes, and Captain America. Journal of Popular Culture, vol. 40, no. 6, Dec. 2007, pp. 915 935. EBSCOhost, xxiv ezproxy.uindy.edu/login?url=http://search.ebscohost.com/login.aspx?direct=true&db=mz h&AN=2009395697&site=ehost-live. Gee, James Paul. What Video Games Have to Teach Us About Learning and Literacy. Palgrave Macmillan, 2008. Gonzalez, Jennifer. Graphic Novels in the Classroom: A Teacher Roundtable. Cult of Pedagogy, Cult of Pedagogy, 18 June 2019, https://www.cultofpedagogy.com/teachinggraphic-novels/. Martin, Daniel. The Americanization of the Hong Kong Kung Fu Hero: Orientalism and Social Class in Marvel Comics Iron Fist. Journal of Popular Culture, vol. 51, no. 6, Dec. 2018, pp. 15211539. EBSCOhost, doi:10.1111/jpcu.12745. Milford, Mike. Veiled Intervention: Anti-Semitism, Allegory, And Captain America. Rhetoric and Public Affairs, vol. 20, no. 4, 2017, pp. 605634. EBSCOhost, ezproxy.uindy.edu/login?url=http://search.ebscohost.com/login.aspx?direct=true&db=mz h&AN=2018870760&site=ehost-live. Miller, P.Andrew. Mutants, Metaphor, and Marginalism: What X-Actly Do the X-Men Stand For? Journal of the Fantastic in the Arts, vol. 13, no. 3 [51], 2003, pp. 282 290. EBSCOhost, search.ebscohost.com/login.aspx?direct=true&db=mzh&AN=2003030073&site=ehostlive. Movie Franchises Box Office History. The Numbers, Nash Information Services LLC, 2020, https://www.the-numbers.com/movies/franchises. Robinson, Ashley Suffl. We Are Iron Man: Tony Stark, Iron Man, and American Identity in xxv the Marvel Cinematic Universes Phase One Films. Journal of Popular Culture, vol. 51, no. 4, Aug. 2018, pp. 824844. EBSCOhost, ezproxy.uindy.edu/login?url=http://search.ebscohost.com/login.aspx?direct=true&db=mz h&AN=2018395430&site=ehost-live. Simon, Rob. Without Comic Books, There Would Be No Me: Teachers as Connoisseurs of Adolescents Literate Lives. Journal of Adolescent & Adult Literacy, vol. 55, no. 6, Mar. 2012, pp. 516526. EBSCOhost, doi:10.1002/JAAL.00061. Williams, Rachel Marie-Crane. Image, Text, and Story: Comics and Graphic Novels in the Classroom. Art Education, vol. 61, no. 6, Nov. 2008, pp. 1319. EBSCOhost, doi:10.1080/00043125.2008.11652072. Wilson, Beth. Teach the How: Critical Lenses and Critical Literacy. English Journal, vol. 103, no. 4, March 2014, pp. 68-75, http://www.ncte.org/library/NCTEFiles/Resources/Journals/EJ/1034mar2014/EJ1034Teach.pdf. Accessed 3 October 2019.  ...

... Functionalization of Silica Gel with a Chiral Moiety By Ion Sochirca Honors Project Thesis Submitted in Partial Fulfillment Of the Requirements for a Baccalaureate Degree With Distinction In the Ron and Laura Strain Honors College of THE UNIVERSITY OF INDIANAPOLIS May 2020 Faculty Advisor: Jonathan N. Stack, Ph.D. Executive Director of Honors: James B. Williams, Ph.D. Ion Sochirca 2020. Copyright in this work rests solely with the author. Please ensure that any reproduction or re-use is done in accordance with relevant national copyright protection. i Abstract The proposed research explores the structural differences of the chiral centers between two enantiomers and how they can be used to accomplish a successful separation. The methods for separating chiral mixtures represent a great importance for drug design or special material synthesis. The non-essential amino acid L-carnitine carries out a remarkable function in the human body, the most understood function being the transport of fatty acids in the mitochondria for oxidation1. L-Carnitine has one chiral center and can only be used by mammals in the L form. The method used for the separation of a racemic mixture of D/L carnitine was directly modifying the silica gel (SIG). The modified silica gel can be applied on a thin layer chromatography (TLC) plate or as a packing material for a high-performance liquid chromatography (HPLC) column. A TLC plate is a sheet of glass coated with a thin layer of adsorbent material. It is normally used for separation of molecules of different polarities. Polar molecules travel upwards, while nonpolar molecules stay at the bottom of the plate.2 Weve explored the modifications of SIG to separate an achiral mixture of amino acids on a functionalized silica gel TLC plate as our starting point and have achieved success as a part of the learning of the process necessary to achieve such modification. We were able to make blank glass TLC plate and coat them with functionalized silica gel. We spotted a set of unfunctionalized silica gel TLC plates and a set of functionalized silica gel TLC plates with the same amino acids. During the application of slurry, the layer turned out to be too thick, therefore we could see the solute traveling upwards on the unfunctionalized silica gel TLC plate, but not on the functionalized one. ii Table of Contents Acknowledgements iii List of Tables iv List of Figures vi Statement of Purpose 1 Introduction 4 Procedure 10 Results 15 Conclusion 19 References 20 iii Acknowledgments I want to thank Dr. Stack for giving me professional advice on how to write the manuscript and the methodology of this research. I also want to thank Dr. Styers-Barnett for supervising me during the laboratory and answer any questions regarding the equipment for the experiment. iv List of Tables Table 1: Adapted from Satinder7. Common classification of chiral stationary phases and their respective composition v List of Figures Figure 1: Structure of the silica gel. Figure 2: Example of chiral and achiral objects. Figure 3: Ball and stick model for the carnitine enantiomers. Figure 4: Skeletal structure of L-carnitine. Figure 5: Oxidative polymerization of chiral aniline with potassium persulfate (KPS). Figure 6: The electrostatic forces holding the chiral aniline polymer and the silica gel in proximity. Figure 7: A visual representation of the blank glass TLC plates. Figure 8: Applied functionalized SIG stationary phase on a glass. Figure 9: The unfunctionalized silica gel TLC plates spotted with the same set of four amino acids. 1 Statement of Purpose The scope of this research is to study the modification of a SIG stationary phase with a chiral moiety, for the purpose of separating a racemic mixture of D/L-carnitine. Silica gel is a solid material used to perform organic and analytical separation of a mixture3. Silica gel consists of an irregular framework of covalently bonded alternating silicon and oxygen atoms with pores and voids at the nanometer scale3. Figure 1. Structure of the silica gel4. A stationary phase is the solid support in a chromatography system, which interacts with the molecules in the mixture that are being separated5. Here, weve used the silica gel as the stationary phase, which we wanted to further modify to observe how that would affect the separation of small chiral molecules. The word chiral is derived from the Greek word cheir, which translates to hand. A chiral object has a nonsuperimposable mirror image6. The most universally common example is the difference between your left and right hand. You cant put your left glove on your right hand for the same reason you cant put your right shoe on your left foot6. A chiral object will have a mirror image that is similar, but not the same due to the different 3dimensional arrangement in space. Molecules may also share the property of chirality6. 2 The presence of a chiral center is needed to induce chirality in a separation or a chemical reaction7. A chiral center can be any atom within a molecule that can make four bonds with four other different atom6. In organic chemistry, a chiral center is usually represented by a carbon atom. A pair of molecules that are mirror images of each other are called enantiomers6, and a mixture of enantiomers is called a racemic mixture. The handedness in chirality relates to the rotation of light by a chiral particle. A particle can exist as left-handed, which means it rotates plane polarized light counterclockwise or it can be right-handed, which means it rotates plane polarized light clockwise8. Figure 2. Example of chiral and achiral objects9. D L Figure 3. Ball and stick model for the carnitine enantiomers. 3 We expected the chiral stationary phase to interact with one of the enantiomers more than the other as the racemic mixture passes over the modified silica gel2. The enantiomer that interacts more favorably with the stationary phase should be retained longer on the silica gel, which promotes the separation2. The anticipated final product of this research project was a functionalized silica gel stationary phase that can separate a racemic mixture of D/L-carnitine enantiomers. 4 Introduction Context and Background information: The property of chirality is extremely important in the pharmaceutical industry because all naturally occurring amino acids are a single enantiomer, which is left-handed. The prefix L means levorotatory because the enantiomer with this absolute configuration will rotate light to the left. L-Carnitine is naturally biosynthesized in humans and rotates plane polarized light to the left. All the D enantiomers are dextrorotatory, and they rotate plane polarized light to the right. Scientists first started to realize the need for chiral separation somewhere between 1957 and 1962, when they developed Thalidomide. This drug was meant to prevent morning sickness in pregnant women. However, it was prescribed as a racemic mixture, which is a mixture that contains both enantiomers of the chiral drug. One of the enantiomers interfered with the proper growth of the fetus, which resulted in thousands of children being born with abnormalities10. Figure 4. Skeletal structure of L-carnitine L-Carnitine is a nonessential amino acid synthesized from lysine side chains1. Its function is to transport long-fatty acid chains from the cytoplasm to the mitochondria, where they are oxidized to produce NADH+, Co-A, and FADH2 1. For this reason, many athletes use it as a supplement to boost their athletic performance11. Currently, there are 5 more than 20 human genetic defects associated with the fatty acid oxidation. These individuals have carnitine deficiency; thus, they can only operate on fatty acid oxidation, which occurs in the endoplasmic reticulum of the liver and kidney cells12. It is currently unknown what pathway is more efficient at generating high energy molecules. However, individuals with this disorder are subject to cardiomyopathy, skeletal muscle weakness, and hypoglycemia13. L-Carnitine is also a component of the seminal fluid responsible for count and motility, therefore it might treat male infertility. A study on mice that were kept under stress conditions concluded that carnitine supplementation promotes fatty acid oxidation, which gives more energy to the sperm cell, increasing its overall quality14. More research asserts the effectiveness of carnitine on reducing the effects of cardiac ischemia15, diabetic neuropathy16, and death of lymphocytes in the HIV infected cells17. We can conclude that L-carnitine is useful as a boost of energy for athletes, but even more important as a medical supplement for individuals with a deficiency of this amino acid. One study concluded that D-carnitine is toxic because it inhibits the activity of Lcarnitine1. Therefore, the investigation of separation methods for small molecules of this type is important to scientific research. There are currently three main ways L-carnitine is produced via either enantioselective synthesis or separation. One method involves an optically pure starting material such as (S)-3-hydroxybutyrolactone. After a series of reactions, the product is Lcarnitine18. A second way to produce L-carnitine is through separating a racemic mixture of both D/L carnitine over a modified silica gel19. There are a couple methods known for 6 synthesis of carnitine as a racemic mixture at the level of industry production, which are the U.S. Pat. No. 3,135,788 and U.S. Pat. No. 4,070,39418. And the last method involves genetically modifying bacteria to instruct its catalytic machinery to generate Lcarnitine20. Our proposed research focused on the second method for several good reasons. A modified silica gel can be packed inside an HPLC column or applied to a TLC plate. The methods applied to modify the silica gel for a chiral separation of D/L-carnitine can also be applied to separate other small molecules. Moreover, the techniques we intend to use for chirally modifying silica gel are based on Siddiqs methods for separating amino acids. This modified silica gel can be retained for further use compared to other chiral separation methods or selective synthesis methods that are only specific for one chiral compound. Those other methods require stoichiometric equivalencies of a chiral reagent. Our proposed method is less expensive, and it can be applied more broadly and repeatedly. Final Literature review. Our proposed methods for generating the chiral stationary phase were derived from a previous article using aniline to modify silica gel for separation of amino acids19. For our proposed separation to work, we will attempt to modify aniline with a chiral moiety. 7 Table 1: Adapted from Satinder7. Common classification of chiral stationary phases and their respective composition. Type Description Example 1 Brush-type DNB-glycine, DNB leucine 2 Polysaccharide Chiralcel-OA, -OB, -OF, -OJ, etc. 3 Inclusion CSPs Cyclobond 1-3, Chiralpak OP, OT 5 Ligand exchangers Proline, OH-proline We were attempting to make the brush type chiral stationary phase and achieved modest success. It has a chiral center that can interact with one enantiomer of carnitine more than the other. Also, it contains charged species available for ionic interactions with the material passing over (see Figure 6). These are characteristics of a brush type stationary phase7. Other similar research has studied the efficiency of amino acids separation over cellulose on a TLC plate. A molecule of cellulose consists of a long chain of repeating (+)-D-glucose units, which act the same way modified silica gel does, however it doesnt contain ionic species7. This is a polysaccharide type stationary phase or simply polymers of sugars. Chiral separation was also achieved by creating large cavities on a chiral stationary phase where certain stereoselective interactions of the materials entering and existing cavities would promote the resolution7. This type of stationary phase is called inclusion CSPs. A stereoselective reaction is one that selects for one of the enantiomers that enters and exits the cavities in the stationary phase. The most common stationary 8 phases for this type of separation are cyclodextrins, crown ethers, polyacrylates, and polyacrylamides7. In another interesting research paper, Davankov and Kurganov have achieved chiral separation through ligand exchange. They have indicated that complexes containing copper (II) and fixed ligands promote efficient separation of alanine, serine, and leucine. When such a complex is used as a stationary phase, it exchanges one ligand with one enantiomer in the racemic mixture that passes over the stationary phase7. The enantiomer that provides the complex with the most stability is favored and thus separated. These types of stationary phases are commonly called ligand exchangers. In a recent study, a polymer of chiral ()-2-sec-butylaniline was successfully synthesized according to the FTIR, UV-Vis, and 1H and 13C NMR spectroscopy. The synthesis was accomplished via two distinct methods. The first method is a classical chemical oxidative polymerization of ()-2-sec-butylaniline with an initiator such as ammonium peroxydisulfate in an acidic aqueous medium21. The second method involves an interfacial polymerization that occurs between two immiscible phases of liquids. In the described study the liquids were hexane and methanesulfonic acid. Either of these methods would produce a polymer of aniline with a chiral alkyl substituent21. However, this research did not take a step further and see how the modification of SIG with a chiral polymer of aniline would give it the functionality to separate a racemic mixture of enantiomers. The possible applications for this polymer are overarching. It can be applied in areas including chiral stationary phase for HPLC, biological sensors, electrodes for enantioselective recognition, membrane separation technology, microwave absorbents21. 9 For our purpose, we want to integrate this research to modify the SIG with a chiral polymer of aniline and investigate how well it can separate enantiomers of carnitine. 10 Procedure A brush-type stationary phase was achieved through one method mentioned in the literature review section. We modified the silica gel by a simple oxidative polymerization of aniline in the presence of SIG19. To initiate the polymerization, a strong oxidizing reagent was added to the slurry dropwise such as ammonium peroxydisulfate19. During the initiation, many positively charged aniline radicals are produced22. Radicals are atoms with unpaired electrons in a solution. Unpaired electrons make atoms very unstable. Thus, one radical of aniline will attack another aniline creating a new bond between two chiral anilines. This new species can get oxidized again and attack another chiral aniline, a process we call propagation in organic chemistry22. Once all the oxidizing reagent is depleted, the reaction terminates. The addition of HCl is necessary after the termination is completed in order to protonate the nitrogen atom and give it a positive charge19. 11 Figure 5. Oxidative polymerization of chiral aniline with potassium persulfate (KPS)22. This scheme represents the generation of a polymer consisting of a long chain of aniline monomers. The chain is kept in proximity to the silica via electrostatic interaction between the positive charge on the nitrogen and the negative charge on the SIG. The free silanol groups on the surface of silica gel are capable of forming hydrogen bonds with electron rich species19. Only then, can the polymer of chiral aniline and the silica gel be kept in proximity via electrostatic forces. 12 Hydrogen bond Figure 6. The electrostatic forces holding the chiral aniline polymer and the silica gel in proximity19. For training purposes, we produced a polymer of achiral aniline and bound it to SIG. The research has been conducted by Siddiq and it was the same in terms of procedure and mechanism of reaction. The amounts of reagents were used from a different paper. Before handling dangerous chemicals, proper safety guidelines were used. We began the procedure by dissolving 2.59g of aniline hydrochloride in 50 mL of water and a similar solution of 5.71g of ammonium peroxydisulfate. Solutions were kept at room temperature. In a 400 mL free standing beaker was added 10g of silica gel with a particle size of 60-200 m and a magnetic stirrer. A total of 100 mL of the prepped solutions was poured over 10g of silica gel.23 The silica gel suspension was stirred for 24 hours at 150 rpm23. The initial color of the solution was black and by the end of the reaction it was slightly green. The product was filtered with a Buchner funnel via vacuum filtration and saved in a moisture free environment. 13 The next obstacle was to apply the newly modified silica gel on a blank TLC. The glass was purchased from Lowes and cut in the Art & Design Department. We want the slurry to bind well to the glass and prevent it from sloughing off the glass slides. Therefore, a mixture containing 1g of CaSO4 and 4g of silica gel was made in 10 mL of water. The resulting slurry was loaded into a syringe and the tip of the syringe was moved back and forth to make sure we delivered even amounts of silica on the glass slides. The thickness of the layer must not exceed 0.25 mm24. This sample size was enough to make four TLC plates. Once the slurry was applied, the plates were left to dry for 1 hour. Further on, the plates were heated in the oven at 176F for 45 min. This gets rid of any remaining water and makes the -OH groups available for electrostatic interactions with the polymer of aniline. Once the functionalized TLC plate was prepared, we moved on to preparing the test solutions. Here we strictly followed the methods described in Siddiqs research. At this point we simply wanted to improve our skills at making the TLC plates, thus no racemic carnitine was used. Instead, known amino acids were used one per plate. We were going to measure the distance they would travel with the mobile phase and compare this to the literature value. If the amino acids travel relatively slower compared to the literature value, then we know that our functionalized silica gel stationary phase acts like a brush. The four chosen amino acids were leucine, isoleucine, norleucine, and phenylalanine. Solutions of amino acids were prepared by dissolving 1g of each in 100 mL of distilled water. Four developing chambers were set up with the mobile phase being a mixture of n-butyl alcohol70% aqueous ethylene glycolethyl acetate ratio 5:3:2 by 14 volume24. Enough solvent was delivered to be right between the pencil line and the bottom of the TLC. In our case we used 200 mL of solvent per 800 mL beaker to make the developing chamber. Each chamber was covered with aluminum foil to let the vapor pressure of the solvent to build up inside the chamber. A capillary tube was held and drawn above a Bunsen burner and made thin enough to spot the TLC. We have spotted the TLC plates three times with the same amino acids per each plate. The TLC plates were put inside the chamber until the mobile phase reached of the height of the TLC plates. This took 2 hours and 30 min. Finally, the TLC were sprayed with ninhydrin solution in butanol and infrared light was applied to detect the how far the amino acids have traveled24. In chemistry we use retention factor values (Rf) to show the distance traveled by the solute relative to the distance traveled by the solvent. The equation for Rf !"#$%&'( $*%+(,(- ./ $0( #1,2$( ('4) is !"#$%&'( $*%+(,(- ./ $0( #1,+(&$ ('4) 15 Results Weve successfully synthesized the functionalized the stationary phase polymer of aniline by carefully following literature procedure. Once the oxidizing agent was added, the color of the reaction mixture has turned into a dark brown, which is due to the formation of radical-cations, and oligomers25. Farther on, around 8 a.m., the color of the reaction was dark green, which was due to the growth of hydrochloric acid within the polyaniline matrix25. The reaction has been completed around 2:30 p.m. and the final color was black. A similar research study performed by Seddique has conducted an oxidative polymerization of the protonated form of aniline, and their observations included that the initial color of the reaction was dark brown (64%7 ~520 nm), and it has changed to a dark green (64%7 ~720 nm)25. What is different about our research is that the oxidative polymerization of protonated aniline has taken place in the presence of silica gel, which could be why it farther produced a black color. The mass of the slurry was 11.7205g. Also, the slurry was easily dried and dissolved in water for application on the glass TLC. As previously mentioned, we bought the glass from Lowes and cut it in the Art & Design Department. Below is a representation of the blank TLC plates. 16 Figure 7. A visual representation of the blank glass TLC plates. We have succeeded at making a set of TLC plates with the functionalized silica gel stationary phase. The slurry was applied on the blank TLC with a syringe and spread as thinly and evenly as possible. The stationary phase was well boned to the glass TLC with the help of calcium sulfate. Below is a picture of the TLC plates before and after the modification. 17 Figure 8. Applied functionalized SIG stationary phase on a glass. Weve successfully spotted the functionalized TLC plates with four different amino acids including leucine, isoleucine, norleucine, and phenyl alanine. Initially, we couldnt see the individual amino acids traveling upwards, but this was due to the functionalized silica gel layer being too thick. We decided to take a set of four unfunctionalized silica gel TLC plates and spot them with the same set of amino acids. They were sprayed with ninhydrin and under infrared light it was visible the trail of individual solutes could be visualized under infrared light. Below is a representation of the unfunctionalized blank TLC, however you cant see the solutes because of the absence of infrared light. 18 Figure 9. The unfunctionalized silica gel TLC plates spotted with the same set of four amino acids. We also had success at finding out how separation of a racemic mixture of carnitine works in literature examples. The higher goals that were set up for the project werent achieved, however the learning of the process that leads to those high-end goals was a success. 19 Conclusion Since this attempt was just a training experiment, the greatest challenge was to find a way to apply the modified silica gel slurry as thin as possible and evenly on the glass TLC plates. Our applied slurry was so thick that we couldnt see how far individual amino acids have traveled. Ideally, we need a TLC applicator that is proficient at applying stationary phases on blank glass TLC plates. If such instrument was available, the rest of the research would be doable, and we could get data measurements. Our main goal was to functionalize the silica gel with a chiral polymer of aniline, which was determined to be ()-2-sec-butylaniline as it was used in a similar study. If our training attempt was entirely successful, then the same protocol was to be applied for the main experiment with the substitution of chiral aniline. The main problem that needs to be addressed to move forward is an efficient way to apply the slurry evenly and thin enough (0.25mm). This will prevent the solute from being simply absorbed and it will move up the functionalized silica gel TLC plate. A successful assembly of a chirally functionalized silica gel TLC plate can promote the separation of a racemic mixture of carnitine amino acids. Moreover, this new silica gel can be packed into an HPLC column to separate larger quantities of material. The stationary phase is not selective for any particular compound; therefore, the array of possibilities is very large. 20 References 1. Harmeyer, J. The Phystiological Role of L-Carntine. Lohmann Information [Online] July 2, 2007, p 1. http://lohmann-information.de/content/l_i_27_article_3.pdf (accessed Apr 15, 2019) 2. Stack, J. Chemistry Department, University of Indianapolis, Indianapolis, IN. Personal communication, Mar. 14, 2019. 3. Harris, D. C.; Lucy, C. A. Quantitative chemical analysis 9th ed.; W.H. Freeman et Company: New York, 2016; pp. 671-673. 4. Schematic silica gel surface. https://commons.wikimedia.org/wiki/File:Schematic_silica_gel_surface.png (accessed Mar. 16, 2019) 5. Harris, D. C.; Lucy, C. A. Quantitative chemical analysis 9th ed.; W.H. Freeman et Company: New York, 2016; p. 609 6. Bruice, P. Y. Organic chemistry; Pearson: Harlow, 2017; pp. 150-152. 7. Ahuja, S. LCGC North America 2007, 25(11), 1112-1128. 8. Helicity, Chirality, Mass, and the Higgs. https://www.quantumdiaries.org/2011/06/19/helicity-chirality-mass-and-thehiggs/ (accessed Apr 15, 2019) 9. Chiral and Achiral Molecules. https://socratic.org/organic-chemistry-1/r-and-sconfigurations/chiral-and-achiral-molecules-1 (accessed Apr 15, 2019 10. Bren L. FDA Consum 2001, 35(2), 24-9. 11. Karlic, H.; Lohninger, A. Nutrition 20 (78): 70915 21 12. Nelson, D. L.; Cox, M. M.; Lehninger, A. L. Lehninger principles of biochemistry 7th ed.; W.H. Freeman and Company: New York, 2017. 13. Stanley C. Ann N Y Acad of Sci 2004, 1033, 4251 14. Ng, C.; Blackman, M.; Wang, C.; Swerdloff, R. Ann N Y Acad of Sci 2004, 1033: 17788. 15. Ferrari, R.; Merli, E.; Cicchitelli, G.; Mele, D.; Fucili, A.; Ceconi, C. Ann N Y Acad of Sci. 1033: 7991. 16. Bene, J.; Hadzsiev, K.; Melegh, B. Nutrition & Diabetes. 2018 8(1): 8. 17. Day, L.; Shikuma, C.; Gerschenson, M. Ann N Y Acad of Sci. 2004 1033: 13946. 18. McCarthy, J. R. (Neurocrine Biosciences, CA., USA). Process for the preparation of L-carnitine. US patent 5,473,104, December 5, 1995. https://patents.google.com/patent/US5473104A/en. (accessed Apr 15, 2019) 19. Siddiq, A.; Ansari, M. O.; Mohammad, A.; Mohammad, F.; El-Desoky, G. E. INT J POLYM MATER PO 2013, 63 (5), 277281. 20. Bernal, V.; Sevilla, .; Cnovas, M.; Iborra, J. L. Microbial Cell Factories 2007, 6(1), 31. 21. Modarresi-Alam, Ali Reza & Amirazizi, Homeyra & Movahedifar, Fahimeh & Farokhzadeh, Abdolkarim & Asli, Gholam & Nahavandi, Hamed. J. Mol. 2015, 1083, 17-26. 22. Sapurina, I.; Stejskal, J. Polym Int 2008, 57, 1295-1325. 23. Stejskal, J.; Trchova, M.; Fedorova, S.; Sapurina, I.; Zemek, J. Surface Polymerization of Aniline on Silica Gel. Langmuir 2003, 19, 3013 3018. 22 24. Mohammad, Ali & Haq, Nazrul & Siddiq, Asma. J. Sep. Sci. 2010, 33, 36193626. 25. Seddique M. Ahmed. Polym 2004, 85, 605-614.  ...

