Bumblebees' Bombus Impatiens

Bumblebees’ Bombus impatiens (Cresson) Learning: An Ecological Context

Hamida B. Mirwan

A Thesis Presented to The University of Guelph

In partial fulfillment of requirements for the degree of Doctor of Philosophy in School of Environmental Biology

Guelph, Ontario, Canada

ABSTRACT

BUMBLEBEES’ BOMBUS IMPATIENS (CRESSON) LEARNING: AN ECOLOGICAL CONTEXT

Hamida B. Mirwan Co-Advisors: University of Guelph, 2014 Professors Peter G. Kevan & Jonathan Newman

The capacities of the bumblebee, Bombus impatiens (Cresson), for learning and cognition were investigated by conditioning with increasingly complex series of single or multiple tasks to obtain the reinforcer (50% sucrose solution). Through operant conditioning, bumblebees could displace variously sized combinations of caps, rotate discs through various arcs (to 180°), and associate rotation direction with colour (white vs. yellow). They overcame various tasks through experience, presumably by shaping and scaffold learning. They showed incremental learning, if they had progressed through a series of easier tasks, single caps with increasing displacement complexity (to left, right, or up) or of balls with increasing masses, but could not complete the most difficult task de novo. They learned to discriminate the number of objects in artificial flower patches with one to three nectary flowers presented simultaneously in three compartments, and include chain responses with three other tasks: sliding doors, lifting caps, and rotating discs presented in fixed order. Pattern recognition and counting are parts of the foraging strategies of bumblebees. Multiple turn mazes, with several dead ends and minimal visual cues, were used to test the abilities of bumblebees to navigate by walking and remember routes after several days. They rapidly learned these mazes and remembered the routes even after

16 days. Bumblebees could learn from each other, socially by imitation, observation, and communication within the nest. They were slower to learn to forage with dead conspecific models than with living ones, whether nest-mate or non-nest-mate. Bees that had no opportunity for social learning were unable to forage. The array of experimental approaches I used for training bumblebees and my results have expanded the scope and understanding of the complexity of invertebrate learning and cognition in the context of comparative psychology, and have application to ecological and ethological principles of the evolution of learning and cognition. ACKNOWLEDGEMENTS

I am thankful to Allah for giving me the health and the strength to finish my study. I am grateful to my country, Libya, for the scholarship granted to me to pursue doctoral studies, and the University of Guelph for the opportunity to do graduate work. My deepest thankfulness goes to my supervisor Professor Dr. Peter Kevan for accepting me and for his brilliance and knowledge from which I benefited during the past years and I have had the privilege to work with him and will forever be grateful for that. His dedication and commitment to science and education is truly inspiring and remarkable.

Special thanks to the members of my graduate committee for great support and for sharing their knowledge on study techniques: Dr. Jonathan Newman; Dr. Georgia Mason for her support, and discussion of several issues, and help with the edition of the final document; Dr.

Ernesto Guzman.

I am grateful to have Dr. Sarah Bates of being always available for questions and help,

Dr. Thomas Woodcock for his time to review some of my work, and John Charles for his help with my final work of my thesis. I also thank Joy Roberts who was the secretary of the graduate studies, and Dr. Hung Lee and Dr. Paul Sibley who were the graduate coordinator, of the School of Environmental Sciences, University of Guelph, for their help of administrative issues during my study, Dr. Francesco Leri at Psychology Department, University of Guelph for his help and comments on some of my papers, Dr. Les Shipp for giving the opportunity to work and gain

Canadian experience at Agriculture and Agri-Food Canada at Harrow Research Centre, Harrow,

Ontario, and Dr. Marisol Amaya Márquez at National University of Colombia. Finally, I thank all the people who work and manage the Honeybee Research Centre at the University of Guelph,

Guelph, Ontario, especially Paul Kelly for providing the food supplements for my bees during all my experiments, and BioBest Biological System, Leamington, Ontario, Canada for providing the bees during all period of my study.

Special thanks to my mother and other members of my family, who believed in and supported me through my doctoral studies. Deep thanks to Mohsin Grera who always helped, supported and encouraged me to find strength and hope. I thank other people who taught me, all my friends for their support and solidarity with me during these years for their friendship.