... PERCEPTIONS OF PARENTAL LEAVE POLICIES IN HEALTHCARE SETTINGS By Sara Skarshaug Honors Project Thesis Submitted in Partial Fulfillment Of the Requirements for a Baccalaureate Degree With Distinction In the Ron and Laura Strain Honors College of THE UNIVERSITY OF INDIANAPOLIS May 2020 Faculty Advisor: Dr. Jacqueline Hess, PsyD Executive Director of Honors: James B. Williams, Ph.D. Sara Skarshaug 2019. Copyright in this work rests solely with the author. Please ensure that any reproduction or re-use is done in accordance with relevant national copyright protection. PERCEPTIONS OF PARENTAL LEAVE POLICIES IN HEALTHCARE SETTINGS i Abstract The discussion of parental leave has been occurring for several decades, and it is important to be aware of the constantly changing views people have. For this study, students in the University of Indianapolis community were given information about parental leave policies and were asked for their opinions. This study educated participants through the use of a fact sheet and an anecdotal passage about parental leave and gathered feedback about the current opinion of students in the University of Indianapolis community regarding parental leave policies. As a future occupational therapist, my goal is to make sure my patients are healthy and have the resources to succeed. An exploration of the current literature suggests that the health of mothers, fathers, and their children can be improved with paid parental leave programs (Daugher et al. 2013). The goal of this study was to see if perceptions about parental leave changed based on the kind of information people receive. Keywords: parental leave, family-friendly programs, perceptions of parental leave, maternal leave, paternal leave, stereotype threat, workplace stereotype threat PERCEPTIONS OF PARENTAL LEAVE POLICIES IN HEALTHCARE SETTINGS ii Table of Contents Abstract i Acknowledgements iii List of Tables iv List of Figures v Introduction, Intellectual Context, Framework and Theory 1 Method 7 Results 16 Discussion 18 Conclusion 22 Reflection/Future Directions 23 References 25 Appendices 27 PERCEPTIONS OF PARENTAL LEAVE POLICIES IN HEALTHCARE SETTINGS iii Acknowledgements I would like to thank Dr. Jacqueline Hess for being my faculty advisor and guide through this process. Her advice and expertise through every step were extremely helpful and made this project possible. Additionally, I would like to thank Dr. Melissa Loria for her guidance during the creation of my initial proposal. Her comments on my drafts aided in the strengthening of my final product. I would also like to thank Dr. Jordan Waldron for her expertise on data analysis, and for helping me with the statistical aspects of this project. Lastly, I would like to thank the Ron and Laura Strain Honors College at the University of Indianapolis for their support of my academic pursuits and this project. PERCEPTIONS OF PARENTAL LEAVE POLICIES IN HEALTHCARE SETTINGS iv List of Tables Table 1: Gender Demographics Table 2: Age Demographics Table 3: Race/Ethnicity Demographics Table 4: Parental Status Demographics Table 5: Reliability of Stereotype Threat Measure Table 6: Reliability of Perceived Career Outcomes Measure Table 7: Reliability of Utilization Measure Table 8: Test of Homogeneity for ANOVA for Gender Differences Table 9: Test of Normality for ANOVA for Gender Differences Table 10: ANOVA for Differences in Condition Table 11: ANOVA for Differences in Gender 8 9 9 9 11 12 13 16 16 17 17 PERCEPTIONS OF PARENTAL LEAVE POLICIES IN HEALTHCARE SETTINGS List of Figures Figure 1: Power Analysis for Main Effect of Gender Figure 2: Power Analysis for Main Effect of Information Type v 7 8 PERCEPTIONS OF PARENTAL LEAVE POLICIES IN HEALTHCARE SETTINGS 1 Introduction, Intellectual Context, Framework and Theory Perceptions of Gender in the Workplace The concept of paid parental leave has been the topic of conversation in businesses, both big and small, for decades. Although it has the potential to be costly for business owners, it benefits the developing bond between the parents and newborn, and the wellbeing of the mother and the child (Daugher et al., 2013). The Family and Medical Leave Act requires companies in the United States with 50 or more employees to provide at least twelve weeks of unpaid leave for family or medical reasons. This program is available to employees who have been with a company for at least one year and have worked a minimum of 1,250 hours. It includes sick leave, parental leave, or leave to take care of a sick family member (FMLA, 2016). The amount of usage of these programs can vary depending on the work place environment and the amount of support mothers and fathers have if they choose to take leave. An issue that frequently arises in the workplace and that may have an effect on leave policies is how men and women are sometimes viewed differently in professional settings. In one study, researchers presented participants vignettes of two managers, one male and one female, who were identical in how they managed. The results showed that the female manager was perceived to be less likable than her male counterpart in identical situations (Heilman & Okimoto, 2007). Similar issues are seen in healthcare settings. Data on referrals for primary care physicians found that female doctors were punished more severely and for longer periods of time than male doctors for the same unfavorable surgical outcomes (Sarsons, 2017). This type of gender bias in the workplace impacts PERCEPTIONS OF PARENTAL LEAVE POLICIES IN HEALTHCARE SETTINGS 2 female employees in many ways including putting a negative stigma on the utilization of family-friendly programs such as leave of absence, part-time work, and job-sharing. Evidence of this was revealed in a 2017 study by von Hippel, Kalokerinos, and Zacher, which focused on how stereotype threat can be a factor in the utilization of familyfriendly programs. They found that gender-based stereotype threat, or the fear of being viewed with a negative stereotype about ones gender, was positively associated with perceived negative career consequences because of the utilization of family-friendly programs (von Hippel et al., 2017). Similarly, Manchester, Leslie, and Park found in 2008 that maternal leave usage, rather than the birth of the child, is what impacts potential wage-growth in white-collar professions. Working mothers who utilize maternity leave policies experience wage-growth of 2.4% less than working mothers who do not utilize these policies (Manchester et al., 2008). Perceptions of Parental Leave Many companies are beginning to adopt family-friendly policies that are offered to parents of both biological and adopted newborns. Some of these policies include leave of absence, part-time employment, working from home, on-site daycare, flexible scheduling, and job-sharing (where two employees share the workload of one employee). Many factors go into whether or not these benefits will actually be utilized by employees. One of the first papers to look into this matter was published in 1999. Thompson, Beauvais, and Lyness looked into work-family culture (the way a company responds to family matters among their employees) and how it impacts work-family benefit utilization. They sent out a questionnaire to alumni of various graduate programs of PERCEPTIONS OF PARENTAL LEAVE POLICIES IN HEALTHCARE SETTINGS 3 northeastern universities and asked them about their perceptions of work-family culture, organizational support, and work-family benefit utilization at their current companies. They found that individuals were more likely to utilize family-friendly policies when they had more supportive work-family cultures. The results also revealed the idea that those who chose to take advantage of these policies had a greater commitment to their company, and were less likely to leave (Thompson et al., 1999). Although this study was done many years ago, its findings are still relevant and supported by current literature. In a more recent study, it was found that companies that are ranked higher on lists such as Forbes Americas Best Employers and Glassdoors Best Places to Work are associated with offering at least some paid maternity leave (Solheid, 2016). For example, before Google increased maternity leave to five months, female employee retention rates were half that of men. After the new implementation, female and male retention rates were equal (West, 2015). Gender Differences in Perceptions of Parental Leave In 2004, Butler, Gasser, and Smart conducted research on work outcome expectancies for men and women who utilized family friendly policies. They distributed a questionnaire to 188 parents asking about perceived career and family outcomes. Findings revealed that women were influenced by their expected negative career outcomes more than men when deciding whether or not to utilize family-friendly programs. These findings suggest that women perceive their societal punishment to be greater in the workplace when they take advantage of family-friendly programs offered than men do (Butler et al., 2004). Von Hippel et al. supported these findings by PERCEPTIONS OF PARENTAL LEAVE POLICIES IN HEALTHCARE SETTINGS highlighting the differences between working mothers and non-working mothers. Their data revealed that working mothers are seen as being less committed to their job when compared to their male counterparts (von Hippel et al., 2017). Additionally, research has shown that non-mothers are more likely to be hired and considered for executive positions than mothers (Bygren et al., 2017). For paternal leave, there is evidence that stay-at-home fathers face societal judgement while working fathers are seen as more competent and warmer than childless fathers. However, this support is decreased when fathers choose to actually take advantage of parental leave. Once they decided to take a 12-week leave, fathers were then viewed as more feminine, less altruistic at work, and warmer but less successful, (Fleischmann & Sieverding, 2015). Policies Implemented in Other Countries Many European countries offer lucrative parental leave policies for mothers, but not for fathers. For example, Ireland offers 25 weeks of high paid (75-100%) maternal leave, but zero weeks of paid paternal leave. Italy has a policy that allows mothers 20 weeks of high paid leave, and an additional 26 weeks of low paid (less than 50%) leave, with fathers getting 26 weeks of low paid leave. Many countries also offer transferrable weeks of paid or partially paid leave. France offers almost 140 weeks of low paid leave. Estonia offers over 60 weeks of fully paid, and another 80 weeks of low paid leave. These transferrable weeks of leave can be used by either the mother or the father at the discretion of the organization they work for (Castro-Garca & Pazos-Moran, 2015). Policies Implemented within Companies without Support of Other Employees 4 PERCEPTIONS OF PARENTAL LEAVE POLICIES IN HEALTHCARE SETTINGS 5 With the arguments for parental leave policies gaining support from families across the world, another viewpoint has entered the discussion. A group known as the Childless Network advocates for employees without children. They argue that childless employees work longer hours, take on more projects, and are unable to take advantage of flexible work scheduling that their coworkers with children are sometimes offered (Parker & Allen, 2001). Negative perceptions of parental leave like these are instilled partly by the idea that parenthood is not a requirement. Parenthood is a morally permissible but not morally mandatory choice that persons make to enrich their own lives (Mills, 2001). With this idea that parenthood is not a requirement of adults, childless employees question their obligation to provide for a seemingly optional leave of absence. However, others argue, and the literature seems to support the idea that parental leave policies are a part of providing adequate care for the child during their first crucial months of life, making it a basic human need for the child (Mills, 2001). Benefits of Family-Friendly Programs Family-friendly programs have been found to benefit companies, the health of the parents, and the health of the child. In addition to previously outlined company benefits, women who took leave were 93% more likely to be at work 9-12 months after the childs birth than women who did not take leave (Houser & Vartanian, 2012). Further, allowing mothers to stay home with their newborn for longer allows them to continue breastfeeding, and literature on the subject estimates that non-breastfed infants require $331 to $475 in additional medical costs when compared to breastfed infants. This suggests that companies providing their employees with family-friendly programs can PERCEPTIONS OF PARENTAL LEAVE POLICIES IN HEALTHCARE SETTINGS 6 potentially save companies money in the long-run by ensuring the health of the child during the first few months of life (West, 2015). Additional literature has supported the benefits of family-friendly programs for children. Evidence shows that parental leave policies can lower mortality rates in children and can shorten hospital stays early in the childs life by 31% (National Partnership for Women and Families, 2015). Mothers who utilize family-friendly programs also experience benefits. Literature shows that women who utilize paid maternity leave are 40% less likely to use public assistance in the childs first year of life compared to women who are unable to take leave (Houser & Vartanian, 2012). Fathers also benefit from family-friendly programs, as literature suggests that fathers who take leave are more involved in their childs care than fathers who do not (National Partnership for Women and Families, 2015). Current Proposal The current study aimed to identify how individuals view the integration of paid parental leave policies in healthcare facilities such as hospitals, and how this perception differed based on the kind of information they received. Additionally, the study investigated how perceptions differ between maternal leave policies and paternal leave policies. The perceptions of stereotype threat and the perceptions of negative career outcomes were also assessed based on random placement in one of three groups. One group received no information, one group received an anecdotal passage regarding parental leave policies, and one group received a fact sheet regarding parental leave policies. PERCEPTIONS OF PARENTAL LEAVE POLICIES IN HEALTHCARE SETTINGS Method Study Approach/Design This study used a survey design, with two experimental conditions and a control condition. The control condition provided no information about parental leave to the participant, the first experimental condition used a fact sheet on parental leave policies, and the second experimental condition provided an anecdotal passage regarding parental leave policies. Participants In order to be eligible to participate in this study, participants had to be students from the University of Indianapolis (UIndy) who were at least 18 years old. They were recruited through SONA Systems and through flyers distributed on campus at UIndy. Data was collected from 101 participants for this study. Figure 1 provides a power analysis for the main effects of gender, and Figure 2 provides a power analysis for the main effects of information type. Assuming a medium effect size with 80% power, 158 participants were needed for this study. Figure 1: Power Analysis for Main Effect of Gender 7 PERCEPTIONS OF PARENTAL LEAVE POLICIES IN HEALTHCARE SETTINGS 8 Figure 2: Power Analysis for Main Effect of Information Type The participants in this study were 91.1% female (n=92) and 8.9% male (n=9). Ninety-five percent of participants were 18-24 years old (n=96), 4% were 25-34 (n=4), and 1% was 35-44 (n=1). Most of the participants in this study were White (88.2%, n=83), followed by Black or African American (7.9%, n=8), then Hispanic or Latino (4%, n=4), Asian or Pacific Islander (3%, n=3), Other (2%, n=2), and Native American or American Indian (1%, n=1). Of those who responded to this question, 5% of participants reported that they had at least one child (n=5), and 95% reported that they did not have any children (n=95). One participant did not respond to this question. See Tables 1-4 for this information. Table 1: Gender Demographics PERCEPTIONS OF PARENTAL LEAVE POLICIES IN HEALTHCARE SETTINGS Table 2: Age Demographics Table 3: Race/Ethnicity Demographics Table 4: Parental Status Demographics 9 PERCEPTIONS OF PARENTAL LEAVE POLICIES IN HEALTHCARE SETTINGS 10 Materials/Measures 1. Demographic The first set of questions was asked before any manipulation was presented and consisted of demographic questions such as age, gender, race/ethnicity, education level, and if the participant has children or not, because the latter could have been a factor in how participants responded. 2. Manipulations After the demographic questions, participants were provided with one of three conditions: no information, a fact sheet, or an anecdotal passage. Following the completion of the randomized condition, all participants were given the same series of measures to gather their perceptions of stereotype threat in the workplace based on gender and parenthood, perceived career outcomes due to their utilization of familyfriendly programs, and the likelihood of their utilization of family-friendly programs in the future. Following the manipulations, the same series of measures was presented to every participant, regardless of the group in which they were randomly placed. The components of the questionnaire are discussed in the sections below. Thirty-three point seven percent of the participants were assigned to the control condition (n=34), 31.7% were assigned to receive the anecdotal passage (n=32), and 34.7% were assigned to receive the fact sheet (n=35). 3. Stereotype Threat A measure was used to determine stereotype threat against mothers and fathers in the workplace. Participants were asked to rate their agreement with statements such as PERCEPTIONS OF PARENTAL LEAVE POLICIES IN HEALTHCARE SETTINGS 11 Women who are mothers have less managerial ability than women who arent mothers, on a Likert scale from one (strongly disagree) to five (strongly agree). The same questions were also asked regarding fathers. There were 32 items for this measure, with 18 items adapted from von Hippel et al. (2017) and 10 adapted from Thompson et al. (1999). An example question was, I would be frustrated if I had to pick up extra shifts or take on new patients at the hospital because a coworker was taking a leave of absence due to family-related reasons, and four items that were created for this study (e.g. Companies should be obligated to offer family-friendly policies to their employees who are fathers). The questions obtained for this measure were from previous studies and were provided to me via personal communication by an author of one of the studies, Dr. Courtney von Hippel or were made available in the 1999 study by Thompson et al. These questions were designed to determine if the participants felt stereotype threat towards parents who utilize family friendly programs. Permission was received to reproduce these questions in this study (see Appendix A for the complete inventory). The Cronbachs alpha for this measure is .946 (see Table 5). Table 5: Reliability of Stereotype Threat Measure 4. Perceived career outcomes A measure with 10 items to test for perceived career outcomes was also utilized. Eight items were adapted from Butler et al. (2004) including, Using family-friendly PERCEPTIONS OF PARENTAL LEAVE POLICIES IN HEALTHCARE SETTINGS 12 policies in the workplace should result in negative career outcomes, and two items were created for this study (e.g. People who utilize family-friendly policies should be able to feel like their job is waiting for them when they return). The questions obtained for this measure were from a previous study and were provided to me via personal communication by one of the authors of that study, Dr. Adam Butler. They were designed to determine whether participants predict a negative career outcome as a result of the utilization of family-friendly programs. Permission was received to reproduce these questions in this study (see Appendix A for the complete inventory). The Cronbachs alpha for this measure is .856 (see Table 6). Table 6: Reliability of Perceived Career Outcomes Measure 5. Utilization of family-friendly programs A measure consisting of six items was used to test for perceived utilization of familyfriendly programs. Five items were replicated from Butler et al. (2004) including, Leave of absence), and one was replicated from von Hippel et al. (2017) including, Unpaid leave. The questions obtained for this measure were from a previous study and were provided to me via personal communication by one of the authors of that study, Dr. Adam Butler. They were designed to determine if stereotype threat and negative perceived career outcomes impact peoples utilization of family-friendly programs. PERCEPTIONS OF PARENTAL LEAVE POLICIES IN HEALTHCARE SETTINGS 13 Permission was received to reproduce these questions in this study (see Appendix A for the complete inventory). The Cronbachs alpha for this measure is .708 (see Table 7). Table 7: Reliability of Utilization Measure Procedures This study was presented to participants through an online survey. Prior to allowing participants to complete the study, they were required to provide their informed consent to ensure that they understood the risks and benefits of their participation. No identifying information was gathered from the participants. Qualtrics randomly assigned participants to one of three conditions: one group received an anecdotal passage from a parent who would benefit from family-friendly programs; one group received a fact sheet on the benefits and successes of family-friendly programs in another country; and one group received no information at all. The group serving as the control condition received no information and was asked about their perceptions through filling out the questionnaire explained in the Materials/Measures section. The first experimental group was presented with statistical information about the benefits of maternal and paternal leave and was then given the same questionnaire as the control group. The final experimental group was presented with anecdotal information about maternal and paternal leave and was given the same questionnaire as the other two groups (see Appendix A for the complete inventory). PERCEPTIONS OF PARENTAL LEAVE POLICIES IN HEALTHCARE SETTINGS 14 Participants with inconsistent data were excluded from analyses. Attention checks were utilized throughout to ensure participants were paying attention. For example, one question was scored on a scale of one to five, with one being a low level of support for family-friendly programs, and another question was scored with one being a high level of support for family-friendly programs, depending on the wording of the questions. Some of the questions were specific to maternity or paternity leave, so to counterbalance any bias by answering a certain set of questions first, I randomly assigned participants to receive the maternity leave questions first, and some to receive the paternity leave questions first. To conclude my study, participants were given the option to request research credit for their time, and they were also given the option to enter their names in a drawing for one of four $25 Amazon gift cards. The four recipients of the gift cards were notified via email, and this is also how they received their gift cards. Participants also received a debriefing document explaining that the purpose of the study was to investigate the different responses depending on what type of information was presented to them. I then thanked them for their participation. The entire study was designed to take approximately 15 minutes for each individual to complete. Qualtrics was used to handle the storing of my data and ensure confidentiality, and all questionnaires were completed using this program. Data Analysis Plan To analyze the data, One-Way ANOVA was used to determine whether there was a main effect of the type of information participants were given, as well as a main effect PERCEPTIONS OF PARENTAL LEAVE POLICIES IN HEALTHCARE SETTINGS 15 of gender. A One-Way ANOVA allowed for the analysis of significance between groups. I predicted that female participants would have more positive perceptions of familyfriendly programs across all three groups and would be more likely to utilize available resources, and that the male participants in the fact sheet group would have more positive perceptions compared to the other male participants. I also predicted