Financial support of this work was by The Ministry of Higher Education and Scientific

Research, Tripoli- Libya, NSERC, CANPOLIN, and The University of Guelph, Guelph, Ontario,

Canada.

DEDICATION

I dedicate this dissertation: To My father Bashir Mirwan, who always engaged and taught me that, the best kind of knowledge is that which is learned for its own sake. To My mother, who taught me that, even the largest task can be accomplished if it is done one step at a time. To My brother Najme, and my nephew Abdul Aziz.

TABLE OF CONTENTS

ACKNOWLEDGEMENTS ...... IV

DEDICATION ...... VI

LIST OF TABLES...... XI

LIST OF FIGURES ...... XII

CHAPTER 1 INTRODUCTION ...... 1

1.1 Introduction ...... 1 1.1.1 Learning definitions ...... 2 1.1.2 Definitions of learning found through a review of the literature ...... 3 1.1.3 The benefits and the cost of learning ...... 4

1.2 Innate behaviour ...... 6

1.3 Learning ...... 6 1.3.1 Non Associative learning ...... 6 1.3.2 Associative learning ...... 7 1.3.3 Cognitive learning ...... 10 1.3.3.1 Insight learning ...... 10 1.3.3.2 Problem Solving...... 10 1.3.3.3 Conditional discrimination and response chaining ...... 13 1.3.3.4 Navigation and maze navigation ...... 16 1.3.3.5 Social learning ...... 18 1.3.4 Objectives ...... 21 1.3.5 Overarching Hypothesis ...... 21

1.4 Glossary ...... 22

CHAPTER 2 COMPLEX OPERANT LEARNING BY WORKER BUMBLEBEES (BOMBUS IMPATIENS CRESSON (HYMENOPTERA: APIDAE)) ...... 25

2.1 ABSTRACT ...... 25

2.2 INTRODUCTION ...... 26

2.3 MATERIAL AND METHODS ...... 27 2.3.1 Experimental set up ...... 28 2.3.1.1 Experimental procedure ...... 30 2.3.2 Operant task 1: Sliding single and “multi” caps (Fig. 2) ...... 30

vii

2.3.2.1 Artificial flowers: Artificial flowers were made from centrifuge tubes (1.5 ml) and 9 cm ...... 31 2.3.2.2 Experimental procedure ...... 31 2.3.3 Operant task 2: Rotating an upper disc ...... 33 2.3.3.1 Operant conditioning 2A: Rotatable an upper blue discs (Fig. 2.3) ...... 33 2.3.3.1.1 Artificial flowers ...... 33 2.3.3.1.2 Experimental procedure ...... 34 2.3.3.2.1 Experiment 2B 1 ...... 35 2.3.3.2.2 Experiment 2B 2 ...... 36 2.3.3.2.3 Experiment 2B 3 ...... 37 2.3.4 Data analysis ...... 37

2.4 RESULTS ...... 40

2.5 DISCUSSION ...... 48

CHAPTER 3 PROBLEM SOLVING BY WORKER BUMBLEBEES BOMBUS IMPATIENS (HYMENOPTERA: APOIDEA) ...... 52

3.1 ABSTRACT ...... 52

3.2 INTRODUCTION ...... 53

3.3 MATERIAL AND METHODS ...... 54 3.3.1 General Methods ...... 54 3.3.1.1 Experimental set-up ...... 56 3.3.1.2 Artificial flowers ...... 57 3.3.1.3 Experimental procedure ...... 61 3.3.2 Experiments ...... 61 3.3.2.1 Experiment 1: sliding and lifting caps (Fig. 3.2) ...... 61 3.3.2.2 Experiment 2: pushing balls (Fig. 3.3) ...... 62 Experiment 2a: increasing of the mass of the ball (six balls of increasing masses) ...... 62 Experiment 2b: Increasing of the mass of the ball (three balls of increasing masses) ...... 63 3.3.3 Data analysis ...... 63