that females would report higher levels of stereotype threat and negative perceived career outcomes, similar to the findings of Butler et al. (2004). The confirmation of these hypotheses would provide insight into what methods are more effective for informing individuals about the benefits of family-friendly policies in the workplace, specifically healthcare settings, and how stereotype threat can be limited to encourage mothers and fathers to utilize available programs. PERCEPTIONS OF PARENTAL LEAVE POLICIES IN HEALTHCARE SETTINGS 16 Results After running a One-Way ANOVA, results showed no significant difference between the experimental conditions for utilization of family-friendly programs (F(2,98)=.734, p=.483), perceived career outcomes (F(2,98)=.585, p=.559), or stereotype threat (F(2,98)=.812, p=.447) (see Table 11). A One-Way ANOVA was also run to test for gender differences, and the assumption of homogeneity was met (see Table 8). Table 9 outlines the results from the test of normality, in which the only significant results were for females in the career measure, and female in the stereotype threat measure. There was a significant difference between male and female participants for utilization of familyfriendly programs (F(2,98)=5.837, p=.018) and stereotype threat (F(2,98)=4.639, p=.034) (see Table 11). There was not a significant difference between male and female participants for perceived career outcomes (F(2,98)=.390, p=.534) (see Table 11). Table 8: Test of Homogeneity for ANOVA for Gender Differences Table 9: Test of Normality for ANOVA for Gender Differences PERCEPTIONS OF PARENTAL LEAVE POLICIES IN HEALTHCARE SETTINGS Table 10: ANOVA for Differences in Condition Table 11: ANOVA for Differences in Gender 17 PERCEPTIONS OF PARENTAL LEAVE POLICIES IN HEALTHCARE SETTINGS 18 Discussion For this study, I predicted that female participants would have more positive perceptions of family-friendly programs across all three groupsthe control group, the group that received an anecdotal passage, and the group that received the fact sheetand would be more likely to utilize available resources. Additionally, I predicted that the male participants in the fact sheet group would have more positive perceptions compared to other male participants. I also predicted that female participants would report higher levels of stereotype threat and negative perceived career outcomes. Results showed no significant differences between the experimental conditions. Although there were not enough male participants in this study to make a formal analysis, there was still a significant difference between male and female participants in regards to their likelihood of utilization and stereotype threat. Limitations One limitation of this study lies within the data themselves. The design of this study was to compare female and male participants, and this comparison would have been more appropriate with an equal number of male and female participants. This limitation may have been due to the fact that the University of Indianapolis has more females in their student population than males. Additionally, more participants who completed the survey were recruited through SONAa system used by students enrolled in Psychology coursesrather than the posted flyers, and the gender distribution is even more heavily female in the Psychology major. This limitation could be reduced by PERCEPTIONS OF PARENTAL LEAVE POLICIES IN HEALTHCARE SETTINGS 19 widening the scope to other universities in order to get a more equal distribution of male and female participants. Another limitation was the short time frame that was available to complete data collection. More time for data collection would have allowed for the recruitment of the number of participants recommended by the power analysis. The power analysis required 158 participants with 80% power, and 101 participants were recruited. If this study were to be replicated with the recommended 158 participants, the results could reveal different relationships than what was found in this one. However, that cannot be said with certainty. Broader Impacts These findings have implications for future research on attitudes toward parental leave policies. Results from this study could indicate the effectiveness of certain types of education regarding parental leave and its importance. One goal of this study was to educate the participants on the benefits of familyfriendly policies in the workplace. Studies like this are necessary because the targeted participantscollege studentswill be entering the workforce very soon and will be impacted by these policies, or lack thereof, should they choose to start a family. Research with college students as participants is important for obtaining information about future generations. College students will soon be choosing whether or not they will utilize programs offered by their future workplace, so it is very important that they are informed on the costs and benefits that come with them, which is what studies like this hope to achieve. PERCEPTIONS OF PARENTAL LEAVE POLICIES IN HEALTHCARE SETTINGS 20 In addition to college students, the populations that these studies should cover could also include members of management teams in companies. Research in that area would be beneficial in further confirming both the effectiveness of family-friendly policies if a company has them. It would also be beneficial to evaluate the reasons why companies choose not to offer paid leave policies, so that education can be adjusted to those attitudes. Evidence has shown that family-friendly policies increase employee retention rates and job satisfaction (West, 2015; Solheid, 2016; Thompson et al., 1999). As highlighted in the literature, many European countries such as France, Estonia, Italy, and Ireland have started to embrace these policies in order to better the lives of workers and their children (Castro-Garca & Pazos-Moran, 2015). Unfortunately, the United States has very limited legislation on this matter. The only actual legislation requiring companies to provide leave for their employees is the Family Medical Leave Act, which provides twelve weeks of unpaid leave to employees who have worked at a company with 50 or more employees for at least one year and a minimum of 1,250 hours. These twelve weeks include all types of leave: sick leave, maternity leave, paternity leave, and leave to take care of family for any reason (FMLA, 2016). In order to bring attention to this, education is necessary, and studies like this need to continue for that to happen. The stereotype measure alone can lead to more research as well. It is well documented that females have been shown to experience stereotype threat in management positions, as well as in general in a range of fields (Heilman & Okimoto, 2007; Sarsons, 2017). Studies on stereotype threat regarding a multitude of topics such as PERCEPTIONS OF PARENTAL LEAVE POLICIES IN HEALTHCARE SETTINGS 21 maternal leave and paternal leave can indicate ways in which our country can improve in terms of negative perceptions of positive legislation. Many individuals, such as the members of The Childless Networkwhich advocates for the rights of employees without childrenhave valid concerns about family-friendly policies. Some of these concerns include the possibility of other employees taking on more work as a result, or that individuals should not be paid for time that they are not there. While these concerns are understandable and important to explore, there is evidence that family-friendly policies are beneficial. There is evidence that these policies benefit not only the parents, but the children, as parental leave has been shown to decrease the length of hospital visits in the infants first year by 31%, and has also shown to lower mortality rates in children (National Partnership for Women and Families, 2015). This evidence creates a situation where parental leave is not only a workplace issue, but also an issue of basic human rights, as it can aid in the childs development early in life. Overall, research like this is important to the progression of society because it allows us to take a look at how we view issues like parental leave, and how we can use this information to educate and make effective change in corporate policies. In this way, we can work to better the lives of workers and their families by assisting in the balance of work and family. PERCEPTIONS OF PARENTAL LEAVE POLICIES IN HEALTHCARE SETTINGS 22 Conclusion The purpose of this study was to investigate the impact that information has on student perceptions of family-friendly policies. Participants were randomly assigned to a group and received either an anecdotal passage, a fact sheet, or no information. This study found gender differences in perceptions of their potential to utilize family-friendly programs as well as stereotype threat. Specifically, female participants were more likely to utilize family-friendly programs such as paid leave, and they also felt stereotype threat more than male participants. However, this study should be replicated with a similar population on a wider scale due to the low number of male participants (n=9), and the low number of participants overall (N=101). PERCEPTIONS OF PARENTAL LEAVE POLICIES IN HEALTHCARE SETTINGS 23 Reflection/Future Directions Although some of the studies mentioned in this paper are several years old, they still hold a significant amount of weight in terms of what has been done, what continues to be done, and what needs to be done. Later research supported the study done by Thompson et al. (1999), which is why it is mentioned in this paper as an important piece of the framework. Future studies could widen the scope to many universities instead of focusing on just one. This would allow for a more diverse sample that would be more conducive to evaluating gender differences as well. The findings of this study can be helpful in determining the current perceptions of students who will soon be entering the workforce and potentially taking advantage of family-friendly programs offered by their future places of business. In addition, future studies can look into the reasons why companies decide not to implement policies beyond what is legally required of them, despite the evidence that family-friendly programs benefit the employer, employee, and the child. Gathering information such as this can be helpful in indicating how attitudes towards this issue can impact implementation by companies and utilization by employees. Due to the time constraints and limitations regarding my recruitment methods, it was difficult to obtain an appropriate number of participants for this study. However, the process of conducting this study gave me the opportunity to apply research skills that will help me in the future, both in graduate school and in my career as an occupational therapist. These skills that I have acquired in data collection, data analysis, and every step of the formation of this paper have given me the knowledge to pursue research PERCEPTIONS OF PARENTAL LEAVE POLICIES IN HEALTHCARE SETTINGS 24 opportunities in OT school and beyond in professional settings. This project was personally important to me because it gave me the opportunity to educate my peers about the current state of parental leave policies in the workplace, and it allowed me to learn about their perceptions as well. PERCEPTIONS OF PARENTAL LEAVE POLICIES IN HEALTHCARE SETTINGS 25 References Butler, A., Gasser, M., & Smart, L. (2004). A Social-Cognitive Perspective on Using Family-Friendly Benefits. Journal of Vocational Behavior, 65(1), 5770., doi:10.1016/s0001-8791(03)00097-6. Bygren, M., Erlandsson, A., & Ghler, M. (2017). Do Employers Prefer Fathers? Evidence from a Field Experiment Testing the Gender by Parenthood Interaction Effect on Callbacks to Job Applications. European Sociological Review, 33(3), 337-348. doi:10.1093/esr/jcx051 Castro-Garca, C., and Pazos-Moran, M. (2015). Parental Leave Policy and Gender Equality in Europe. Feminist Economics, 22(3), 5173., doi:10.1080/13545701.2015.1082033. Dagher, R. K., McGovern, P. M., & Dowd, B. E. (2013). Maternity leave duration and postpartum mental and physical health: Implications for leave policies. Journal of Health Politics, Policy and Law, 39(2), 369-416. doi:10.1215/03616878-2416247 Fleischmann, A., and Sieverding, M. (2015). Reactions Toward Men Who Have Taken Parental Leave: Does the Length of Parental Leave Matter? Sex Roles, 72(9-10), 462476., doi:10.1007/s11199-015-0469-x. FMLA (Family & Medical Leave). (2016, May 06). Retrieved April 09, 2018, from https://www.dol.gov/general/topic/benefits-leave/fmla Heilman, M.E., and Okimoto, T.G. (2007). Why Are Women Penalized for Success at Male Tasks?: The Implied Communality Deficit. Journal of Applied Psychology, 92(1), 8192., doi:10.1037/0021-9010.92.1.81. Houser, L., & Vartanian, T. (2012). Pay Matters: The Positive Economic Impacts of Paid Family Leave for Families, Businesses and the Public. Rutgers Center for Women and Work,1-17. Manchester, C., Leslie, L., & Park, T.-Y. (2008). Screening for commitment: The effect of maternity leave usage on wages (Technical Report). Minneapolis, MD: University of Minnesota. Mills, C. (2001). Workplace Wars: How Much Should I Be Required to Meet the Needs of Your Children? Philosophy and Public Policy Quarterly, 21(1), 1520. National Partnership for Women and Families. (2015, March). Children Benefit When Parents Have Access to Paid Leave. PERCEPTIONS OF PARENTAL LEAVE POLICIES IN HEALTHCARE SETTINGS 26 Parker, L., and Allen, T.D. (2001). Work/Family Benefits: Variables Related to Employees' Fairness Perceptions. Journal of Vocational Behavior, 58(3), 453 468., doi:10.1006/jvbe.2000.1773. Sarsons, H. (2017). Interpreting Signals in the Labor Market: Evidence from Medical Referrals. Harvard University. Solheid, M. (2016). Do Parental Leave Policies Affect Employee Engagement at Firms in the United States?. University of Minnesota. Thompson, C.A., Beauvais, L.L., & Lyness, K.S. (1999).When Work-Family Benefits Are Not Enough: The Influence of Work-Family Culture on Benefit Utilization, Organizational Attachment, and Work-Family Conflict. Journal of Vocational Behavior, 54(3), 392415., doi:10.1006/jvbe.1998.1681. von Hippel, C., Kalokerinos, E. K., & Zacher, H. (2017). Stereotype Threat and Perceptions of Family-Friendly Policies among Female Employees. Frontiers in Psychology, 7. doi:10.3389/fpsyg.2016.02043 West, K. J. (2015). Is the U.S. Missing Out? The Effects of Parental Leave Benefits on Business. Undergraduate Review, 11, 120-122. PERCEPTIONS OF PARENTAL LEAVE POLICIES IN HEALTHCARE SETTINGS 27 Appendices Appendix A: Inventories Italicized questions were written specifically for this study. Please indicate your agreement with the following statements on a scale of 1 to 5. 1 strongly disagree 2 disagree 3 neutral 4 agree 5 strongly agree 1. Using family-friendly policies in the workplace should result in negative career outcomes. 2. Using family-friendly programs should hurt ones status at work. 3. Using family-friendly programs shows that one is capable of balancing work and family. 4. Using family-friendly programs improves the lives of the individuals child(ren). 5. People who utilize family-friendly policies should be paid less than those who dont. 6. People who utilize family-friendly programs will not be able to meet the demands of their job like someone who does not utilize such programs. 7. People who utilize family-friendly programs are not able to succeed in their jobs and at home. 8. People who utilize family-friendly programs find a way to balance their work and home life. 9. People who utilize family-friendly policies should be able to feel like their job is waiting for them when they return. 10. Offering family-friendly policies should be the standard in the workplace. Below is a list of different family-friendly programs that are offered at some companies. Please indicate how likely you are to use these programs if they were offered at your job/future job. 11. Leave of absence 1 very unlikely 2 unlikely 3 neutral 4 likely 5 very likely 2 unlikely 3 neutral 4 likely 5 very likely 2 3 4 5 12. Flexible scheduling 1 very unlikely 13. On-site daycare 1 PERCEPTIONS OF PARENTAL LEAVE POLICIES IN HEALTHCARE SETTINGS very unlikely unlikely 28 neutral likely very likely 2 unlikely 3 neutral 4 likely 5 very likely 2 unlikely 3 neutral 4 likely 5 very likely 14. Telecommuting/Working from home 1 very unlikely 15. Unpaid leave 1 very unlikely 16. Job-sharing (two part-time employees completing the work of one full-time employee) 1 very unlikely 2 unlikely 3 neutral 4 likely 5 very likely Butler et al. (2004), von Hippel et al. (2017) Please indicate your agreement with the following statements regarding maternal leave. 1 strongly disagree 2 disagree 3 neutral 4 agree 5 strongly agree 17. Women who are mothers have less managerial ability than women who arent mothers. 18. Women who are mothers are less committed to their careers as women who arent mothers. 19. Women who are mothers are limited in how high they can advance in their careers. 20. Women who are mothers have the same abilities in the workplace as women who arent mothers. 21. Women who are mothers have less to contribute in the workplace than women who arent mothers. 22. Women who are mothers and women who arent should be paid the same amount for the same work. 23. Whether or not a woman is a mother should influence her career growth. 24. Using family-friendly programs should have no impact on how a woman grows with a company. 25. Using family-friendly programs suggests that a mother is not as serious about her career as employees who dont use these programs. 26. Companies should be obligated to offer family-friendly programs to their employees who are mothers. PERCEPTIONS OF PARENTAL LEAVE POLICIES IN HEALTHCARE SETTINGS 29 27. Companies should have the right to not consider women who are mothers for promotions because they are mothers. Please indicate your agreement with the following statements regarding paternal leave. 1 strongly disagree 2 disagree 3 neutral 4 agree 5 strongly agree 28. Men who are fathers have less managerial ability than men who arent fathers. 29. Men who are fathers are less committed to their careers as men who arent fathers. 30. Men who are fathers are limited in how high they can advance in their careers. 31. Men who are fathers have the same abilities in the workplace as men who arent fathers. 32. Men who are fathers have less to contribute in the workplace than men who arent fathers. 33. Men who are fathers and men who arent should be paid the same amount for the same work. 34. Whether or not a man is a father should influence his career growth. 35. Using family-friendly programs should have no impact on how a man grows with a company. 36. Using family-friendly programs suggests that a father is not as serious about his career as employees who dont use these programs. 37. Companies should be obligated to offer family-friendly to their employees who are fathers. 38. Companies should have the right to not consider men who are fathers for promotions because they are fathers. von Hippel et al. (2011) Imagine yourself as the head of a successful hospital. You are responsible for hiring, firing, and promoting doctors and other medical personnel. As the head, please rate how likely you are to have the following opinions or make the following decisions for the benefit of your company as a whole. 39. I would view medical personnel that put their patients before their family more favorably than those who put their family first. 1 very unlikely 2 unlikely 3 neutral 4 likely 5 very likely PERCEPTIONS OF PARENTAL LEAVE POLICIES IN HEALTHCARE SETTINGS 30 40. I would have the opinion that medical personnel who utilize family-friendly programs are less likely to advance in their careers than medical personnel who dont. 1 very unlikely 2 unlikely 3 neutral 4 likely 5 very likely 41. I would expect medical personnel to put their jobs before their families and other obligations to maximize productivity. 1 very unlikely 2 unlikely 3 neutral 4 likely 5 very likely 42. If an employee turns down a promotion for family-related reasons, I would not want to offer them promotions in the future. 1 very unlikely 2 unlikely 3 neutral 4 likely 5 very likely 43. I would understand if employees had to leave work to take care of family matters. 1 very unlikely 2 unlikely 3 neutral 4 likely 5 very likely 44. I would view employees who work part-time by choice as less committed than full-time employees. 1 very unlikely 2 unlikely 3 neutral 4 likely 5 very likely 45. I would be sensitive to employees family matters, even if it might cost the company money. 1 very unlikely 2 unlikely 3 neutral 4 likely 5 very likely Respond to the next questions as if you were a coworker to an individual who utilizes family-friendly programs. 46. As a fellow health professional at a hospital, I would resent a father who takes extended leaves to care for a newborn or adopted child. 1 very unlikely 2 unlikely 3 neutral 4 likely 5 very likely PERCEPTIONS OF PARENTAL LEAVE POLICIES IN HEALTHCARE SETTINGS 31 47. As a fellow health professional at a company, I would resent a mother who takes extended leaves to care for a newborn or adopted child. 1 very unlikely 2 unlikely 3 neutral 4 likely 5 very likely 48. I would be frustrated if I had to pick up extra shifts or take on new patients at the hospital because a coworker was taking a leave of absence due to family-related reasons. 1 very unlikely 2 unlikely 3 neutral 4 likely Thompson et al. (1999) Appendix B: CITI Training Completion Date 18-Sep-2017 Expiration Date 18-Sep-2019 Record ID 24427289 This is to certify that: Sara Skarshaug Has completed the following CITI Program course: Human Subjects Research (HSR) Group 2: Health Related Research 1 - Basic Course (Curriculum Group) (Course Learner Group) (Stage) Under requirements set by: University of Indianapolis Verify at www.citiprogram.org/verify/?wd12aff55-7575-42cb-be72-52dab091dda8-24427289 5 very likely PERCEPTIONS OF PARENTAL LEAVE POLICIES IN HEALTHCARE SETTINGS 32 Completion Date 25-Sep-2018 Expiration Date 24-Sep-2023 Record ID 28847678 This is to certify that: Sara Skarshaug Has completed the following CITI Program course: Responsible Conduct of Research (RCR) Group 8: Responsible Conduct of Research 1 - Basic Course (Curriculum Group) (Course Learner Group) (Stage) Under requirements set by: University of Indianapolis Verify at www.citiprogram.org/verify/?wc015b8e5-8122-4012-a4d8-9c03de3fac87-28847678 Appendix C: IRB Approval Dear Dr. Hess, The UIndy HRPP has reviewed and approved exempt status for your new study titled, Perceptions of Parental Leave Policies in Healthcare Settings" (UIndy Exempt Study #0957). I have attached to this email the official approval certification letter. For your convenience, I have also attached the following files: Word copy of the clean ICF and a pdf copy of the APPROVED ICF. Please be sure to update the entire ICF on your Qualtrics questionnaire with the finalized ICF, including the header. I have also attached copies of the approved recruitment materials. Please feel free to edit the formatting but please do not change the content. If you have any questions, please let me know. Sincerely, Erin Fekete Associate Professor of Psychology Director of Psychological Sciences Faculty Co-Chair, Human Research Protection Program College of Applied Behavioral Sciences University of Indianapolis PERCEPTIONS OF PARENTAL LEAVE POLICIES IN HEALTHCARE SETTINGS Appendix D: Recruitment Flyer Participate in a UIndy Research Study The purpose of this study is to gather information about student perceptions of parental leave policies. After completing the study, you will have the chance to enter a drawing for one of four $25 Amazon gift cards. Please go to https://uindy.co1.qualtrics.com/jfe/form/SV_6fdoTqMRyg2zjdX to participate in the study! Go to https://uindy.co1.qual trics.com/jfe/form/SV _6fdoTqMRyg2zjdX Go to https://uindy.co1.qual trics.com/jfe/form/SV _6fdoTqMRyg2zjdX Go to https://uindy.co1.qual trics.com/jfe/form/SV _6fdoTqMRyg2zjdX Go to https://uindy.co1.qual trics.com/jfe/form/SV _6fdoTqMRyg2zjdX Go to https://uindy.co1.qual trics.com/jfe/form/SV _6fdoTqMRyg2zjdX Go to https://uindy.co1.qual trics.com/jfe/form/SV _6fdoTqMRyg2zjdX Go to https://uindy.co1.qual trics.com/jfe/form/SV _6fdoTqMRyg2zjdX Go to https://uindy.co1.qual trics.com/jfe/form/SV _6fdoTqMRyg2zjdX Go to This study has been approved by the UIndy Human Subjects Protections Program (Exempt UIndy Study #0957; Approval Date: December 17, 2018). https://uindy.co1.qual trics.com/jfe/form/SV _6fdoTqMRyg2zjdX Go to https://uindy.co1.qual trics.com/jfe/form/SV _6fdoTqMRyg2zjdX 33  ...

... VALUES IN 4 PARTS: Music and text bound by the limitless constraints of the Oulipo By Lucy Shirley Honors Project Thesis Submitted in Partial Fulfillment Of the Requirements for a Baccalaureate Degree With Distinction In the Ron and Laura Strain Honors College of THE UNIVERSITY OF INDIANAPOLIS May 2020 Faculty Advisor: John Berners, Ph.D. Executive Director of Honors: James B. Williams, Ph.D. Lucy Shirley, 2020. Copyright in this work rests solely with the author. Please ensure that any reproduction or re-use is done in accordance with relevant national copyright protection. Acknowledgements I would like to express my deepest gratitude to my advisor, Dr. John Berners, who first introduced me to the Oulipo and their philosophies and guided me every step of the way. Dr. Berners has been instrumental in my compositional growth as a music major, working with me closely through weekly lessons and helping me craft this idea of a cross-disciplinary melding of literature, language, music, and the methods of organization favored by the Oulipo. Thank you to Gavin Craig for his enthusiasm, his indispensable counsel regarding composing for the saxophone, and for his willingness to perform the final product. Thank you to the Strain Honors College for giving me the opportunity to share creative work in an honors platform. Thank you to Dr. Sharon Parr for her mentorship in all aspects of artistic growth. Thank you to my family and friends for their love and support. i Table of Contents Acknowledgements .......................................................................................................................... i Table of Contents ............................................................................................................................ ii Abstract .......................................................................................................................................... iii Introduction ......................................................................................................................................1 Project Development ........................................................................................................................4 Performance Notes ...........................................................................................................................4 Score ................................................................................................................................................5 Program Notes ...............................................................................................................................12 Conclusion .....................................................................................................................................20 Bibliography ..................................................................................................................................21 Vita.................................................................................................................................................22 ii Abstract Values in 4 Parts aims to explore the vast potential of language and music and the relationships therein through the methods laid out by the French literary group the Oulipo (French acronym for The Workshop of Potential Literature). Values is a collection of voice and saxophone studies in language and expression, based on the principles set forth by Raymond Queneau and primarily employed by the French literary group the Oulipo. Each of the four movements of Values contains a poem or a fragment of text that has been manipulated by or interpreted with a certain Oulipian technique or principle. The technique is then translated into a musical process and paired with the text to create a set of self-contained studies in verbal and musical language. NOTE: Audio of Values in 4 Parts is available online: Values in 4 Parts by Lucy Shirley. Featuring Lucy Shirley, soprano, and Gavin Craig, saxophone. Recorded November 2019. https://www.youtube.com/watch?v=AUBbOS5B6jI iii Introduction Throughout history, philosophers have posed questions of whether artistic creation takes place primarily as a result of spontaneous inspiration or primarily through measured technique1 and whether creativity flourishes most under the context of creative freedom or creative constraint. The French literary group the Oulipo was created in 1960 by Franois Le Lionnais and Raymond Queneau2 as an anti-Modernist literary and mathematical workshop to answer such a question.3 Still operating today, the Oulipo seeks to redefine the structural confines of creativity,4 not through reducing structure as one might expect, but through increasing it to the utmost degree in an argument that rules can stimulate better writing than the absence thereof.5 Creativity, they argue, is able to expand under the binds of formality in a way it is unable to do without structure. In eliminating random spontaneity and tumultuous bouts of inspiration,6 the many writers of the Oulipo construct a world in which various specific, personally-imposed rules guide creative choices. Recent cognitive research shows that placing certain constraints on creative processes does not stifle creativity as one might expect, but in fact stimulates greater quantities of creative activity, a finding consistent with the claims of the Oulipos members.7 In limiting oneself to Chris Andrews. "Inspiration and the Oulipo, Studies in 20th & 21st Century Literature: Vol. 29: Iss. 1, Article 2 (2005): 1-2. https://doi.org/10.4148/2334-4415.1590. 2 All That is Evident is Suspect: Readings from the Oulipo: 1963-2017. Edited by Monk, Ian, and Daniel Levin Becker. (San Francisco: McSweeneys, 2018), 7. 3 Daniel Levin Becker. Many Subtle Channels: In Praise of Potential Literature. (Cambridge: Harvard University Press, 2012), 7. 4 All That is Evident, Monk, 7. 5 Levin Becker, Many Subtle Channels, 14. 6 James, Alison. Constraining Chance: Georges Perec and the Oulipo. (Chicago: Northwestern University Press, 2009), 14. 7 Colin Symes. "Writing by Numbers: OuLiPo and the Creativity of Constraints." Mosaic: An Interdisciplinary Critical Journal 32, no. 3 (1999): 87-107. http://www.jstor.org/stable/44029802. 1 1 specific rules governing the process of creation (such as writing a novel without using the letter e, as in Ernest Vincent Wrights Gadsby or Oulipo member Georges Perecs La Disparition), creative potential is exponentially increased. Since its birth, the Oulipo has numbered forty-one members across the globe (although, as member Paul Fournel explains, half have been excused from group activities on the grounds of being deceased)8 and has spawned over a dozen more semi-official sister groups for other artforms such as painting, filmmaking, etc. Franois de Lionnais, one of the original founders of the Oulipo, originally envisioned the principles of the group extending to a variety of other disciplines, specifically music, and in 2011, an official Oumupo (French acronym for The Workshop of Potential Music) organization was founded.9 Although varied in discipline and medium, each new group bearing the Ouxpo moniker cites the original Oulipo purpose statement of exploration in potentialities for an artform and rejection of purely unpredictable, unruly and spontaneous inspiration as their individual raisons detre. Project Development For this project, I wrote four studies (three containing texts), exploring various aspects of style and meaning in language, society, mathematics, and music. I selected a text by the cofounder of calculus Gottfried Liebniz, another by the 19th century poet Thomas Hood, and wrote one text of my own, then set these texts to music. My goal was to incorporate techniques of the Oulipian movement,10 drawing musical implications from the literary techniques (and in some 8 All That is Evident, Monk, 7. M. Andreatta, M. Granger, T. Johnson and V. Villenave. Music, mathematics and language: chronicles from the Oumupo sandbox. (2017), 1. http://oumupo.org/wiki/images/e/e3/Oumupo_sandbox_chronicles_2017.pdf. 10 Techniques taken from The Oulipo Compendium. Ed. By Mathews, Harry, and Alistair Brotchie. Atlas Press, 2005. 9 2 cases, employing musical techniques favored by the composers of the Oulipos newly-founded sister group, the Oumupo11), resulting in an oeuvre of four musical studies created through the type of restrictions under which the Oulipo operates. The title Values is open to interpretation. The studies in the work focus on the word value in its many different connotations, whether in culture, length of time, numbers and mathematics, color (literal hue or in musical timbre), memory, language, musical and textual preference, confidence, morals, etc. The meanings have continued to evolve throughout the entire process of crafting this project. Values is truly about values and the constraints therein and thereout. 11 Techniques taken from Andreatta, Music, mathematics and language, 3-18. 3 Values in 4 Parts Instrumentation: Soprano voice and Saxophone Duration: c. 8:00 Performance Notes: Regarding the second movement, each letter should be voiced and unpitched consonants should be emphasized before smoothly transitioning to a neutral vowel. (E.g., "T" should be sung "T-t-uh" so that the unvoiced sound may clearly be heard at the onset of the note before transitioning to a sung pitch on a neutral vowel such as "uh." The same may be done for consonants such as "C," "D," "X," "G," "H," and W). Voiced consonants should be sung on pitch. (E.g. "M" should be sung as a closed-mouth "M-m" without the addition of a vowel. The same should be done for "N," "R," and "S" to sound as a "Z"). Vowels should be sung as written. 4 5 6 7 8 9 10 11 Program Notes Part I: A study in beatboxing, allusions, and finishing phrases Forget it, Jake, it's China town, down in an earlier roundoff backhand spring, summer, winter, follow the leader ship packages through Amazon Rain (Forest, run!) in Spain falls mainly on the plain-Jane Seymour's my friends with benefits like a glovebox-ing ring ring ring ring banana phoneCall of Duty. I love rock n' roll with the punch and Judy Bluming Garland of Flowers for Algernon na na na, hey hey hey, Goodbye, Columbus Ohi, Mark ZuckerBurger king in the North Pole dancing Queen Victoria Justice is blind faith without works is dead with a capital D and that rhymes with P and that stands for Postmodernist Critique. Lucy Shirley This first movement is a fun, lighthearted study in references to popular culture and runon sentences. A delightful Oulipian technique is that of a perverb,12 a proverb that has been crossed with another proverb (all roads gather no moss13). In this same vein, I have created a text made out of a chain of paranomastic14 perverbs, so-to-speak. The language blogger Arnold Zwicky coined the term Phrasal overlap portmanteaus15 (or POPs for short) to describe the technique of combining two independent phrases through a pivot word (or in some cases, multiple words). Although not specifically a product of the Oulipo, per se, the method of constructing POPs is extremely Oulipian in practice and is in fact just a more flexible variation of the perverb. For example, the phrases rock and roll and roll with 12 Oulipo Compendium, Mathews, 210. Ibid, 65. 14 Ibid, 209. 15 Zwicky, Arnold. Phrasal overlap portmanteaus. Arnold Zwickys blog. 13 May, 2010. https://arnoldzwicky.org/2010/05/13/phrasal-overlap-portmanteaus/ 13 12 the punches share the word roll in common, so the two phrases could be elided into rock and /roll/ with the punches. The two phrases overlap through the use of their common pivot word roll, creating a paranormastic POP. In this exercise, I wrote continuous streams of POPs, brimming with allusions to pop culture, history, songs, novels, plays, and movies, with every phrase overlapping to create new references. Take, for example, the section halfway through the second stanza: Goodbye, Columbus /Ohi, Mark ZuckerBurger / king in the North. Goodbye, Columbus is the title of a controversial 1959 novel by Phillip Roth. Columbus, Ohio, is simply a cool city. Oh, hi, Mark, is a quote from the 2003 cult classic movie The Room (simply google the phrase oh, hi, mark and examples abound). Mark Zuckerberg is the founder of Facebook and the subject of the Oscar-winning 2010 drama film The Social Network. The fast food restaurant Burger King was founded in Jacksonville, Florida in 1954. King in the North refers to the name given to Robb Stark in the popular series of novels A Game of Thrones, and later in the television adaptation of the same name. Every phrase is separate, yet shares common words with each phrase around it, creating chains of pop-culture POPs. POPs have been employed by many poets and musicians, although I first became acquainted with the technique at a young age through the music of Jason Webley and Amanda Palmer, who often use such overlaps in the lyrics of their songs. For example, in a song called Myspace from their collaborative project Evelyn Evelyn, Webley and Palmer employ this technique across lines of their songs, creating a series of social media-themed POPs in conjunction with poetic enjambment, such as in the lyric When youre looking for a friendly 13 face / book a ticket to another place.16 I desired to emulate such phrasal portmanteaus in the first movement of Values. Once I had finished crafting the text, I began thinking of ways in which I could represent a musical equivalent to the phrasal portmanteaus in the text. The obvious answer was to craft a short, repeated musical line wrought with musical elisions, shifting meters, and phrasal uncertainty, so that the listener is unaware when exactly the line will end and restart. Just as the text seems like a run-on collection of phrases, so, too, does the music feel like a run-on series of motives. The whole text has no overall meaning apart from the collections of meaning present in each individual phrase that makes up the whole, and similarly, the musical line does not have a clear emotional trajectory or climax. To match the modern nature of the text and the references therein, the saxophone serves as a sort of beatboxer, employing extended techniques in the form of pitched and unpitched pops to enhance the rhythmic aspects of the melody as the voice speaks rhythmically and freely above the saxophone, never quite lining up properly with the music and further contributing to a sense of unfinished overlap. Part II: A study in counting, letters, and conversation Music is a hidden exercise in arithmetic where the mind is unaware it is counting. Gottfried Leibniz For the second movement, I took inspiration primarily from the techniques of the Oumupo,17 the musical sibling of the Oulipo. The Oumupo, although currently existing in several iterations, has been instrumental in the exploration of ways to musically signify numbers18 and Evelyn Evelyn. Myspace, 2009, track 11 on Evelyn Evelyn, 11 Records, 2010, LP. Andreatta, Music, mathematics and language, 3-18. 18 Ibid, 4-7 16 17 14 letters.19 Inspired by their use of various numeric bases to generate musical signifiers, I began working through different ways to encode my selected text into the musical line itself. I began by composing the saxophone part. To generate the musical phrases, I took each letter of the sentence and translated it into the standard HTML Unicode UTF-8 (the number system used to signify letters in computer coding). After I had a number to represent each letter, I translated it into a base-4 code (utilizing the values 0, 1, 2, 3). Each letter was then represented by a four-value unit (for example, the letter I now is represented by the value [1021]). I then assigned each value a musical interval: [0] means the pitch of the note must remain the same as the previous note, [1] means the pitch of the note must move up or down (depending on my preference) an interval of one half step from the previous note, [2] means the note must move up or down an interval of one whole step from the previous note, and [3] means the note must move up or down an interval of a minor third from the previous note. Similarly, I assigned each value a length of musical time: [0] means the length must remain the same as the previous note length, [1] means the length must be an eighth note, [2] means the length must be a quarter note, and [3] means the length must be a half note. For example, examine at the saxophone part in the second beat of the fourth measure of the second movement (second system of page 6). The letter I is represented by the base-four value [1021], and four notes represent the musical choices in the four numerals. The first note is an eighth note on F (a [1] means the note must be a half-step higher or lower than the previous note and must be the length of an eight note). The second note is another eighth note on F (a [0] means the pitch must be the same as the previous note, and the length must be the same as the previous notes length). The third note is a quarter note on G (a [2] means the pitch must be a 19 Ibid, 7. 15 whole step higher or lower than the previous note, and the length must be a quarter note). The fourth note is an eighth note on G# (a [1] means the note must be a half-step higher or lower than the previous note and must be the length of an eighth note). So, the value [1021] is represented by a four-note figure, with each note representing a distinct intervallic and temporal choice. Once I had crafted a pleasing musical line for the saxophone, carefully choosing the direction and shape of the line, and often choosing to displace pitches at the octave for a better effect, I turned to composing the vocal line. I decided to use a different system to compose the vocal line, arranging the letters of the text in order of frequency of use. The 12 most commonly used letters in the sentence were assigned the 12 pitches of the octave, with the remaining 4 letters present in the sentence (X, W, O, G) serving as quartertones. Every letter would be represented by one pitch, and the length of the note would be determined by the length of the corresponding figure of the saxophone. For example, when the saxophone plays the representation of the letter I in the section mentioned above, which is the second beat of the fourth measure of the second movement (second system of page 6), the voice sings the letter I directly above for the length of two beats and a half, which is exactly the same length it takes the saxophone to play the four-note figure that represents [1021]. What results is two different, concurrent musical representations of the letters in the given sentence, with the voice further contributing to this effect by phonetically sounding out each individual letter in every word. Parts of the sentence are then spoken in segments by the vocalist and saxophonist, who take turns fragmenting the sentence to enhance understandability and emphasize the idea of counting. 16 Part III: Une tude de style and stuff [Il ny a pas du texte en cet ouvrage pare que laudience a besoin de sentir nouveau laise.] [There isnt any text in this piece because the audience needs to feel comfortable again.] One of the many techniques employed by the Oulipo is a deliberate deviation from imposed constraints: it is called a clinamen.20 A clinamen is a careful choice by a creator to break the rules that they have themselves made on the grounds of aesthetic preference. If somehow the method created works but is not aesthetically pleasing in practice, a clinamen may be employed as a deliberate deviation, but only because the result will be better than if the rules were to be followed without ceasing. This piece has no text, and therefore no Oulipian textual manipulation. I could have added text to this movement, but I chose not to on the grounds of aesthetic preference, employing a clinamen trump card. The prior second movement is extremely unique in nature, and therefore I found it best to insert a text-less, energetic, French-inspired, modern saxophone folk fantasia-etude. This movement, although text-less, is not entirely removed from Oulipian inspiration. The first section is a swirling, bipolar journey through multiple styles of saxophone writing. In its composition I took inspiration from Debussys Rapsodie Orchestra and Alto Saxophone,21 then melded the airy, French figures with the raw, growling music of popular modern saxophonists, several of whom who have achieved great success on the Internet. This blending of old and new saxophone styles seeks to mimic the blending of the old and new ideas of the Oulipo: although sixty years old, the ideas proposed by its members are still relevant to literary discourse. The 20th century French swirling figure finally escapes from the coarser growls and slaps of American popular saxophone writing and develops in the middle section, as 20 21 Oulipo Compendium, Mathews, 125. Debussy, Claude Achille. Rapsodie pour Orchestre et Saxophone. 1903. Paris: Durand & Cie., 1919. Plate D. & F. 9597. 17 the figure is expanded in a rising sequence, floating in a classical halo of lyric melody. The growls soon return, however, and the opening section is repeated. A poignant coda reprises the swirling melody one last time before melting away. The name of the movement, une tude de style and stuff, is a play on Raymond Queneaus seminal 1947 book Exercices de Style22 (trans. Exercises in Style). The pun, then, is in the translation of the word tude: in French, the word tude may be translated as either a study or an exercise (most often used musically, as in Debussys many tudes for piano). The addition of the word stuff is simply an English perversion of the reference. The movement is a deviation from the literary basis of the project, employing a clinamen to instead focus solely on the musical aspects of style. It is a blend of old and new, French and American, classical and popular, just as the Oulipo itself is a blend of analytical constraints and creative unorthodoxy. Part IV: A study in breath, phonation, and silence There is a silence where hath been no sound, There is a silence where no sound may be, In the cold graveunder the deep deep sea, Or in the wide desert where no life is found, Which hath been mute, and still must sleep profound. No voice is hushdno life treads silently, But clouds and cloudy shadows wander free, That never spoke, over the idle ground: But in green ruins, in the desolate walls Of antique palaces, where Man hath been, Though the dun fox, or wild hyena, calls, And owls, that flit continually between, Shriek to the echo, and the low winds moan, There the true Silence is, self-conscious and alone. Silence by Thomas Hood 22 Raymond Queneau. Exercises in Style. Trans. by Barbara Wright. (New York: New Directions, 1981). 18 The last movement relies on silence. The saxophone begins playing without any sound, just the squeaks of the key depressions and the sound of breath into the mouthpiece. The voice begins reciting the poem, yet without employing the vocal chords, so that the only sound that may be heard is air being expelled through the hissing and smacking of consonants like S and C. Slowly, the saxophone begins adding more tone to the repetitive musical figure. The vocalist coughs and clears their throat, as if warming up their voice to likewise produce real tone. Finally, the saxophone suddenly stops before reaching any semblance of a loud volume, and the vocalist whispers to a seemingly empty space the last line of the poem: There the true Silence is, self-conscious and alone. This movement plays with the listeners ability to distinguish the music from the white noise of empty space and challenges them to begin constructing meaning from toneless music and text that is undiscernible. The same music is being repetitively played by the saxophone, yet its true sound is only revealed toward the end of the movement. The vocalist is speaking the text of the poem, but without phonation, the words lack meaning. When the speaker finally is able to produce sound, they are left in silence and the movement ends. Oulipian techniques primarily concern the meaning of language and that meanings potential alteration through certain techniques of manipulation. I was inspired to create a short study in silence that raises these very same ideas. In manipulating text by removing phonation, this final movement explores the question of whether meaning can be transmitted without clarity of sound. 19 Conclusion In defending their methods, many Oulipo members have argued that all art is created following certain rules and structures (for example, the specific number of lines in a Shakespearean sonnet or the traditional rise and fall in the plot of a classic novel).23 They choose, instead of blindly and subconsciously following the rules for art as dictated by society, to craft their own rules, thereby allowing authentic innovation and creativity to flow forth, solving the age-old debate of inspiration vs. technique. They craft inspiration as a result of technique, not because it is the most effective, but because they believe it is the only effective method for true creativity. Creativity cannot exist without boundaries. This project seeks to be an outgrowth of such an idea, experimenting with artistic creation by putting rules and restrictions on the process of musical and textual composition, showing how the walls that can cause creative impediment may be shattered through the very process of constructing new ones. 23 Symes, Writing by Numbers, 89. 20 Bibliography All That is Evident is Suspect: Readings from the Oulipo: 1963-2017. Edited by Monk, Ian, and Daniel Levin Becker. San Francisco: McSweeneys, 2018. Andreatta, M., M. Granger, T. Johnson and V. Villenave. Music, mathematics and language: chronicles from the Oumupo sandbox. (2017). http://oumupo.org/wiki/images/e/e3/Oumupo_sandbox_chronicles_2017.pdf. Andrews, Chris. "Inspiration and the Oulipo, Studies in 20th & 21st Century Literature: Vol. 29: Iss. 1, Article 2 (2005). https://doi.org/10.4148/2334-4415.1590. Debussy, Claude Achille. Rapsodie pour Orchestre et Saxophone. 1903. Paris: Durand & Cie., 1919. Plate D. & F. 9597. https://ks.imslp.net/files/imglnks/usimg/4/40/IMSLP13914Debussy_-_Rapsodie_for_Orchestra_and_Saxophone_(sax._and_piano).pdf Evelyn Evelyn. Myspace, 2009, track 11 on Evelyn Evelyn, 11 Records, 2010, LP. James, Alison. Constraining Chance: Georges Perec and the Oulipo. Evanston: Northwestern University Press, 2009. Levin Becker, Daniel. Many Subtle Channels: In Praise of Potential Literature. Cambridge: Harvard University Press, 2012. Matthews, Harry, and Raymond Queneau. The Oulipo Compendium. Atlas Press, 2005. Queneau, Raymond. Exercises in Style. Translated by Barbara Wright. New York: New Directions, 1981. Symes, Colin. "Writing by Numbers: OuLiPo and the Creativity of Constraints." Mosaic: An Interdisciplinary Critical Journal 32, no. 3 (1999): 87-107. http://www.jstor.org/stable/44029802. Zwicky, Arnold. Phrasal overlap portmanteaus. Arnold Zwickys blog. 13 May, 2010. https://arnoldzwicky.org/2010/05/13/phrasal-overlap-portmanteaus/ 21 VITA Lucy Shirley is a composer, pianist, and teacher based in Indianapolis, Indiana. She has worked with composers such as John Berners, Christopher Dietz, Chen Yi, Douglas Knehans, and Anthony Plog. In 2019, she won 1st place in the University of Indianapolis Concerto Competition and 2nd prize in the Charles Joray Piano Competition. She has presented research at musicology conferences throughout the Midwest and was named a 2019 Shaheen Leadership Fellow for her scholarly work. Her compositions have been played by university and community ensembles in the greater Indianapolis area, and she has participated in festivals and workshops across the nation. A lover of poetry and literature as well as music, Lucys works have been praised for their quirky neo-classicism and often focus on the idiosyncrasies of language and the way music relates to the spoken word. Lucy is currently a senior at the University of Indianapolis and will be attending graduate school for music composition at the University of MissouriKansas City Conservatory this fall. 22  ...