3.4 RESULTS ...... 64

3.5 DISCUSSION ...... 71

CHAPTER 4 CONDITIONAL DISCRIMINATION AND RESPONSE CHAINS BY WORKER BUMBLEBEES (BOMBUS IMPATIENS CRESSON (HYMENOPTERA: APIDAE)) ...... 75

4.1 ABSTRACT ...... 75

4.2 INTRODUCTION ...... 76

viii

4.3 MATERIAL AND METHODS ...... 78 4.3.1 General methods ...... 78 4.3.2 Experiment 1: Conditional discrimination: Testing bumblebees’ abilities to recognize elements in a numerical array ...... 80 4.3.2.1 Artificial flowers ...... 80 4.3.2.2 Experimental set up ...... 80 4.3.2.3 Experimental Procedure ...... 81 4.3.3 Experiment 2: Response chains: Bumble bees manipulating multiple obstacles ...... 85 4.3.3.1 Material ...... 85 4.3.3.2 Experimental set up ...... 86 4.3.3.3 Experimental procedure ...... 86 4.3.4 Data analysis ...... 89

4.4 RESULTS ...... 89 4.4.1 Conditional discrimination: Recognizing the number of nectaries (Experiment 1) ...... 89 4.4.2 Response chains: Manipulating multiple obstacles ...... 90

4.5 DISCUSSION ...... 95

CHAPTER 5 MAZE LEARNING AND ROUTE MEMORIZATION BY WORKER BUMBLEBEES (BOMBUS IMPATIENS (CRESSON) (HYMENOPTERA: APIDAE) ..... 99

5.1 ABSTRACT ...... 99

5.2 INTRODUCTION ...... 99

5.3 MATERIAL AND GENERAL METHODS ...... 102 5.3.1 Experimental set-up ...... 102 5.3.1.1 Description of the Maze (Fig. 5.2) ...... 105 5.3.1.2 Experiment procedure ...... 107 5.3.2 Experiments ...... 108 5.3.2.1 Experiment 1: Maze Navigation ...... 108 5.3.2.2 Experiment 2: Maze re-navigation after several days ...... 109 5.3.3 Data analysis ...... 109

5.4 RESULTS ...... 109

5.4.1 Possible Influence of Trail-marking ...... 109 5.4.2 Maze Navigating ...... 110 5.4.3 Maze re-navigation (remembering the maze) after several days ...... 115

5.5 DISCUSSION ...... 123

CHAPTER 6 SOCIAL LEARNING IN BUMBLEBEES (BOMBUS IMPATIENS): WORKER BUMBLEBEES LEARN TO MANIPULATE AND FORAGE AT

ARTIFICIAL FLOWERS BY OBSERVATION AND COMMUNICATION WITHIN THE COLONY...... 128

6.1 ABSTRACT ...... 128

6.2 INTRODUCTION ...... 129

6.3 MATERIAL AND METHODS: ...... 130 6.3.1 Experimental Procedures ...... 130 6.3.1.1 Artificial flowers: ...... 131 6.3.1.2 The experimental groups: ...... 134 6.3.2 Experiments: ...... 134 6.3.2.1 Experiment 1: Control experiment ...... 135 6.3.2.2 Experiment 2: Using a model dead bee and observing nest-mate ...... 135 6.3.2.3 Experiment 3: Observing none nest-mate and communication possibilities ...... 136 6.3.3 Data Analyses:...... 138

6.4 RESULTS ...... 138

6.5 DISCUSSION ...... 148

CHAPTER 7 CONCLUSION ...... 151

REFERENCES ...... 160

LIST OF TABLES

Table 2.1 Types of sliding and lifting artificial flowers occluded by caps, as used in

Experiment 1……………………………………………………………………………………64

Table 2.2 Types of pushing balls artificial flowers occluded by balls. Subject bees were inexperienced, except for foraging successfully on simple artificial flowers without barriers to the reinforcer (INE) or experienced (E) for Experiment 2a and Experiment 2b……………….65

Table 4.1 Learning rates of worker bumblebees for the different obstacles to be manipulated as calculated by the difference in time between first and tenth encounters, and by the exponent of the power function for the learning curve over 10 trials………………………………………..95

2 Table 5.1 Chi-square values (χ ) of first turn choice on entering the maze (H0: equal likelihood of turning left or right) for the first and last 5 trials of three colonies combined (cells A & B) and separate (cells C – N), for tested on day 1, and re-tested on days 6, 11 and 16………………121

Table 7.1 Statistical values from repeated one-way Analysis of Variance of the findings from experiments in which dead bees were used as models to aid in the learning process for foraging by living bee, in which living bees were able to watch other living bees (nest-mates and non- nest-mates) forage to aid in the learning process, and in which living bees which had no opportunity to observe models or other living bees foraging learned to forage by within-colony communication………………………………………………………………………………….145

LIST OF FIGURES

Figure 2.1 Experiment setup with hive, holding area, flight cage testing arena, patch of artificial flowers and mesh tube routes with gates by which the bees were allowed to enter and exit the flight cage………………………………………………………………………………………...28

Figure 2.2 Operant conditioning 1: Multi capped artificial flowers. The tube’s cap was cut off and attached to the stick of a cotton swab (3 cm long + 1.8cm cap of the centrifuge tube). The cotton end was pinned down as the anchor/hinge at the edge of the flower to allow the cap to be swung to the left or right…………………………………………………………………………31

Figure 2.3 Operant conditioning 2A: Rotatable disc artificial flowers. Artificial flowers were made with centrifuge tubes (1.5 ml), and 7cm blue acetate disc………………………………...37

Figure 2.4 Operant conditioning 2B. Artificial flowers of rotatable white and yellow upper discs, operant condition 2B (similar to those shown in Figure 3) with white and yellow Eva foam upper discs………………………………………………………………………………………………38

Figure 2.5 Operant conditioning 1 sliding multi caps. The mean (±SE) of the durations (seconds) taken by the bees, after landing at the centre of the disc (time = 0), to slide capped or multi capped flowers and start feeding, for 10 trials, each trial being a single flowers’ patch visit

(N=16)……………………………………………………………………………………………40

Figure 2.6 Operant conditioning 2 experiment 2A. The mean (±SE) of the durations (seconds) taken by the bees, after landing at on the lower disc (time = 0), to rotate the Eva foam, upper, disc over 5 arcs for 10 trial………………………………………………………………………42

Figure 2.7 Operant conditioning 2B, experiment 2B1. The mean (±SE) of the durations (seconds) taken by the bees, after landing, to rotate coloured discs 45° in their correct direction………………………………………………………………………………...………..44

xii

Figure 2.8 Operant conditioning 2B, Experiment 2B2. The mean (±SE) of the durations

(seconds) taken by the bees, after landing, to rotate coloured discs 45° in their correct direction, white discs clockwise and yellow discs counter clockwise……………………………………..45

Figure 2.9 Operant conditioning 2B, Experiment 2B2. The proportion of correct choices the experimental bees made in rotating discs on either yellow (= rotate right) or white (= rotate left) in array (patch) of both artificial flower colours (four of each) presented simultaneously the day after they had been initially shaped in arenas of four flowers of single colours (yellow or white)……………………………………………………………………………………………46

Figure 3.1 Experimental set-up with hive, holding area, flight cage testing arena, patch of artificial flowers and mesh tube routes with gates by which the bees were allowed to enter and exit the flight cage………………………………………………………………………………54

Figure 3.2 Artificial flowers A and their increasing complexity. Sliding caps……………..…58

Figure 3.3 Artificial flowers B and their increasing complexity. Balls of different massed occluding the syrup reservoir………………………………………………………………...... 59

Figure 3.4 Sliding and lifting caps. a Shows the mean time of sliding and lifting caps ±(SE) for ten trials on artificial flowers. b Shows the mean time of lifting caps ±(SE) for ten trials on type e) artificial flowers IEN inexperienced bees (N = 4) and caps upside down and could be lifted but with difficulty because of the lack of anywhere to gain purchase E: experienced bees

(N = 12)…………………………………………………………………………………………66

Figure 3.5 The mean (±SE) of the duration (seconds) taken by the bees, after landing, to push balls away from the centrifuge tube entry……………………………………………………...67