... Post-damage responses in a mixed species hay field: interactions of tolerance and interspecific competition on yield and nutrition By Lily Schoolcraft Honors Project Thesis Submitted in Partial Fulfillment Of the Requirements for a Baccalaureate Degree With Distinction In the Ron and Laura Strain Honors College of THE UNIVERSITY OF INDIANAPOLIS May 2020 Faculty Advisor: Dr. Daniel R. Scholes, Ph.D. Executive Director of Honors: Dr. James B. Williams, Ph.D. Lily Schoolcraft 2020. Copyright in this work rests solely with the author. Please ensure that any reproduction or re-use is done in accordance with relevant national copyright protection. i Abstract Tissue removal by animal herbivores is a fundamental stressor in the evolutionary history of plants. In the agriculture of perennial biomass crops (e.g. grasses as feedstock), harvesting is a necessary damage event that one hopes is tolerated such that subsequent harvests are not reduced in quantity or nutrition. The objective of this study was to determine how a community of mixed plant species is affected by harvest for hay production. Plants in the study field were harvested once in May and again in September, which is a common practice for this type of field. Another field was harvested once in the fall, acting as a control. Comparisons of the biomass and nutrient profiles of first vs. second harvests allowed inference to the response of plants to damage. The results of this study indicate that a strategy of two harvests provides more biomass and nutrition in total relative to one late-season harvest. This is due in part to the apparent tolerance of some species to damage, such that their biomass was maintained or increased in the second harvest relative to the first. Another factor was the substantial recruitment of a species late in the season that was absent in the first harvest, which contributed to the maintenance of biomass and nutrition when comparing the two harvests. Overall, this study provides an assessment of the response of a mixed plant community to tissue removal within the context of perennial agriculture, with inference to the evolutionary strategies employed and potential for application to livestock feeding. ii Table of Contents Cover Page .... Abstract ... i Table of Contents ... ii Acknowledgements ... iii List of Figures .... iv Statement of Purpose .. 1 Introduction 2 Methods/Procedure . 4 Product Produced 8 Results 8 Analysis/Conclusion . 11 Reflection . 17 Bibliography . 19 Appendix .. 23 Appendix A - CITI Training Certificate .. 23 iii Acknowledgements Thank you to the Strain Honors College at the University of Indianapolis for making this research project possible through their financial support. Special thanks to Dr. Daniel Scholes for serving as my project supervisor, providing guidance and mentorship throughout the duration of this project. Thank you to my family for supporting me throughout this project and for the use of their field. Finally, thank you to SureTech Laboratories in Indianapolis for conducting the near infrared reflectance spectroscopy analyses of forage samples. iv List of Figures Figure 1: Total mass and nutrient mass of early first, late second, and late first harvests . 8 Figure 2: Total mass and nutrient mass of the sum of early first and late second harvests and the late first harvest .. 9 Figure 3: Species masses by harvest 10 1 Statement of Purpose The objective of this study was to determine how a community of mixed plant species for hay production was affected by mechanical damage. This allowed for inference to their damage response strategies, which theoretically could include resistance or tolerance. Plants belonging to the experimental group were harvested, allowed to re-grow over a four-month period, and then harvested again, which is a common practice for forage crops. Another plot was harvested at the end of the season only, acting as a control for seasonality of the experimental plot harvests. Comparisons of the biomass and nutritional quality of first vs. second harvests should have allowed inference to the response of plants to damage, where plants should regrow larger (tolerate) or less nutritious (resist). An unforeseen late-season species increased in abundance dramatically, however, being absent in the early-season plots, such that late-season plots were intermixed with previously damaged and undamaged plants of different species. Our ability to assess individual tolerance and resistance responses at the species level was therefore impaired. Results indicate that while some individual species appeared to tolerate damage in terms of biomass production, the plant community-maintained biomass and nutritional quality overall between first and second harvests. Plots that were only harvested once in the late-season produced less yield and were less nutritious relative to the sum of the two harvests of comparison plots. These results indicate that community composition may change from early- to late-season, but that at least for the field assessed, a two-harvest strategy maximizes yield and nutrition in part due to species tolerance abilities. 2 Introduction The ability of plants to adapt and respond to damage is crucial to their survival. Without this ability, they may succumb to even the most minor threats to their fitness (e.g. small-scale tissue consumption by insects). Plants frequently undergo assaults from a wide range of sources such as weather damage (e.g. damage from hail, frost, or flooding), trampling, and herbivory. The ways in which a plant responds to damage vary greatly between species and even between populations and individuals of a single species (Byington et al. 1994; Agrawal 1998). Evolutionarily, a plants response to damage should be one that will give the plant the best chance at survival and reproduction (i.e. maintain its fitness) given its adverse experience as promoted by natural selection. Because damage by animal herbivores is a fundamental (i.e. common among plant taxa, frequent, and sometimes severe) stress, herbivores are a major driving force of natural selection for plant defense strategies (Mauricio and Rausher 1997; Fornoni et al. 2004). Generally, plants combat and/or compensate for tissue removal via two responses: resistance and tolerance. Resistance responses encompass any strategy that prevents future damage, and typically includes structures (e.g. thorns, trichomes, and lignins) that make plants difficult to consume or chemical compounds that make plants toxic or unpalatable (Westbrook et al. 2011; Barton 2016). A plant may utilize chemical resistance strategies such as digestion inhibitors and toxins that interfere with an animals physiological functions (Par and Tumlinson 1999; War et al. 2012). These responses may be constitutive (i.e. always present, even without herbivory) or may be induced following an herbivory event (Agrawal 1998; Agrawal 2007; Mithfer and Boland 2012). Additionally, the plant may employ physical deterrents such as 3 thorns, spines, and tough, unpalatable leaves that defend against herbivore attack (Westbrook et al. 2011; Barton 2016). Tolerance responses, in contrast, describe efforts to mitigate the effects on biomass and/or seed yield of damage that has already occurred (Karban and Myers 1989; Stowe et al. 2000). Tolerance strategies are particularly effective for severe damage events, as is most typical by mammalian herbivores. Large mammals, such as deer and elk, impose substantial selection pressure on plants for regrowth strategies due to the severe but predictable damage imposed on plants along their migration routes (Dublin et al. 1990; Cadenasso and Pickett 2000; Goheen et al. 2004). This removal causes significant damage and gives the plant the challenge of regaining its former biomass, or total size, quickly in order to maximize its fitness in the now muchshortened growing season. Because the chemical resistance strategies that may ward off a small insect may not necessarily have the same effect on a larger herbivore, or they may be relatively impervious to small-scale physical deterrents like trichomes, the plants ability to tolerate damage becomes imperative. In order to increase its tolerance to damage, a plant may employ an increased net photosynthetic rate leading to high re-growth rates and potentially increased branching, and/or the storing of high levels of carbon in roots prior to damage and shunting those carbon stores and nutrients to the regrowing shoots following damage (Strauss and Agrawal 1999; Yamaji and Ma 2014). Regardless of the mechanism, the response employed aims ideally to achieve comparable fitness to when undamaged, though myriad genetic and environmental factors cause individuals to be more or less successful at this (Weis et al. 2000; Wang et al. 2003). Responding to damage may intuitively seem less desirable than preventing damage, but numerous studies over the years have shown tolerance to be an effective response to damage that 4 may actually increase plant biomass and/or nutritional quality (McNaughton 1979; Tiffin 2000; Wilson et al. 2007; Fornoni 2011; Yamaji and Ma 2014). Longstanding theory posits that resistance and tolerance strategies are redundant (i.e. a resistant plant does not get damaged to the point that it evolves tolerance ability, and vice versa), and so plants under limited resources should exhibit one but not both strategies due to the necessity and limitation of nutrient allocation (van der Meijden et al. 1988; Poorter and Nagel 2000; Zeng 2003). In an agricultural setting, plants are harvested frequently and/or severely (e.g. the removal of all aboveground biomass in potentially repeated damage events), such that their response in terms of biomass and nutrient distribution following damage is of great importance. Given that plants will largely employ resistance or tolerance strategies, but not both, and that tolerance is more desirable in repeatedly harvested agricultural crops, to what degree is tolerance in terms of biomass and nutrition observed among hay forage? For example, if nutrients are devoted towards resistance, the regrowth should be of lesser nutritional quality to the herbivore (i.e. livestock, in an agricultural setting) because more nutrients are being used to produce defenses that prevent further damage and yield likely will be reduced with subsequent harvests. If nutrients are instead invested into tolerance, the regrowth should be of similar or greater nutritional quality relative to initial growth with limited-to-no reduction in yield. Methods/Procedure Experimental Design This study was conducted on a mixed-species hay field near Martinsville, IN, USA. In order to investigate the effect that damage has on the plant community in relation to harvest quantity and quality, two separate damage events were carried out using mechanical harvesting of aboveground biomass as is standard for hay production. In early spring, the field was divided 5 into two sections. The experimental section underwent two harvests throughout the course of the year: an early first harvest in May and a late second harvest in September. The first harvest acted as the initial damage event to assess natural, undamaged plant quality and biomass, while the second harvest served to assess these characters in regrown plant tissue based on their response to the first harvest. The other field section underwent only one late first harvest at the same time as the second cutting of the experimental group. This section served as a control for the effects of seasonality (e.g. to determine if nutritional quality and biomass change late relative to early regardless of damage) in comparison to the experimental group. Prior to the first experimental harvest in May, twelve 11-meter plots within the experimental and control field sections were randomly selected and marked using metal stakes and twine. These plots remained marked in this manner for the duration of the experiment, with the exception of harvest days. In order for the mower to easily maneuver over these plots during harvests, the twine was removed, and the stakes were driven deep enough into the ground to avoid the mower blades. After cutting had been completed, the stakes were pulled back up and the twine replaced so that the plots were once again clearly visible and the plants from these plots were collected. These procedures were repeated in the late second harvest, allowing the same plots of plants to be directly compared in the early first harvest of undamaged growth relative to the late second harvest of regrowth. The procedures were also repeated in the control field section for a late first harvest. Each field section underwent cutting, raking, and baling on the same days so as to prevent any difference between sections based on timing of these events. As a measure to prevent intermixing of plant material between sections, all remaining plant material was removed from the harvesting equipment before using it on the next section. 6 Data Collection Biomass & Plant Quality An identical data collection procedure occurred twice: once in late May (first harvest of the experimental section) and a second time in September (second harvest of the experimental section and first harvest of the control section). On the day the hay was harvested, all plants within the 11-meter sections were cut to near ground level using shears. The material from each plot was bagged for transport and kept separate from material from other plots. After removal from the field, the material in each bag was spread out on a tarp and allowed to dry for approximately five days, which is typical for hay production. Once fully dried, the material from each plot was sorted by species and weighed in order to determine the total dry biomass of each species within each 11-meter plot. Then, species mass totals were combined to obtain the total biomass produced by the sample plot. Having data on the pre- and post-damage biomass of the experimental section allowed us to compare the overall change in biomass between the early first and late second harvest. After the material from each 11-meter section had been fully dried, sorted, and weighed, they were prepared for nutritional analysis. This process consisted of cutting all plant material from each individual plot into small pieces and thoroughly mixing them to create a mixture representative of the plot. Then, a 500-gram sample from each plot was individually bagged and labeled. The samples were then sent to SureTech Laboratories (Indianapolis, IN, USA) for analysis. Near infrared reflectance (NIR) spectroscopy analysis was conducted to measure protein, available protein, calcium, phosphorus, magnesium, potassium, and lignin, all measures of nutritional quality for livestock consumption (Ball et al. 2001). In total, 36 total samples were analyzed for biomass, species composition, and nutrition: 12 early first harvest samples, 12 late second harvest samples (which were from the same plots as the early first harvest samples), and 12 late first harvest samples as a control for the timing of 7 harvest. Nutritional analyses yielded measurements of the percentage of each nutrient in a standardized amount of sample; these percentages were summed to estimate the total nutrient percentage of each plot on a per-mass basis. A variable that incorporates both aspects of quality and quantity, termed nutrient mass, was calculated for each plot by multiplying its biomass with its total nutrient percentage. Data Analysis Total biomass was compared between early first and late second harvests via a paired samples t-test, matching the 11-meter plots of each harvest. Nutrient percentage and nutrient mass were compared between the early first and late second harvests similarly. Comparisons of these measures were performed between early first and late first harvests, and additionally late second and late first harvests, via independent samples t-tests. In order to compare attributes of the two harvests overall relative to the single harvest, the sum of each measure (total biomass, nutrient percentage, nutrient mass) was calculated and compared with the measures of the late first harvest via independent samples t-tests for each measure. In order to assess the species composition of each harvest, and therefore to infer the response of each individual species to damage inflicted within the growing season, the masses of each species were compared between early first and late second harvests by paired samples ttests. Species masses were similarly compared between the sum of early first and late second harvests and the late first harvest by independent samples t-tests. Unexpectedly, a species (creeping foxtail) appeared in the late season harvests that was absent from the early first harvest. For this species, its increase in mass relative to the early season (0) was analyzed via one-sample t-test. 8 Product Produced This manuscript is the primary product produced from this study. In addition, a poster summarizing the study will be produced and displayed during the 2020 University of Indianapolis Scholars Day research colloquium. Results Total mass and nutrition by harvest When harvested twice during a single growing season, the harvests (early first and late second) were not significantly different in mass (t(11) = 0.30, p = 0.77, Figure 1). Although the second harvest had the same mass as the first, however, it was marginally statistically significantly more nutritious on a per-mass basis (early first: 26.6%; late second: 30.1%; t(11) = 1.98, p = 0.074). Considered together, the total nutrient mass of the two harvests did not differ statistically (t(11) = 0.85, p = 0.416; Figure 1). 1200 Total Mass 1000 Nutrient Mass Mass (g) 800 600 400 200 0 Early First Late Second Late First Harvest Figure 1: Total mass (white) and nutrient mass (gray) of early first, late second, and late first harvests. Plotted are means standard errors. 9 The field section harvested only once late in the season (late first) yielded more total biomass than either of the two harvests of the other field section independently (late first vs. early first: t(22) = 4.42, p < 0.0001; late first vs. late second: t(22) = 4.41, p < 0.0001; Figure 1). The single harvest was more nutritious on a per-mass basis than the early first harvest (late first: 29.0%; early first: 26.6%; t(22) = 2.09, p < 0.05), but was comparable in nutritional percentage to the late second harvest (late second: 30.1%; t(22) = 0.54, p = 0.60). Together, the field section harvested twice yielded significantly more total biomass and total nutrient mass than the field section harvested once (total biomass: t(22) = 4.13, p < 0.0001; nutrient mass: t(22) = 4.04, p < 0.001; Figure 2). 1600 Total Mass 1400 Nutrient Mass Mass (g) 1200 1000 800 600 400 200 0 Sum of Early First and Late Second Late First Harvest Figure 2: Total mass (white) and nutrient mass (gray) of the sum of early first and late second harvests and the late first harvest. Plotted are means standard errors. Species composition of harvests by mass When damaged, nearly all species regrew to a lesser mass for the late second harvest than they had at the time of the early first harvest (all p < 0.001 except red clover: t(11) = 0.22, p = 0.833; Figure 3). The greatest decrease occurred for tall fescue, which was the most prominent 10 species (by mass) in the early first harvest (493.6 g on average per plot) with a dramatic reduction in the late second harvest (81.7 g; Figure 3). The decrease in tall fescue does not seem to be a natural occurrence due to seasonality (i.e. a natural late-season decline), since it maintained overwhelmingly greater mass than any other species in the late first plots (Figure 3). Instead, a new species which was not present during the early season harvest appeared in both of the late season harvests (late first and late second), which increased in mass dramatically in the late second harvest (t(11) = 11.8, p < 0.0001) concurrently with the substantial reduction in tall fescue mass (Figure 3). While creeping foxtail likewise appeared in the late first harvest (i.e., in the field section that had not been previously harvested and in which tall fescue remained the most massive species by far), creeping foxtails mass was far lower in the late first harvest than in the late second harvest (t(22) = 11.2, p < 0.0001; Figure 3). 900 Early First 800 Late Second Mass (g) 700 Late First 600 500 400 300 200 100 0 Tall Fescue Orchard Grass Timothy Grass Red Clover Creeping Foxtail Species Figure 3: Species masses by harvest (early first in white, late second in light gray, late first in dark gray). Plotted are means standard errors. 11 Analysis/Conclusion The results of this study indicate that a strategy of two harvests provides more biomass and nutrition in total relative to one late-season harvest. This is due in part to the apparent tolerance of some species to damage, such that their biomass increased in the second harvest relative to the first. Another factor was the substantial abundance of creeping foxtail in the late second harvest relative to the early first harvest, which contributed to the maintenance of biomass and nutrition when comparing the two harvests. The recruitment of creeping foxtail into the plant community may have been aided by competitive release provided by the first harvest, whereby damage may have provided the availability of light and soil nutrients to the undamaged creeping foxtail during the time of stress of the other species following the first harvest. Without this early damage (i.e. in the late first harvest plots), creeping foxtail recruitment occurs but to a much lower abundance, likely due to competition with already-established plant species. Overall and from an agriculturalists viewpoint, it seems best to perform two harvests within a single growing season, despite the fact that the species composition within each harvest may vary greatly and the individual harvests are less massive than a single late-season harvest. A two-harvest strategy yielded individual harvests which were smaller than the late season single harvest strategy. Initially harvesting in May reduced the overall seven month growing period (March September) into two 3.5-month periods, and incurred stress on the plants as they lost substantial biomass, aboveground stored resources, and photosynthetic capacity, and were likely more susceptible to subsequent transpirational water loss and microorganism infection. Despite these setbacks and the necessity to regrow in a much-reduced timeframe before seasons end, regrowth in the second half of the growing season, following damage, was comparable to that of the first half of the growing season. This indicates that 12 growth of the plant community as a whole was not delayed substantially by damage. A change in the plant community, mainly the recruitment of other species late in the season (e.g. creeping foxtail) that were not present early in the season, contributed to the maintenance of growth rates following the initial harvest. Although smaller individually, the sum of the individual harvests biomass was greater than that of a single harvest, however. This is likely due to a plateau in biomass accumulation (i.e., a decrease in growth rates) late in the season, as late in the season is typically when growth rates decline and instead resources are used for reproduction, and/or as light and soil nutrients become less available as competition increases due to shoot and root crowding. Finally, it does appear that some individual species (red clover) displayed the ability to tolerate damage in terms of biomass production, having approximately the same mass in the early first and late second harvests. The degree to which our perception of tolerance is affected by factors like interspecific competition is unknown (such that, hypothetically, tall fescue may have been able to tolerate damage when alone but not when in competition with creeping foxtail). Tolerance has been reported in natural mixed-species settings, typically in fields with a long history of severe, repeated, and predictable damage events (as is often the case in the agriculture of biomass crops; Dublin et al. 1990; Lennartsson et al. 1998; Cadenasso and Pickett 2000; Goheen et al. 2004). On a per-mass basis, the late season second harvest was marginally more nutritious than the early season first harvest. This indicates that despite achieving a lesser total mass, the postdamage regrowth actually grew to be more nutritious gram-for-gram than the initial growth. This is characteristic of a community tolerance response for nutrition. There are a few main mechanisms of tolerance which are widely recognized, one of which is the ability to rapidly move nutrients and carbon stores from roots to aboveground tissues following damage (Strauss 13 and Agrawal 1999). Therefore, the higher concentration of nutrients in aboveground tissue, which is what was analyzed, may be in part responsible for this increase in nutrition. Another possible explanation for this phenomenon could lie in the appearance and prevalence of creeping foxtail within the late season second harvest. If creeping foxtail were a naturally more nutritious species than the species which made up the early season harvest, it would be reasonable to conclude that the increase in nutritional percentage was at least in part due to its presence. The rapid, widespread presence of creeping foxtail in the field post-damage may be an example of competitive release. Competitive release is a fundamental concept within ecology that helps to explain species diversity and abundance within a community. It refers to the idea that when a competitor is removed or inhibited, a species that was previously suppressed will take hold (Huston 1979). In other words, the competitor is able to suppress another species based upon some sort of advantage, whether that be faster growth rates which allows it to crowd out space for later maturing species, or simply superior characteristics (e.g. shade tolerance) that allow it to outcompete other species for limited resources. Once the competitor is disturbed, other species are released from its suppressive effects and are able to utilize resources more fully than when in the competitors presence. This phenomenon has been observed many times within grassland communities, including those in Mediterranean grasslands (Segre et al. 2016) and tallgrass prairies (Jutila & Grace 2002; McCain et al. 2010). One major mechanism leading to competitive release is herbivory, whether in the form of grazing or largescale harvest events such as those regularly carried out in the agriculture industry. In our case, the early harvest released creeping foxtail from the suppressive effects of its competitors, likely the highly abundant tall fescue in particular. When the field was harvested in the early season, creeping foxtail was not large or abundant enough to appear in the harvest. The disturbance to its 14 competitors allowed the repressed, later maturing creeping foxtail to rapidly increase in abundance. When the field was left undisturbed, resources (e.g. space, light, soil nutrients) were apparently limited enough by competition with established species to limit its size and abundance within the field. Creeping foxtail and tall fescue have likewise been observed in other studies to suppress the performance of other species in grassland communities. Specifically, Moyer and Boswall (2002) found that when in the presence of either creeping foxtail or tall fescue, a less competitive species, foxtail barley, had significantly reduced abundance. This study supports our findings of tall fescue and creeping foxtail being the most abundant (or massive, at least) species likely via their abilities to outcompete others. In addition, it helps provide an explanation for creeping foxtails rapid success, not only as a result of competitive release, but as a combined effort with its natural competitive ability. Not all assessed species appeared to tolerate damage based on their decrease in biomass regrowth following harvest. All species studied are common crop grasses which have a long history of undergoing frequent largescale damage events (i.e. harvests). Based upon that fact, natural selection would favor variants within species that are capable of tolerating damage, so a resistance response in these species would be unexpected. Our findings with tall fescue, which was the most prominent species by mass during the first harvest of the experimental field, seemed to indicate that this species does not tolerate damage in terms of biomass production. This was based upon the fact that tall fescue, despite being prominent in the first harvest, decreased significantly in mass production by the second harvest. However, its decrease in abundance might not actually have been due to its lack of tolerance ability, but as a result of being outcompeted while regrowing from a damaged state. Another species, orchard grass, which also theoretically should have tolerated damage given its history of being a repeatedly 15 harvested biomass crop (i.e., of artificial selection promoting its tolerance ability), also appeared to not tolerate damage. Orchard grass is known to be less competitive against other species as compared to either tall fescue or creeping foxtail (Moyer & Boswall 2002). The naturally weaker competitive ability of orchard grass combined with the presence of creeping foxtail, which is a naturally better competitor, could actually be the explanation for orchard grasss decrease in abundance rather than a lack of a tolerance response. Studies assessing these species grown as monocultures would eliminate the contribution of interspecific competition on the results and help determine the true tolerance abilities of these species. Limitations to this study were primarily related to the logistics of collecting data within an agricultural operation. Despite collecting samples from twelve randomly dispersed plots per harvest, we only had access to one field for data collection. In addition, prior to data collection, the field had to be divided in half, designating the control (late season first harvest) and experimental (early season first and late season second harvests) sections. This was a necessary design due to the maneuverability of the large equipment used to harvest the fields, rather than interspersing all plots throughout the field which would have better randomized environmental variables across the plots. Although sections of the field were not observably different, it is possible that there is variation in species composition along with water and nutrient availability between sections. Other limitations included the difficulties of controlling experiments outside of a laboratory setting. Throughout the season, species composition changed in the field as made evident by the late appearance of creeping foxtail. In addition, in a community setting different species will compete for resources such as space, light, water, and nutrients. This competition within our mixed species plots added another layer of unpredictability to the study where not only the ability of individual species to tolerate damage had to be taken into consideration, but 16 also their ability to compete against other species within the community and as the community changed. These are all dynamics of real hay crop communities, however, and so our results are reflective of the myriad changes and interactions occurring in this field. To understand the tolerance responses of species in the initial plant community, future experiments in the field setting could carefully monitor plots and remove any new species that appear. Had creeping foxtail been removed, more inferences could be made about the tolerance abilities of the original four focus species (tall fescue, orchard grass, timothy grass, and red clover), since they all incurred damage but may not have all regrown back to the masses they had at the time of the first harvest. This method of removal could also be followed to eliminate the factor of interspecific competition (i.e. competition for the same resources among different species in a community) by removing all plants from plots with the exception of a single focal species. Moving forward, future experiments seeking to further understand the responses of individual plant species to damage could benefit from performing the experiment in a controlled lab setting where other species are unable to interfere with rates of growth or outcompete the species of interest, and/or where environmental factors (e.g. water availability) could be manipulated. However, if considering these responses in relation to their practicality for agriculturalists, the optimal research setting is a natural field as used in this experiment but expanded to larger and more numerous fields as well as with different species compositions and/or environmental factors (e.g. soil types). In total, this study yielded information on plant tolerance to damage in the context of harvesting for hay production, providing insight to the dynamic responses of a mixed-species community in relation to species composition, interspecific competition, regrowth abilities, total yield, and nutrition. 17 Reflection Overall, despite logistical and physical difficulties with completing this project, combined with the unexpected appearance of a species late in the season, I believe that this project was successful in generating interesting results that reflect the complexity of not only individual species responses to damage but also the interactions and competition among individuals within a mixed species community. This study also produced information that is valuable to agriculturists as they consider which species to include in their fields and the optimal harvest schedule. The planning and proposal phase of the project went smoothly and with the guidance of Dr. Scholes, I was able to produce a project proposal that was accepted on the first attempt. Data collection was very time consuming given that throughout the course of the experiment, I cut a total of thirty-six 11-meter sections of grass by hand and then sorted each section by species so that they could be individually weighed. The first round of data collection went well, but during the second round of collection in the fall, difficulties arose. This difficulty was the unexpected appearance of a new species of grass in the late season harvest that interfered with our ability to fully answer our original research questions. Fortunately, this setback did not ruin the project and questions were able to be reframed. The planning, execution, and completion of this project were both personally and professionally impactful. This has arguably been one of the most challenging, yet rewarding, projects of my academic career. It fostered personal growth by forcing me to work harder, think deeper, and believe in my ability to complete something that my freshman honors student self once viewed as daunting. It further taught me how to persevere through difficulties and how to take those setbacks and turn them into something worthwhile. Having completed an independent 18 research project was undoubtedly beneficial while applying and interviewing for graduate programs and aided in my acceptance to multiple schools, including my top choice. Not only that, but this project has given me experience in carrying out research from start to finish which will be invaluable as I transition into graduate school. 19 Bibliography Agrawal A. 1998. Induced responses to herbivory and increased plant performance. Science, 279(5354): 1201-1202. Agrawal A. 2007. Macroevolution of plant defense strategies. Trends in Ecology and Evolution, 22(2): 103-109. Ball D, Collins M, Lacefield G, Martin N, Mertens D, Olson K, Putnam D, Undersander D, Wolf M. 2001. Understanding Forage Quality. American Farm Bureau Federation Publication, 1(01): 1-17. Barton KE. 2016. Tougher and thornier: general patterns in the induction of physical defence traits. Functional ecology, 30(2): 181-187. Byington TS, Gottschalk KW, McGraw JB. 1994. Within-population variation in response of red oak seedlings to herbivory by gypsy moth larvae. American Midland Naturalist, 132: 328-339. Cadenasso ML, Pickett STA. 2000. Linking forest edge structure to edge function: mediation of herbivore damage. Journal of Ecology, 88(1): 31-44. Dublin HT, Sinclair AR, McGlade J. 1990. Elephants and fire as causes of multiple stable states in the Serengeti-Mara woodlands. The Journal of Animal Ecology, 59(3): 1147-1164. Fornoni J. 2011. Ecological and evolutionary implications of plant tolerance to herbivory. Functional Ecology, 25(2): 399-407. Fornoni J, Valverde PL, Nunez-Farfan J. 2004. Population variation in the cost and benefit of tolerance and resistance against herbivory in Datura stramonium. Evolution, 58(8): 16961704. 20 Goheen JR, Keesing F, Allan BF, Ogada D, Ostfeld RS. 2004. Net effect of large mammals on Acacia seedling survival in an African savanna. Ecology, 85(6): 1555-1561. Huston M. 1979. A general hypothesis of species diversity. The American Naturalist, 113(1): 81-101. Jutila HM, Grace JB. 2002. Effects of disturbance on germination and seedling establishment in a coastal prairie grassland: a test of the competitive release hypothesis. Journal of Ecology, 90(2): 291-302. Karban R, Myers J. 1989. Induced Plant Responses to Herbivory. Annual Review of Ecology and Systematics, 20(1989): 331-348. Lennartsson T, Nilsson P, Tuomi J. 1998. Induction of overcompensation in the field gentian, Gentianella campestris. Ecology, 79(3): 1061-1072. Mauricio R, Rausher MD. 1997. Experimental manipulation of putative selective agents provides evidence for the role of natural enemies in the evolution of plant defense. Evolution, 51(5): 1435-1444. McCain KN, Baer SG, Blair JM, Wilson GW. 2010. Dominant grasses suppress local diversity in restored tallgrass prairie. Restoration Ecology, 18: 40-49. McNaughton SJ. 1979. Grazing as an optimization process: Grass-ungulate relationships in the Serengeti. The American Naturalist, 113: 691-703. Mithfer A, Boland W. 2012. Plant defense against herbivores: chemical aspects. Annual Review of Plant Biology, 63: 431-450. Moyer JR, Boswall AL. 2002. Tall fescue or creeping foxtail suppresses foxtail barley. Canadian Journal of Plant Science, 82(1): 89-92. 21 Par PW, Tumlinson JH. 1999. Plant volatiles as a defense against insect herbivores. Plant Physiology, 121(2): 325-332. Poorter H, Nagel O. 2000. The role of biomass allocation in the regrowth response of plants to different levels of light, CO2, nutrients and water: a quantitative review. Australian Journal of Plant Physiology, 27: 595-607. Segre H, DeMalach N, Henkin Z, Kadmon R. 2016. Quantifying competitive exclusion and competitive release in ecological communities: a conceptual framework and a case study. PloS One, 11(8). Stowe K, Marquis R, Hochwender C, Simms E. 2000. The evolutionary ecology of tolerance to consumer damage. Annual Review of Ecology & Systematics, 31: 565-595. Strauss S, Agrawal A. 1999. The ecology and evolution of plant tolerance to herbivory. Trends in Ecology and Evolution, 14(5): 179-185. Tiffin P. 2000. Mechanisms of tolerance to herbivore damage: what do we know? Evolutionary Ecology, 14: 523-536. Van der Meijden E, Wijn M, Verkaar HJ. 1988. Defence and regrowth, alternative plant strategies in the struggle against herbivores. Oikos, (1988): 355-363. Wang W, Vinocur B, Altman A. 2003. Plant responses to drought, salinity, and extreme temperatures: towards genetic engineering for stress tolerance. Planta, 218(1): 1-14. War AR, Paulraj MG, Ahmad T, Buhroo AA, Hussain B, Ignacimuthu S, Sharma HC. 2012. Mechanisms of plant defense against insect herbivores. Plant Signaling & Behavior, 7(10): 1306-1320. Weis AE, Simms EL, Hochberg ME. 2000. Will plant vigor and tolerance be genetically 22 correlated? Effects of intrinsic growth rate and self-limitation on regrowth. Evolutionary Biology, 14(4-6): 331-352. Westbrook JW, Kitajima K, Burleigh JG, Kress WJ, Erickson DL, Wright SJ. 2011. What makes a leaf tough? Patterns of correlated evolution between leaf toughness traits and demographic rates among 197 shade-tolerant woody species in a Neotropical forest. The American Naturalist, 177: 800-811. Wilson JR, Yeates A, Schooler S, Julien MH. 2007. Rapid response to shoot removal by the invasive wetland plant, alligator weed (Alternanthera philoxeroides). Environmental and Experimental Botany, 60(1): 20-25. Yamaji N, Ma JF. 2014. The node, a hub for mineral nutrient distribution in graminaceous plants. Trends in Plant Science, 19(9): 556-563. Zeng B. 2003. Aboveground biomass partitioning and leaf development of Chinese subtropical trees following pruning. Forest Ecology and Management, 173(1-3): 135-144. 23 Appendices Appendix A - CITI Training Certificate  ...

... Spermiogenesis within the testis of the Mexican Whiptail Lizard, Aspidoscelis costata By Jaime Rhodes Honors Project Thesis Submitted in Partial Fulfillment Of the Requirements for a Baccalaureate Degree With Distinction In the Ron and Laura Strain Honors College of THE UNIVERSITY OF INDIANAPOLIS May 2020 Faculty Advisor: Kevin M. Gribbins, Ph.D. Executive Director of Honors: James B. Williams, Ph.D. Jaime Rhodes 2020. Copyright in this work rests solely with the author. Please ensure that any reproduction or re-use is done in accordance with relevant national copyright protection. 2 Jaime Rhodes Abstract To date, there are multiple studies detailing the ultrastructure of spermatozoa within the genus Aspidocelis and within the family Teiidae. However, there are no complete data on the process of spermiogenesis within a species of Whiptail Lizard or within Ground Lizards. The purpose of this investigation is to provide the first detailed study on the entire ultrastructural process of spermiogenesis within Aspidocelis costata. These results can then be compared to the spermatozoa data that is available for Teiidae. Testicular tissues of A. costata were fixed, dehydrated, embedded in epon812, sectioned with an ultramicrotome, and viewed using transmission electron microscopy (TEM). Lizard spermatid morphology in recent studies have been found to be relatively conserved to one another within Squamata. Unlike spermatozoa, however, small differences during the development of the spermatids within lizards of the same genus have been observed with Transmission Electron Microscopy (TEM). Thus, we hypothesize from the preliminary data collected to date that differences in acrosome formation, midpiece morphology, nuclear morphology, and manchette morphology will be observed between the spermatids of A. costata and others within Teiidae. We hope that the data provided here will add to the growing information about spermatid development in lizards. These data will provide new species and familial characteristics that can be compared to that of the previously studied saurian taxa. 3 Jaime Rhodes Table of Contents Acknowledgements Statement of Purpose Introduction Methods Results Figure 1 Figure 2 Figure 3 Figure 4 Discussion/Conclusion Reflection References Appendices 4 4 5 9 11 12 13 14 15 16 19 22 23 4 Jaime Rhodes Acknowledgements I would like to thank Dr. Gribbins for both his guidance and support throughout the completion of this Honors Project. I would also like to thank the UIndy Biology Department for providing me with equipment and other resources that were needed in order to conduct this research. Statement of Purpose Lizards and snakes, formally known as Squamata, are organized phylogenetically by both morphological and functional characteristics between species (Gribbins, 2011). Morphological classification (distinguishing differences relative to form and structure) plays a key role in this process. One effective method by which morphology can be studied is through the observation of changes to germ cells as they progress through spermatogenesis and spermiogenesis. Spermatogenesis is the formation of sperm cells, while spermiogenesis is the final maturational phase of spermatogenesis, where haploid spermatids differentiate into mature (and motile) spermatozoa. Existing literature regarding the morphological and genetic classification of squamates presents extensive research in regards to the study of mature spermatozoa and its use alongside genetic and molecular data to produce phylogenetic hypotheses on species relationships within the families of Squamata: Gekkonidae (Rheubert et al., 2011), Phrynosomatidae (Gribbins, et al., 2013), Dactyloidae (Rheubert et al., 2010), Lacertidae, Viperidae (Tourmente et al., 2008), Gymnopthalmidae (Teixeira et al., 1999c), Tropiduridae, Lepidosauria (Teixeira, 1999a), and Serpentes (Tavares-Bastos et al., 2007; Tourmente, 2008). However, very little research exists that focuses on the specific analyses of the ontogenic steps of spermiogenesis (Gribbins et al., 2014). Thus, its not surprising that research on spermiogenesis within the Whiptail Lizard, Aspidoscelis costa, has not been explored to date. Dr. Gribbins lab 5 Jaime Rhodes at the University of Indianapolis has conducted several studies analyzing spermiogenesis within different squamate species. He has worked alongside colleagues to more clearly classify the morphological steps of spermiogenesis within the Mediterranean Gecko, Hemidactylus turcicus (Rheubert et al., 2011), the Jamaican Gray Anole, Anolis lineatopus (Rheubert et al., 2009), and many other squamates (Gribbins and Rheubert, 2015). Gribbins et al. have also shown that spermiogenic data provides morphological differences between species within the same genus (such as Sceloporus bicanthalis, and Sceloporous variabilis, Gribbins et al., 2013). In contrast, classically utilized sperm ultrastructure rarely detects differences below the family level (Gribbins, 2011). Thus, this project aims to describe spermatid ultrastructure within the testis of Aspidoscelis costa and then compare it to that of other lizards studied to date. Furthermore, it has been suggested that observations of the morphology and ultrastructure of developing spermatids could also provide morphological data that can be combined with what is known of sperm morphology to further differentiate and classify squamate species phylogenetically (Gribbins, 2011). This honors project supports this long term goal by providing needed spermiogenic data for a species and genus that currently lacks any information regarding spermiogenesis. Ultimately, the expectations are that the present novel data will serve as a template for future spermiogenesis studies, and to support the future use of spermiogenic ultrastructural data for phylogenetic classification among squamates. Introduction Reptilian Spermatogenesis More accurate ways to classify organisms phylogenetically are continuously being explored during research. Spermatogenesis, the study of germ cells, is becoming an increasingly 6 Jaime Rhodes useful method by which ultrastructural changes of organisms can be observed (Gribbins et al., 2014). Within the reptilian class, studies focusing on the progression of germ cells are limited. Few research data sets exist that describe the complete ontogenetic steps of reptilian spermatogenesis using transmission electron microscopy (TEM). In other words, research that analyzes mitosis, meiosis, and spermiogenesis within the Reptilia is sparse. Study of spermatogenesis within amniotes (birds, mammals, and reptiles), is significant, due to the fact that during development, a change in testicular morphology takes place. This structural change is unique to amniotes, and is also indicative of their unique ability to rely upon internal fertilization (Gribbins et al., 2014). That being said, this structural change within seminiferous tubules of the testicles of amniotes also differs between different clades (reptiles vs. birds vs. humans). Comparisons between clades can aid in our richness of understanding of male reproduction and the utilization of these morphological characters may be useful in the classification of reptiles in future phylogenetic studies. Squamate Spermatogenesis Cycles Most amniotes that breed seasonally share common developmental changes in morphology and ultrastructure during sperm development. These changes are clearly observed within the epithelium of the seminiferous tubules (within the testis), due to the proliferation of older generations of germ cells that migrate towards the lumen of the seminiferous epithelium (Gribbins and Rheubert, 2014). Therefore, at any given time, up to five generations of germ cells can be viewed within the testis of an amniote. Squamates, however, are unique in their ability to house up to nine generations of germ cells within their epithelia, providing a more comprehensive record of the development of cells generationally within monthly samples of testes. Due to their ability to produce many generations of germ cells within the same 7 Jaime Rhodes reproductive cycle, reptiles are considered an ideal model for studies on ultrastructure of spermiogenesis. During reptilian spermiogenesis, five to nine spermatids can be observed within the seminiferous epithelium at the same time, as opposed to three that are consistently present in birds and mammals (Gribbins, 2011). Spermatogenesis in squamates is broken down into three phases: the proliferative phase, meiotic phase, and spermiogenic phase. During proliferation, spermatogonia undergo mitotic divisions that lead to the development of many germ cells. The meiotic phase involves spermatocytes that undergo meiotic division, segregation, and genetic recombination. The spermiogenic phase consists of haploid spermatids that specialize into mature sperm cells that have the ability to fertilize eggs (Gribbins, 2011). A number of complete ultrastructural and morphological studies of squamate spermatogenic cycles in lizards exist: Podarcis muralis (Gribbins et. al., 2013), Anolis lineatopus (Gribbins and Gist, 2003), Scincella lateralis (Rheubert et al., 2010), Hemidactylus turcicus, Sceloporus consobrinus, and Sceloporus mucronatus (Rheubert et al., 2011). The testes of amniotes consist of seminiferous tubules, small tubes in which spermatozoa develop, which are lined with thick epithelia that are composed of Sertoli cells. Germ cells undergo cytological and ultrastructural changes within these tubules. These developmental changes in squamate seminiferous tubules are significant, due to the fact that most squamates are seasonal breeders. Thus, the way that spermatids develop within their seminiferous tubules differs from other amniotes that do not breed seasonally, and from anamniotes. That being said, how exactly seasonal breeding is reflected within ultrastructural development of spermatozoa is unclear (Gribbins and Gist, 2003). Non-seasonally breeding amniotes typically undergo a spermatogenic cycle that is continuous, while spermatogenesis of many temperate anamniotes is known to be episodic or seasonal. Further studies of these cycles of sperm development in 8 Jaime Rhodes squamates can potentially aid in our understanding of the morphological and cytological development within the testes of seasonally-breeding organisms. Studies of some reptile species such as Trachemys scripta (Slider Turtle), have displayed a spermatogenesis pattern that is unique, consisting of only a single spermiation event at the end of the breeding cycle (Gribbins et al., 2003). This research indicates that some squamates may undergo a type of spermatogenesis that serves as an intermediate between episodic anamniotic development, and that of continuously breeding amniotes. Other reptilian species, specifically Agkistrodon piscivorus leucostoma (Western Cottonmouth Snake), also exhibit breeding patterns that are episodic, similar to their non-mammalian amniotic counterparts. These episodes, however, are not singularthey occur more than once throughout a calendar year. For example, Cottonmouths hold a continuous supply of spermatogonia within their seminiferous tubules year-round, however they practice a bimodal method of episodic spermatogenesis that takes place at the end of each breeding season (Gribbins et al., 2008). Ultrastructural Data on spermiogenesis in Lizards Of the over 30 families of lizards (roughly 9,000 species), approximately 10 possess complete morphological data in regards to spermiogenesis (Gribbins and Rheubert, 2014; Pyron, 2013). Thus, familial relationships as far as common morphology during spermiogenesis between lizard species are still widely unexplored, which makes phylogenetic classification based on spermiogenesis alone difficult to impossible. Additions of unstudied species can potentially aid in the validity and reliability of phylogenetic classification based on the ontogenic steps of spermiogenesis. Recent studies have begun to explore spermiogenic similarities and differences between two species within the same genus. Sceloporus variabilis and Sceloporus bicanthalis, two lizards from the Sceloporus genus, have been studied comparatively, and 9 Jaime Rhodes notably produced six main character differences in spermatid development (Gribbins et al., 2014). Presence/absence of a mitochondrial complex near the Golgi apparatus during development, nuclear indentation size, level of chromatin condensation, spiraling or nonspiraling chromatin, presence/absence of microtubules, and presence/lack of nuclear lacunae are differences have been discovered between these two lizards. The significance of these differences remains unclear. However, they aid in potentially supporting the claim that spermiogenic development is species specific (Gribbins et al., 2014). These developmental studies in lizards not only aid in a better understanding of their spermatid morphologies, but potentially provide more comprehensive data in regards to spermatid morphologies in amniotes as a whole. Histological analysis of Aspidoscelis costa will serve to add more data to the growing information available for lizards, as this study is the first of its kind for the Aspidoscelis genus. Methodology Animal collection: The six A. costalis specimens were collected just outside the city of Veracruz, Mexico during April of 2016 by Dr. Gribbins and colleagues. April represents the peak of spermiogenesis at this latitude in this species of Aspidoscelis (Gribbins, personal communication). Tissue Preparation for Electron Microscopy: After lizard collection, they were humanely anesthetized using MS222. The lizards were dissected and reproductive tracts were collected and fixed in Trumps fixative for at least 48 hours. The entire lizard was then placed in a solution of formaldehyde fixatives stored at the National Autonomous University of Mexico (UNAM). The testes were cut into multiple squares approximately 3x3 millimeters in size and post fixed in osmium tetroxide for two hours. The 10 Jaime Rhodes samples were then washed three times in distilled water and placed in small tubes of 70% ethanol solution until ready to embed. Embedding started with the dehydration process (a series of ascending ethanol washes), which created a favorable environment to begin replacing the ethanol with plastic. Placement of the tissue in plastic and then letting the plastic harden in an oven (for 48 hours) also provided support for the tissue while allowing very thin sections to be cut via ultramicrotomy. Sections cut with the Leica ultramicrotome were then placed onto grids and stained with uranyl acetate for 20 minutes and lead citrate for 5 minutes. TEM Analysis Grids were viewed using a Zeiss 109 TEM and pictures were taken of spermatid structure. These pictures were then scanned and analyzed using Photoshop CS7. Required Resources A comprehensive list of supplies needed for this research is supplied in Appendix B. Tissue samples of Aspidoscelis costa were obtained from Dr. Oswaldo Hernandez, a collaborator of Dr. Gribbins, from Universidad Autonoma del Estado de Mexico. A tissue embedding kit, along with tissue embedding media were utilized in order to prepare the samples for sectioning. Dr. Gribbins provided the ultra microtome, the device used for sectioning tissue samples. This machine utilizes glass knives in order to slice the sections. Dr. Gribbins supplied materials needed to prepare, stain, and observe the sectioned tissuesmicroscope slides, stains, a light microscope, and electron microscope. Timeline I began my project at the start of the 2019 spring semester, and continued working through winter of 2019. I completed all steps of tissue processing, along with photography of my samples by March of 2019. I utilized the following summer to analyze, organize, and label the 11 Jaime Rhodes photos that were taken. Following this step, I began to compare and contrast my findings of A. costa to those of other lizard speciesgiving me a solid framework to write my final paper and prepare my poster presentation (found in Appendix B). I prepared and finalized my final paper and poster over the Fall semester 2020, and had my project ready to present at the UIndy Honors Showcase (April, 2020). Results: Spermiogenesis begins in the round stage (Figure 1). An acrosome vesicle and acrosome granule form, and eventually flatten out the apex of the nucleus (Figure 1: A, B). Vesicles floating in the cytoplasm that originated from the Golgi apparatus (GA) of the cell simultaneously fuse with the membrane of the acrosome vesicle. As the acrosome granule (AG) grows, it travels to the inferior end of the acrosome vesicle, ending up against the apex of the nucleus once it reaches maximum size (Figure 1: A, B). Due to close association of the endoplasmic reticulum (ER) and mitochondria (Mi) to the Golgi apparatus throughout this stage of spermiogenesis, the two are labeled the mitochondrial complex. The proximal/distal centrioles (DC) at the inferior end of the nucleus (Figure 1: C) aid in flagellar formation. Towards the end of round stage spermiogenesis, the shoulders of the acrosome vesicle will collapse and then 12 Jaime Rhodes flatten the nuclear apex (Figure 1: C). An early and underdeveloped manchette (Ma) is also visible at this time (Figure 1: B). Figure 1: Aspidoscelis costata, round spermatids in early stages of spermatogenesis. A: Early acrosomal formation stage. An acrosome granule (AG) is located inside of the acrosome vesicle (AV). Apex of the nucleus (N) showing a shallow fossa accommodating the acrosome (NF). Golgi apparatus (GA, black arrow), endoplasmic reticulum (ER, black arrow), and mitochondria (Mi) are visible. B: Acrosomal vesicle (AV) has collapsed over the apex of the nucleus (N), acrosomal granule (AG) still sits inside of the acrosomal lumen. Manchette microtubules (Ma) are already visible just outside of the nucleus. C: Elongation begins at the final stage of round spermatids, and chromatin condensation begins and is filamentous. Both distal centrioles (DC) and annulus (An) are visible. Nucleus begins to constrict at the apex just underneath the acrosome (A) where the manchette microtubules (Ma) are most concentrated, and flagellum (F) and principle piece (PP) elongate. Mitochondria (Mi) are present. 13 Jaime Rhodes Next, spermatids begin to elongate, and nuclear length increases (Figure 2). The nucleus of the spermatid begins to shift upward in order to remove the cytoplasmic gap between the acrosome membrane and the spermatid cell membrane (Figure 2: A, B). Chromatin within the nucleus condenses filamentously, and in this case, produces light spiraling (Figure 2). A manchette and manchette microtubules also begin to increase in development (Figure 2: C). Figure 2: Aspidoscelis costata, elongating spermatids undergoing chromatin condensation. A: Acrosome vesicle (AV) and Acrosomal lucent ridge (Alr,) have wrapped around the nucleus (N). Manchette microtubules (Ma) become visible. B: Acrosome granule (AG) rests inside of the acrosome vesicle (AV). Subacrosomal space becomes visible as it wraps around the nuclear 14 Jaime Rhodes shoulders and apex of the nucleus (sAs). C: Chromatin in the nucleus is non-spiraling and filamentous. Manchette microtubules (Ma) are well-developed and visible at this stage. Spermatids continue to elongate as chromatin becomes fully condensed (Figure 3). Manchette microtubules continue to surround the nucleus (Figure 3: C). A nuclear fossa (NF) is now fully formed (Figure 3: A), and the proximal centriole lies just inside of it, followed by the presence of a distal centriole inferiorly that runs approximately 90 degrees to the proximal centriole. A fibrous sheath begins to develop around the distal centrioles, and large mitochondria and dense bodies surround them within the midpiece (Figure 3: B). Elongation thus produces both the midpiece of the spermatid and the principle piece, separated by the annulus of the flagellum (Figure 3: B). 15 Jaime Rhodes Figure 3: Aspidoscelis costata, in the late elongation phase. A: Elongating spermatid is approaching maximum nuclear (N) length. Nuclear fossa (NF) forms at the superior end of the midpiece (MP). Large mitochondria (Mi) and dense bodies (DB) are visible throughout the midpiece. Principle piece (PP), and annulus (An) is developed. B: Magnification of midpiece region. Proximal centrioles (PC) and distal centrioles (DC) are visible just inferior to the nuclear fossa. Annulus (An) sits at the superior end of the principle piece. C: Nucleus (N) is visible with a developed manchette (Ma). Once spermiogenesis is complete (Figure 4: A), the mature spermatid exhibits a developed acrosome vesicle at the superior end, followed by subacrosomal space that envelopes the nuclear rostrum (Figure 4: B, C, I). The acrosome consists of a perforatorium (Figure 4: A, B, I), subacrosomal space, an acrosomal lucent ride (Alr) (Figure 1: A), acrosomal cortex (Figure 4: B, C), and a basal plate (BP) located at the basal end of the perforatorium. The elongated nucleus is now surrounded by fully developed manchette microtubules (Figure 4: E). Microtubules (Figure 4: F) within the axoneme of the flagellum have peripheral fibers 3 and 8 extending from the distal centriole Figure 4: F, G). Peripheral fibers 3 and 8 do not extend into the principal piece and endpiece. Mitochondria of A. costatus are relatively large and well-developed surrounding the distal centrioles and within the midpiece; they are cylindrical in sagittal section (Figure 4: F, H). A fibrous sheath continues through the principle piece and terminates at the superior end of the endpiece. 16 Jaime Rhodes Figure 4: Aspidoscelis costata, in the final step of elongation and condensation of spermiogenesis. A: Sagittal section of mature spermatid. Perforatorium (Pf, black arrow), epinuclear lucent zone (EN), and nuclear rostrum (R) are visible. B: Cross section through the perforatorium (Pf), acrosome medulla (Am), and the acrosome cortex (AC). C: Cross section through the superior end of the nuclear rostrum (N), subacrosomal space (sAs), acrosome cortex (AC), and epinuclear lucent zone (EN). D: Cross section through nuclear rostrum (*), subacrosomal space (sAs) and the acrosome (A). E: Cross section through the nucleus (N) and the microtubules of manchette (Ma). F: Cross section through the midpiece, mitochondria (Mi), dense bodies (DB), and axoneme (Ax). G: Cross section through endpiece, axoneme of principle piece (Ax), fibrous sheath (FS) and microtubules of the manchette (Ma). H: Magnified view of the midpiece, fibrous sheath (FS), dense bodies (DB), and mitochondria (Mi). I: Magnified view of the acrosome vesicle (AV), perforatorim (Pf), epinuclear lucent zone (EN), nuclear rostrum (R), and nucleus (N). Discussion/Conclusion: The acrosome formation and development of round spermatids within Aspidoscelis costata formed a shallow indentation across the nucleus, which is also conserved across acrosome formation processes within S. variabilis (Gribbins et al., 2013), Sceloporus undulatus, and S. consobrinus (Rheubert et al., 2017). This contrasts with the deep acrosomal indentation formed by other squamates such as S. bicanthalis (Gribbins et al., 2013), A. lineatopus (Rheubert et al., 2010), and H. turcicus (Rheubert et al., 2011). This difference in nuclear indentation is potentially due to the difference in the duration of time that it takes for the acrosome vesicle to form and meet the nuclear membrane (Gribbins et al., 2014). A shallow indentation across the apex of the nucleus may indicate that the large vesicle took more time to fuse with the spermatid nuclear membrane. The acrosomal granule travels to the basal part of the acrosome vesicle, which is conserved across most studied squamates such as S. bicanthalis, S. variabilis (Gribbins et al., 2014), H. turcicus (Rheubert et al., 2011), S. consobrinus, S. undulatus (Rheubert et al., 2017), and P. muralis (Gribbins and Gist, 2003). Aspidoscelis costata exhibits a perforatorium within the acrosomal vesicle, which has been determined to be unique to squamates, and sympleisomorphic within this taxon (Teixeira et al., 1999a,c; Tavares-Bastos et al., 2007; Tourmente et al., 2008). 17 Jaime Rhodes The midpiece of A. costata consists of a fibrous sheath surrounded by large round and cylindrical mitochondria and dense bodies, which is also seen in S. variabilis (Gribbins et al., 2013), A. lineatopus (Rheubert et al., 2010), H. turcicus (Rheubert et al., 2011), S. bicanthalis (Gribbins et al., 2013), S. consobrinus, and S. undulatus (Rheubert et al., 2017). The peripheral fibers 3 and 8 of the midpiece do not extend into both the principle piece and the endpiece, a feature that is also seen in A. lineatopus (Rheubert et al., 2010) and S. consobrinus (Rheubert et al., 2017). This differs from other squamates whose peripheral fibers 3 and 8 extend into both the principle and end pieces, such as in S. variabilis (Gribbins et al., 2013), H. turicus (Rheubert et al., 2011), and S. bicanthalis (Gribbins et al., 2013). Chromatin of A. costata condenses in a filamentous formation with little spiraling. This pattern is seen in S. variabilis (Gribbins et al., 2013), and contrasts the well-developed spiraling condensation of other squamates such as H. turcicus (Rheubert et al., 2011), A. lineatopus (Rheubert et al., 2010), and S. bicanthalis (Gribbins et al., 2011). An epinuclear lucent zone is also present in the developing A. costata spermatids, which can be observed in all lizards studied to date (Gribbins and Rheubert, 2015). Aspidoscelis costata spermatids lack a nuclear lacuna, which is also the case for H. turcicus (Gribbins et al., 2011) and S. variabilis (Gribbins et al., 2014). Other lizards such as S. bicanthalis, S. consobrinus, and S. undulatus (Rheubert et al., 2017) exhibit a nuclear lacunae extending within the body of the nucleus. Historically, the nuclear lacunae have been considered absent, due to the fact that it does not always extend through the entire nucleus, and is thus not visible in cross section (Rheubert, 2016). During the elongation phase of A. costata, a manchette develops around the nucleus that does not extend far away from the nucleus cytoplasmically. A functional manchette is present in all squamates studies, which is thought to aid in stretching of the nucleus, except in the study of 18 Jaime Rhodes A. lineatopus (Rheubert et al., 2010). Historically, elongation of the nucleus is thought to be aided by the microtubules of the manchette, which are also present in A. costata (Gribbins et al., 2014). Circum-cylindrical and parallel microtubules make up the manchette of A. costata, and play a key role in the development and elongation of the nucleus throughout mid and late spermiogenesis. This is similar to the microtubule structure of S. variabilis, S. consobrinus (Rheubert et al., 2017), H. turcicus (Rheubert et al., 2011), S. bicanthalis, and S. undulatus (Rheubert et al., 2012). The extent of this microtubule scaffolding appears more compact and less developed in A. costata than within these other taxa of lizards. Thus, there seems to be much diversity to the appearance, consistency, and arrangement of microtubules within the manchette across squamates. It would be interesting in future studies to actually focus on the manchette and try to work on the functional significance of these differences across saurian taxa. After observing key steps of spermiogenesis in A. costata (acrosome formation, midpiece morphology, nuclear morphology, and manchette morphology), it appears that several A. costata characters are conserved across many existing saurian studies. As the first study to date on the spermiogenesis of A. costata, a conclusion of notable characteristics includes the development of a shallow nuclear indent on the acrosome vesicle (seen in S. variabilis), a midpiece surrounded by microtubules and peripheral fibers (3, 8) (seen in S. variabilis, H. turcicus, S. bicanthalis, S. consobrinus, and S. undulatus), and an underdeveloped manchette compared to some of the other taxa, similar to S. variabilis and S. bicanthalis (Gribbins et al., 2014). These characters also differ from some existing squamate studies. A deep acrosomal indentation is formed by S. bicanthalis, H. turcicus, and A. lineatopus. Peripheral fibers 3 and 8 extend into the principal and end pieces of H. turcicus, S. variabilis, and S. bicanthalis. An array of microtubule shapes is observed in S. undulatus and S. bicanthalis. This, in turn, supports the claim that the processes of 19 Jaime Rhodes spermiogenesis can allow for a comparative study between morphologies of different species, as some characteristics of A. costata are conserved across some studies, and different from others. Further saurian studies will be able to add to the growing list of similarities and differences between these characters in order to aid in phylogenetic classifications of lizards. The purpose of this study, to determine if steps of spermiogenesis studied in other members of the saurian taxa can be applied and compared to those of A. costata, was fulfilled. It was found that A. costata exhibits characteristics that are both present and absent in existing studies. Comparative analysis between A. costata and other species also aided in a more comprehensive understanding of the process of spermiogenesis within the lizards. It can be concluded that the characters studied in A. costata will be useful tools to continue comparative studies in other species. Moving forward, I plan to continue working with Dr. Gribbins with the comparative study of A. costata, in hopes to eventually publish this research in a peer-reviewed journal. Reflection: Personally, this study was important in order to not only gain knowledge in regards to the processes of spermiogenesis, but in order to exercise scientific practices that I believe will aid me in my future endeavors (i.e. good practice in scientific research, analyzing existing research, and practice utilizing tools such as a microtome, microscope, and working with stains). Fixation, sectioning, and staining of A. costata specimen was completed by Dr. Gribbins, graduate students, and I. This was my first experience practicing these skills. That being said, there was a learning curve. For semester II 2018-2019, I spent multiple days per week with graduate students working in Dr. Gribbins lab. I was able to observe and practice tissue embedding, sectioning, and staining. Learning how to embed tissue samples was a tedious, but 20 Jaime Rhodes interesting process that I was taught by both Dr. Gribbins and his graduate students while working in the lab. This was my first experience embedding tissues in a series of ethanol washes in order to eventually transition them into plastic. Although time consuming, graduate students guided me through the processes, which made for adequate results. Following ethanol and plastic washes, I was able to place my samples into an oven in order to properly fix them. Following this, Dr. Gribbins saturated the samples with uranyl acetate. As for sectioning tissue samples using a microtome, I found the process to be both frustrating and tedious. Dr. Gribbins sectioned the majority of my samples, as I was unable to produce adequate sections of the specimen. Tissue sectioning involves variables such as hand-steadiness, and ability to refine a microtome in order to produce clear sections. I continued to practice with other tissue samples throughout the semester, as it is a skill that I hope to master in the future. Creating my own analysis of A. costata specimen extended over the periods of semesters I and II 2019-2020. During this time, I expanded my literature knowledge in regards to spermiogenesis, and the process of labeling and analyzing photos for scientific journal presentation. From the construction of my Honors Project proposal, I had some experience analyzing TEM photos. However, in order to complete this Honors Manuscript, I needed to expand my knowledge significantly through reading existing literature and with the help of Dr. Gribbins. Specifically, the analyzing process improved my ability to pay close attention to detail, and comparatively study existing research with my own. I used my photos and closely compared and contrasted them to published work of Dr. Gribbins and his colleagues. This allowed me to prepare my findings for the discussion portion of this Manuscript. In regards to labeling my photos, this process took several attempts in order to produce adequate figures. This was my first 21 Jaime Rhodes experience using a photo editing software (similar to Photoshop). Creating appropriately sized and placed labels for each of my micrographs took many hours to complete. I spent semester II 2018-2019 presenting the preliminary research that I conducted on A. costata. I presented at the Sigma Zeta conference, annual meeting of the Indiana Academy of Science, and Butler University Undergraduate Research Conference. This was my first time presenting undergraduate research. I constructed a poster with the guidance of Dr. Gribbins and graduate students. These experiences allowed me to greatly improve my public speaking skills. I spent weeks preparing for these conferences by becoming even more familiar with the Proposal portion of my Honors Project, and organizing my findings in a way that I could present to judges. A final notable challenge that I faced while completing this Honors Project was making final edits and changes to the Manuscript throughout quarantining in my hometown of Columbus, Ohio due to COVID-19. Ideally, I would have been on campus for the duration of semester II 2019-2020, and would have access to both Dr. Gribbins and graduate students in the lab. My communication with Dr. Gribbins upon returning home was achieved solely through email, which slowed down the process of completing the manuscript. However, this newfound free time due to all classes shifting to an online format did allow me to expand my literature research even further in order to continue analyzing my results of A. costata. This Honors Project was my first experience conducting a study of spermiogenesis, and my first time conducting a comparative scientific study. Dr. Gribbins is an experienced herpetologist with a concentration in spermiogenesis, so I was fortunate enough to have a mentor that provided me with an abundance of knowledge on the subject. Through working with him 22 Jaime Rhodes over the course of my undergraduate career at the University of Indianapolis, I received an abundance of guidance that will aid me in my further schooling. References Gribbins, K. 2011. Reptilian spermatogenesis: A histological and ultrastructural perspective. Spermatogenesis 1:250-69. Gribbins K. M., Gist D. H. 2003. Cytological evaluation of spermatogenesis within the g erminal epithelium of the male European Wall Lizard, Podacris muralis. Journal of Morphology 258: 296-306. Gribbins K. M., Gist D. H., Congdon J. D. 2003. Cytological e valuation of spermatogenesis and organization of the germinal epithelium in the male Slider Turtle, Trachemys scripta. Journal of Morphology 255: 337-346. Gribbins, K. and Rheubert, J. The ophidian testis, spermatogenesis and mature spermatozoa. Reproductive Biology and Phylogeny of Ophidia, edited by Aldridge, D. and Sewer, D. 2011. SP Publishers, 183-264. Gribbins K., Matchett C., DelBello K., Rheubert J., Villagrn-Santa Cruz M., GranadosGonzlez G., Hernndez- Gallegos O. 2013b. The ultrastructure of spermatid development during spermiogenesis within the Rosebelly Lizard Sceloporus variabilis (Reptilia, Squamata, Phrynosomatidae). Journal of Morphology 274(6): 344-360. Gribbins, K. and Rheubert, J. The architecture of the testis, spermatogenesis, and mature spermatozoa. Reproductive Biology and Phylogeny of Lizards and Tuatara, edited by Rheubert, J., Siegel, D., and Trauth, S. 2014.CRC Press, 340-424. Gribbins, K., Rheubert J. L., Collier, M. H., Siegel, D. S., Sever, D. M. 2008. Histological analysis of spermatogenesis and the germ cell development strategy within the testis of the male Western Cottonmouth Snake, Agkistrodon piscivorus leucostoma. Annals of Anatomy 190(5): 461-476. Rheubert, J. L., Siegel, D. S., Venable, K. J., Sever, D. M. and Gribbins, K. M. 2011. Ultrastructural description of spermiogenesis within the Mediterranean Gecko, Hemidactylus turcicus (Squamata: Gekkonidae). Micron 42: 680690. Rheubert, J. L., Wilson, B. S., Wolf, K. W. and Gribbins, K. M. 2010. Ultrastructural study of spermiogenesis in the Jamaican Gray Anole, Anolis lineatopus (Reptilia: Polychrotidae). Acta Zoologica 91: 484494. Rheubert J., Messak J. A., Siegel D. S., Gribbins K. M., Trauth S. E., Sever D. M. 2017. Interand intraspecific variation in sperm morphology of Sceloporus consobrinus and Sceloporus 23 Jaime Rhodes undulatus (Squamata: Phrynosomatidae). Biological Journal of the Linnean Society 121(2):355364. doi:10.1093/biolinnean/blw043. Pyron, R.A., Burbrink, F.T. and Wiens, J.J., 2013. A phylogeny and revised classification of Squamata, including 4161 species of lizards and snakes. BMC evolutionary biology, 13, 93. Teixeira, R. D., Colli, G. R. and Bao, S. N. 1999a. The ultrastructure of the spermatozoa of the lizard Micrablepharus maximiliani (Squamata, Gymnophthalmidae), with considerations on the use of sperm ultrastructure characters in phylogenetic reconstruction. Acta Zoologica (Stockholm) 80: 4759. Teixeira, R. D., Vieira, G. H. C., Colli, G. R. and Bao, S. N. 1999b. Ultrastructural study of spermatozoa of the neotropical lizards, Tropidurus semitaeniatus and Tropidurus torquatus (Squamata, Tropiduridae). Tissue and Cell 31: 308317. Teixeira, R. D., Colli, G. R. and Bao, S. N. 1999c. The ultrastructure of spermatozoa of the lizard Polychrus acutirostris (Squamata, Polychrotidae). Journal of Submicroscopic Cytology and Pathology 31: 387395. Tavares-Bastos, L., Cunha, L. D., Colli, G. R. and Bao, S. N. 2007. Ultrastructure of spermatozoa of scolecophidian snakes (Lepidosauria, Squamata). Acta Zoologica (Stockholm) 88: 189197. Tourmente, M., Giojalas, L. and Chiaraviglio, M. 2008. Sperm ultrastructure of Bothrops alternatus and Bothrops diporus (Viperidae, Serpentes), and its possible relation to the reproductive features of the species. Zoomorphology 127: 241248. Appendices Appendix A- CITI Training Certificate 24 Jaime Rhodes 25 Jaime Rhodes 26 Jaime Rhodes 27 Jaime Rhodes Appendix B- Poster 28 Jaime Rhodes Appendix C- List of Items Needed/Budget Item Tissue embedding kit Embedding media Electron microscope Light microscope Obtained Microscope slides Stains A. costa tissue sample Knife machinery Lecia Ultracut S Microtome Glass rods X X X X X Appendix D- Lab Safety Cost $120 X X $250 29 Jaime Rhodes 30 Jaime Rhodes 31 Jaime Rhodes 32 Jaime Rhodes 33 Jaime Rhodes 34 Jaime Rhodes 35 Jaime Rhodes 36 Jaime Rhodes 37 Jaime Rhodes 38 Jaime Rhodes 39 Jaime Rhodes 40 Jaime Rhodes  ...

... Identifying Plant Species by Leaf Venation Patterns Using Machine Learning By Evan Parduhn Honors Project Thesis Submitted in Partial Fulfillment Of the Requirements for a Baccalaureate Degree With Distinction In the Ron and Laura Strain Honors College of THE UNIVERSITY OF INDIANAPOLIS April 18th, 2020 Faculty Advisor: Paul Talaga, Ph.D. Executive Director of Honors: James B. Williams, Ph.D i Abstract Invasive species can be devastating to native vegetation and can be difficult to identify unless properly trained. The best method for combating invasive species is to quickly identify and contain them. This project looks at a possible solution to difficult identification by creating a trained artificial neural network using TensorFlow that can categorize plant species by their venation patterns. This network could then be integrated into an application that identifies if a particular plant was invasive and allows for faster containment measures to be taken. For this project an artificial neural network was set up and tested using five different plant species; Acer campestre, Acer ginnala, Acer griseum, Acer platanoides, Acer negundo. Some of these species are invasive, but this project focused on categorizing species rather than if they are invasive or native. The images used for this project were pulled from the open-source Leafsnap dataset [1]. Leafsnap is a similar project that uses machine learning to categorize leaf species by leaf shape. The network was trained on a total of 135 images, 27 for each species, and tested with 35 images, 7 images for each species. After training, the network had an overall accuracy of 94% when categorizing by venation pattern. The preliminary results of this project show that with a larger dataset and a more refined artificial neural network, a reliable application can be created that will quickly identify invasive plants. ii Table of Contents Abstract i Table of Contents ii List of Tables iii List of Figures iii Introduction 1 Statement of Purpose 1 State of the Art & Related work 2 Current State of the Art 2 Leafsnap 3 Machine Learning 4 Justification and Honors Worthiness 6 Solution 6 Dataset 7 Neural Network 9 Discussion 10 Results 10 Reflection 11 Future Work 12 Neural Network 12 Dataset 12 Conclusion 13 Literature Cited 14 Appendix 15 Appendix A: Acknowledgements 15 Appendix B: leaf_id.py 15 iii List of Tables Table 1: Accuracy by Species 11 List of Figures Figure 1: Artificial Neural Network 5 Figure 2: Acer Campestre Lab Image 7 Figure 3: Acer Campestre Segmented Image 7 Figure 4: Acer Campestre Masked Image 8 1 Introduction Statement of Purpose While some invasive plants are distinctive and easily recognized, many others are difficult to distinguish from one or more species of our native flora[2]. That was from the book Mistaken Identity, a book about invasive plants that look similar to native ones. As more and more invasive species pop up and compete with local plants, it becomes imperative to find methods of dealing with these plants. Rejmnek states that there are three objectives to dealing with invasive species; prevention, early detection, and containment/eradication [3]. Although plant recognition may not be as helpful in preventing an invasion of a non-native species, it could become imperative in identifying and eradicating existing invasive plants. Rejmnek states that to achieve the second objective, early detection, proper field experience and relevant sampling techniques (e.g. adaptive sampling, Thompson & Seber 1996) are necessary [3]. Artificial neural networks are the leading technology in artificial intelligence and are used in many technologies such as facial recognition and self-driving. Facial recognition works by processing an image at low resolution to reject non-face regions and then evaluates the challenging regions at higher and higher resolutions [4]. The purpose of this project was to use neural networks to identify plant species by their venation patterns. It could be used to categorize invasive species and as a method for early identification and help prevent the damage caused by invasive plants. 2 State of the Art & Related work In this section I will discuss three things; how leaf identification is being done or how it has been done, describe one existing automated leaf identification method, and provide a background of machine learning. Current State of the Art There are many resources that people can rely upon to identify plants, but there is always a need for the user to either be trained in identification or have an extensive amount of time to dedicate to identification. For example there are many books and websites that catalog species with information on the plant. Go Botany[5], for example, asks you several questions about the plant and returns a list matching those descriptors. The site only covers so many features before the user has to begin comparing one by one, which can still end up being tedious and time consuming. Another issue with this site is that it categorizes plants by location, which could cut out species that might not normally exist in that area such as invasive species [5]. A method of finding invasive plants and determining impact in the area is aerial hyperspectral photos. Hyperspectral imaging works by collecting information from the electromagnetic spectrum. Certain objects have spectral fingerprints, which can be used to identify a certain material. After an aerial hyperspectral image is taken, someone can map where and how widespread invasive species are [6].This method of identification is crucial for early detection and containment. This method is not efficient because the cost for taking 3 hyperspectral images is high and an expert that can differentiate species is required to create the mappings. The methods currently being used are not ideal for curtailing invasive species because they are either too slow or require extensive training. A possible solution to this problem is to create an application that categorizes plant species automatically from cell phone images. Leafsnap An Automated Identification Tool One such technology for plant identification that is already widely available is Leafsnap [1]. A leaf recognition app that was developed using facial recognition software and artificial intelligence. Kumar states why this software could help save time: Without visual recognition tools such as Leafsnap, a dichotomous key (decision tree) must be manually navigated to search the many branches and seemingly endless nodes of the taxonomic tree. Identifying a single species using this process by answering dozens of oftenambiguous questions, such as, are the leaves flat and thin? may take several minutes or even hours. This is difficult for experts, and exceedingly so (or even impossible) for amateurs. [1]. Manual leaf identification requires extensive training and experience before someone can be reliably trusted to identify a species. An artificial neural network, on the other hand, would help speed up the identification of invasive plants and also make it easier for amateurs to find invasive plants. Leafsnap works by finding the outline of the leaf from an image, extracting curvature based features, and then identifies the leaf using a nearest neighbor search algorithm[1]. The nearest neighbor algorithm starts at one point in a map of comparable data, where each node is 4 connected by similar features to another node. Data that is extracted from the leaf is then compared to the nodes connected to the starting point and moves the node with the closest match. This repeats until a threshold for a certain match is reached. Leafsnap then returns the top 25 matches and boasts that 96.8% of queries have a match in the top five results. Leafsnap is able to discern 189 different species, so they are able to reduce the possible species by 98% and have a 96.8% accuracy. Our work used a much smaller number of species, so our performance will be based on our top predicted species rather than top 5. Machine Learning Leafsnap solves the problem of identification by implementing an algorithm that compares existing data on features to an unidentified leaf. Another possible method for solving this problem is using an artificial neural network. Essentially, a machine is given a number of Tasks (T), which it is then rated on by its performance (P). The machine is supposed to improve with experience (E) [7]. For example, if we wanted to train a machine to differentiate between plant leaves, the machines tasks would be to take a picture and identify it. The machine would then be rated by whether it got the identification right or wrong. If the machine gets it right, it is rewarded making that outcome more likely. If the machine is incorrect, the connections leading to that outcome are inhibited and make the outcome less likely. In comparison, Leafsnap uses a nearest neighbor approach, which is equivalent to a neural network with no hidden layers which can only separate species if their feature space is linearly separable. Neural networks with hidden(middle) layers are able to discern a feature space with more complex boundaries. 5 In a little more detail, we understand that the machine is a neural network, where each node takes multiple inputs, and has one output. So by having many nodes at the beginning and slowly decreasing the number of nodes, we can filter a lot of binary input into one simple output. Each of these nodes has weights, which really means how much is needed to flip the nodes switch and change the output [8]. This is shown in Fig 1. Each of the circles is a node, and each of the arrows is a connection to another node. Each of these connections has a weight to it that determines where the information gets sent next until it reaches the end. The node in the output layer that has the most activation is the answer. Figure 1: Artificial Neural Network For this project the number of nodes at the beginning is 256. This will then boil down to five nodes at the end, matching the number of species this network was being trained on. Data is sent to each node, weighted to decide what node it will send this data to next. This continues until one of the final nodes is reached, which is the machine's final answer. When training, if the network outputs a wrong answer, there is a feedback mechanism that weakens the weights of the nodes associated with that incorrect answer. This makes that specific wrong answer less likely to happen. The same feedback mechanism occurs when the network outputs a right 6 answer. This time, the feedback mechanism strengthens those paths making the answer more likely to happen. After many training samples the network should learn and improve its performance. Justification and Honors Worthiness Invasive species have been a plight in the world for as long as people have been travelling. These species take away from native plants resources, leaving native plants starved and dying. From the U.S. Forest Service, Invasive species have contributed to the decline of 42% of U. S. endangered and threatened species [9]. The only way to prevent this from happening is early detection and eradication. One issue with early detection is that many people are not trained to differentiate plants and some invasive plants look similar to native plants. This issue could be ameliorated by using a neural network, which is easier and faster to train than humans, to identify these plants. I chose to create a system that identified plant species by their venation patterns. Based on my research, I found that most currently available applications identify leaves based on the outline of the leaf. Through this research I will determine if plants can be identified by their venation patterns, and if this method of identification could be used and incorporated with existing methods to bolster accuracy. Solution This project had two goals. The first goal was processing the Leafsnap dataset so that the leaf is isolated from the background. The second goal was to create the neural network and train it to identify plant species using the vein patterns in the leaf. 7 Dataset Figure 2: Acer Campestre Lab Image Figure 3: Acer Campestre Segmented Image The network I created needed to categorize leaves by their venation patterns. As it is, the Leafsnap dataset is not conducive to this type of categorization and needs to be changed to fit those needs. The Leafsnap dataset contains the original picture taken in a lab setting and a segmented image of that picture that Leafsnap uses in their current system. Figure 2, above, is an example of an original lab image. This image is not ideal for training a network because there is a lot of superfluous data such as the color bars on the bottom and right side. Since the network needs to categorize by vein patterns, this extra data might interfere with the categorization process and produce a network that identifies the leaves by the color bar instead of the vein pattern. Figure 3, above, is an example of a segmented image. Figure 2 was processed by Leafsnaps application and turned into Figure 3. This allowed Leafsnap to identify the species by the shape of the leaf. Figure 3 is ideal for Leafsnaps system, but not for this system because it got rid of the venation patterns in Figure 2. This project needed an image that 8 preserved the contents of the leaf but got rid of any arbitrary data around the leaf. To accomplish this, the segmented images were used as masks for the original image. Figure 4: Acer Campestre Masked Image Figure 4, above, is an example of what we have done to the dataset to make it ideal for the network I created. Figure 4 was created using a software called GIMP, an open source image editing software. GIMP has a feature called Python-Fu which allows for a script that uses GIMP functions to run on multiple images. The script I wrote takes the original image and the segmented image and puts them on top of each other, resizing when necessary. Then the script creates a grayscale mask of the segmented image and inverts it. This cuts out the white space 9 and leaves a cutout of the leaf. Then the original image and the mask are combined to create something similar to Figure 4. This script was run for all of the images in the Leafsnap dataset. Once the dataset was processed to meet the needs of the network, a script was developed that created a neural network, generated a histogram of the images, and trained the network using those histograms, described below. Neural Network There were two types of data that were used as input to the neural network; a histogram of the image, and a ratio of the area of veins to the total area of the leaf. A histogram of the image gets the color distribution of the image. Since the veins of the leaf are a much different color than the rest of the leaf, a histogram will show the concentration of veins to the rest of the leaf. The ratio contains the same information, but it is more focused. We needed to keep both because the ratio does not contain all of the veins and the histogram is not as focused on vein patterns. The histogram was generated by flattening an image array into a list of numbers. The ratio was generated by finding the area of the veins and dividing that by the total area of the leaf. To find the area of the edges, an image was first sent to Canny, a function from the library cv2. Canny detects edges by finding a large difference in color and then creates a black and white image of just the edges. These two pieces of data were chosen because they both contain a lot of focused information. The neural network is trying to categorize the information that is given to it and it decides what is useful or not within the information. For example, if I try to train a neural 10 network to identify whether something in an image is either a dog or a dolphin, the neural network might use information in the background of the image, rather than the animal. Most pictures of dolphins are going to have a blue background because they live in the ocean. Pictures of dogs, on the other hand, will most likely not have a lot of blue in the background. If this data is not filtered out, the neural network might decide that all images with a blue background have dolphins in them. That is why it is important to choose data that is relevant to what the neural network should be trained on. A histogram of the image preserves the color distribution and contains information of just the leaf on vein concentration. The ratio contains that concentration as well and has the bonus of being impervious to changes in the zoom of the image. Discussion Results Five different species were arbitrarily chosen to train and test the network. In total, there were 135 samples in the training set, 27 per species, and 35 samples in the testing set, 7 per species. The network achieved a 94% accuracy with its first choice. Since any given network can be saved and used again, there is no need to discuss average accuracy across different networks. Instead we will discuss the network that achieved the highest accuracy. Acer campestre Acer ginnala Acer griseum Acer platanoides Acer negundo 100% 71% 100% 100% 100% Table 1: Accuracy by Species 11 Table 1 shows the accuracy that the neural network makes the correct match. Acer ginnala is the only leaf that the network does not categorize correctly. The network sometimes categorizes an Acer ginnala leaf as either Acer platanoides or Acer negundo. There is no discernible reason for this error. Reflection This project set up a dataset that can be used for future machine learning applications and created a neural network that can identify three species at an 94% success rate. The original goal for this project was to use machine learning to create an application that identifies leaves as either invasive or non-invasive. This project is a large step in the direction of this goal since it sets up an artificial neural network that can be trained to identify leaves. The missing link is a larger dataset that includes leaves in different orientations and ones that are taken in the field. The reason that I was not able to get to the application phase of this project was largely due to the issues faced while creating the dataset. The original goals were created with the assumption that I would be able to use the Leafsnap dataset right out of the box. However, the Leafsnap dataset was not formatted in a way that I could easily set up a neural network. This meant that I had to spend much more time analyzing the dataset and creating a new dataset that allowed for training based on vein patterns. Through this issue, though, I have learned a lot about image processing and using GIMP scripts as a tool to process a lot of images quickly and efficiently. This is a valuable skill for anyone to have and could be useful for me specifically as I enter the professional world where this skill would set me apart. I have also learned a lot about machine learning, and specifically about how to use Tensorflow, which is a machine learning tool. These skills are valuable because there are many 12 problems in the world that could be solved with machine learning. Now that I have the skillset to use this tool, I am better prepared to face professional problems that machine learning can solve. Future Work Neural Network Currently the maximum accuracy that the neural network can reach is 94% with five species. This is comparable to Leafsnap as they claim that they have a 96% accuracy within the first five results across 189 different species. The next steps for this project on the machine learning side will be to train a network with more species and ensure that an accuracy rating of 94% or higher is maintained. It might be beneficial to explore different types of data for the training set which would raise the accuracy. This would include finding more information about the veins, such as the angles between them, their length, and their curve. This data could be difficult to find as lines are harder to map across pixels, but it would be very valuable and could raise the accuracy. Dataset The biggest next step for this project would be obtaining a more comprehensive dataset that includes other types of images. Currently the data set I have created includes only lab images. These images ensured the leaves were flat in varying degrees of light which allow for the best data to be extracted. This network would fail on images that were taken in the field 13 that are either curved, have an indistinguishable background, or do not contain the contrast necessary to retrieve useful data. Conclusion Invasive species have a destructive impact on native species that can cause plant species to become endangered. The methods that are currently being used to contain invasive plants rely on early detection. Without extensive training, properly identifying invasive species is not fast enough. This project is a step towards solving the problem of not being able to identify invasive species quickly or accurately. This project uses machine learning to train an artificial neural network to categorize five species by their venation patterns. The network achieved an average accuracy rating of 94% which is 2% lower than the Leafsnap project. To complete this project, a larger dataset is needed that includes more image types and more invasive species. Once a more comprehensive dataset has been compiled, an application can be made that could be used by anyone to identify invasive species and help stop their plight. 14 Literature Cited [1] "Leafsnap: A Computer Vision System for Automatic Plant Species Identification," Neeraj Kumar, Peter N. Belhumeur, Arijit Biswas, David W. Jacobs, W. John Kress, Ida C. Lopez, Joo V. B. Soares, Proceedings of the 12th European Conference on Computer Vision (ECCV), October 2012. [2]Sarver, Matthew, et al. Mistaken Identity?: Invasive Plants and Their Native LookAlikes: an Identification Guide for the Mid-Atlantic. Delaware Department Agriculture, 2008. [3] Rejmnek, Marcel. "Invasive plants: approaches and predictions." Austral ecology 25.5 (2000): 497-506. [4] Shah, Jay. Neural Networks for Beginners: Popular Types and Applications. Stats and Bots, Stats and Bots, 16 Nov. 2017, blog.statsbot.co/neural-networks-for-beginnersd99f2235efca. [5] Native Plant Trust, 14 April 2020, https://gobotany.nativeplanttrust.org/simple/woodyplants/woodyangiosperms/#_filters=family,genus,habitat_general,state_distribution,plant_habit_wa,l eaf_type_general_wa,leaf_arrangement_wa,leaf_blade_margin_general_wa,leaf_durati on_wa,armature_wa&habitat_general=terrestrial&plant_habit_wa=tree&leaf_type_gen eral_wa=simple&leaf_arrangement_wa=1%20leaf%20per%20node&leaf_blade_margin _general_wa=toothed&leaf_duration_wa=deciduous%20or%20marcescent&armature_ wa=absent&_view=photos&_show=plant%20form 15 [6] Lawrence, Rick L., Shana D. Wood, and Roger L. Sheley. "Mapping invasive plants using hyperspectral imagery and Breiman Cutler classifications (RandomForest)." Remote Sensing of Environment 100.3 (2006): 356-362. [7] Goodfellow, Ian, Yoshua Bengio and Aaron Courville. Deep Learning. Cambridge, Massachusetts, MIT Press, 2016. http://www.deeplearningbook.org [8] Nielsen, Michael A. Neural networks and deep learning. Vol. 25. USA: Determination press, 2015. [9] Invasive Plants. Forest Service Shield, U.S. Forest Service, www.fs.fed.us/wildflowers/invasives/. Appendix Appendix A: Acknowledgements I would like to acknowledge Dr. Paul Talaga for his help on this project. As my advisor he has been a great resource for troubleshooting or just to discuss new ideas. This project would not have been possible without him. Appendix B: leaf_id.py # -*- coding: utf-8 -*"""leaf_id.ipynb Author: Evan Parduhn Description: This script is used to train an artificial neural network to categorize leaves according to their venation patterns Automatically generated by Colaboratory. Original file is located at https://colab.research.google.com/drive/1goioTPOEeN8p4QkvPZLDrbbJsnrn_BoS """ 16 from __future__ import absolute_import, division, print_function, unicode_literals # Install TensorFlow import tensorflow as tf from google.colab import drive import os import math import numpy as np import cv2 import pandas as pd import matplotlib.pyplot as plt import PIL # This needs to be catered to the users Google drive. the ls only serve as a manual check drive.mount('/content/drive') !ls "/content/drive/Shared drives/" !ls "/content/drive/Shared drives/EvanLeafSnap/" !ls "/content/drive/Shared drives/EvanLeafSnap/leafsnap-dataset" !ls "/content/drive/Shared drives/EvanLeafSnap/Vein_Identifier" # This function takes a filename and parses out the extra characters and returns just the name # of the species. This was made specifically for the data set I created so it may need to be changed def parseName(filename): fl = filename[:-5] if(fl[-2].isdigit()): fl = fl[:-2] elif(fl[-1].isdigit()): fl = fl[:-1] return fl fp = "/content/drive/Shared drives/EvanLeafSnap/Vein_Identifier/Test" iname = "" im_dict = {} # Keeps track of the number samples for each species found tr_dict = {} # Contains a list of the samples for each species in the training list te_dict = {} # Contains a list of the samples for each species in the testing list dy = {} # This is the answer key i=0 for filename in os.listdir(fp): fl = parseName(filename) 17 if(fl not in im_dict.keys()): im_dict[fl] = 1 tr_dict[fl] = [] te_dict[fl] = [] dy[fl] = i i+=1 elif(fl in im_dict.keys()): im_dict[fl] += 1 # This find the minimum number of samples to a species # I did this so that the number of samples across both lists would remain the samples # to hopefull prevent scewed results i_min = min(im_dict.values()) train_num = int(i_min * .8) test_num = int(i_min * .2) # print(train_num,test_num) Used to debug # print(dy) # Finds the histogram of and image and returns it def genHist(fn): im = PIL.Image.open(fn) im = np.array(im) hist,bins = np.histogram(im,bins=255, range=(0,255)) return hist # Finds the edges in an image and creates a new image with just the edges # Returns the sum of the new image which is the area of the edges def genEdges(fn): img = cv2.imread(fn) gray = cv2.cvtColor(img,cv2.COLOR_BGR2GRAY) edges = cv2.Canny(gray,20,50,apertureSize = 3) return np.sum(np.asarray(edges)) # Uses genEdges to find the area of veins and returns that over the area of the leaf def genRatio(fn): im = PIL.Image.open(fn) im = np.array(im) area = 0 18 for arr in im: for p in arr: if(p.any() != 0): area += 1 return int(genEdges(fn) / area) curr_name = "" # x arrays contain samples, y arrays contain the names x_train = [] x_test = [] y_train = [] y_test = [] # Evenly distributes samples into the training and testing arrays for filename in os.listdir(fp): fl = parseName(filename) if(len(tr_dict[fl]) < train_num): x_train.append(np.append(genHist(fp + '/' + filename),genRatio(fp + '/' + filename))) y_train.append(dy[fl]) tr_dict[fl].append(genHist(fp + '/' + filename)) elif(len(tr_dict[fl]) >= train_num and len(te_dict[fl]) <= test_num): x_test.append(np.append(genHist(fp + '/' + filename),genRatio(fp + '/' + filename))) y_test.append(dy[fl]) te_dict[fl].append(filename) # You have to convert these to numpy arrays so they work with TensorFlow # Yes I know that numpy has an append function, but its confusing and I hate it x_train = np.asarray(x_train) x_test = np.asarray(x_test) y_train = np.asarray(y_train) y_test = np.asarray(y_test) # print(len(x_train),len(x_test),len(y_train),len(y_test)) # for debugging # These are useful functions if you want to send whole images to TensorFlow # I did not do that for the final implementation, but they are here for future implementations # Finds and returns the max shape in the training and testing arrays def findMaxShape(arr_tr, arr_ts): ms = [0,0] 19 for im in arr_tr: if(im.shape[0] > ms[0]): ms[0] = im.shape[0] if(im.shape[1] > ms[1]): ms[1] = im.shape[1] for im in arr_ts: if(im.shape[0] > ms[0]): ms[0] = im.shape[0] if(im.shape[1] > ms[1]): ms[1] = im.shape[1] return ms # adds padding to the samples according to the max shape returned by findMaxShape # Funny how the function names say exactly what they do # This is necessary because TensorFlow likes its training samples to all be the same size def addPadding(arr_tr, arr_ts): ms = findMaxShape(arr_tr,arr_ts) n_tr = [] n_ts = [] for im in arr_tr: hd = ms[0] - im.shape[0] wd = ms[1] - im.shape[1] nim = np.pad(im,((0,hd),(0,wd)),'minimum') n_tr.append(nim) for im in arr_ts: hd = ms[0] - im.shape[0] wd = ms[1] - im.shape[1] nim = np.pad(im,((0,hd),(0,wd)),'minimum') n_ts.append(nim) return np.asarray(n_tr),np.asarray(n_ts) # print(len(x_train),len(x_test),len(y_train),len(y_test)) # more debugging # Creates the training model in TensorFlow. Trains and then tests. I also added a predict # to see more detailed results model = tf.keras.models.Sequential([ tf.keras.layers.Flatten(input_shape=(256,)), tf.keras.layers.Dense(128, activation='relu'), tf.keras.layers.Dropout(0.2), 20 tf.keras.layers.Dense(len(dy), activation='softmax') ]) model.compile(optimizer='adam', loss='sparse_categorical_crossentropy', metrics=['accuracy']) model.fit(x_train, y_train, epochs=200) model.evaluate(x_test, y_test, verbose=2) model.predict(x_test)  ...