Figure 3.6 The mean (±SE) of the durations (seconds) taken by the bees, after landing, to push balls away from the centrifuge tubes entry…………………………………………………….68

xiii

Figure 4.1 a) Flower patch A and B, b) Experiment setup with hive, holding area, flight cage testing arena divided into a three compartment maze…………………………………………..83

Figure 4.2 Bumble bees solving sequential obstacles experiment setup: a) front view of the black box, b) pushing up cap artificial flower, c) Eva foam rotatable white disc mounted on the testing arena wall to block the entry, d) Experiment setup with hive, holding area, flight cage testing arena, an artificial flower patch, mesh tube routes with gates by which the bees were allowed to enter and exit the flight cage……………………………………………………………………87

Figure 4.3 The mean number of bees (N=24 bees in each of 10 trials) landing on one of the three training flower types. The flower type rewarded depended on the number of nectaries (1N, 2N and 3N (tubes)) and the compartment (Fig. 1) in which they were presented ………………… 91

Figure 4.4 The mean (±SE) of the durations (seconds) taken by the bees, to manipulate each obstacle (N=10) during 10 trials and power function best fits to the curves of results…………92

Figure 5.1 Experiment setup showing hive, holding area, flight cage testing arena, feeding area and mesh tube (tunnels) routes with gates by which the bees were allowed to enter and exit the flight cage……………………………………………………………………………………….103

Figure 5.2 Multiple-turn maze with several dead ends made from a solid block of white high density polyethylene so that it had the same white colour for the walls and the floor…………105

Figure 5.3 The mean (±SE) of the durations (seconds) taken by the bees from three different colonies to navigate through the maze from their first encounter (trial 1) to their 20th trial…...111

Figure 5.4 The mean (±SE) numbers of mistakes made by bees from three different colonies when navigating through the maze. The findings are presented by groups of trials (trials 1 - 4 combined; 5 - 8 combined; to 17 – 20 combined)…………………………………………..….112

xiv

Figure 5.5 The percent (±SE) of bee-choices for the correct (to the right) first turn by tested bees entering the maze. presented by groups of trials (trials 1 - 4 combined; 5 - 8 combined; to 17 – 20 combined). See also results of statistical tests in Table 1, cells A and B……………………..113

Figure 5.6 The means (±SE) of the durations (seconds) taken by groups of bees to navigate through the maze………………………………………………………………………………115

Figure 5.7 The mean (±SE) numbers of mistakes, grouped by four trials (1 – 4; 5 – 8; 9 – 12; 13

– 16; 17 – 20) that the tested bees made when navigating through the maze………………….118

Figure 5.8 The percent (±SE) of bees’ choices of the correct (right-hand) first turn on entering the maze…………………………………………………………………………………..……119

Figure 5.9 The mean (±SE) durations (seconds) bumblebees spent to navigate the maze successfully and mean numbers of mistakes………………………………………………..…120

Figure 6.1 Experiment setup with hive, holding area, flight cage testing arena, patch of artificial flowers and mesh tube routes with gates by which the bees were allowed to enter and exit the flight cage……………………………………………………………………………………….131

Figure 6.2 Artificial flowers were made of 1.5mL centrifuge tubes inserted into the centres of blue plastic discs, 7 cm in diameter ………………………………………………………..…132

Figure 6.3 The learning curve (time/ sec) ( +/- SE) taken to access and forage on syrup) for 10 initially naive workers of Bombus impatiens allowed to forage freely, but only one at a time, at artificial flowers with and without dead bees present…………………………………………141

Figure 6.4 The learning curve (time/ sec) ( +/- SE) taken to access and forage on syrup) for 9 workers of Bombus impatiens allowed to watch experienced foragers at artificial flowers for 10 hours and held incommunicado overnight…………………………………………………..…142

Figure 6.5 The learning curve (time/ sec) ( +/- SE) taken to access and forage on syrup) for 9 workers of Bombus impatiens allowed to watch experienced foragers ( from different colony

(colony A)) at artificial flowers for 10 hours and held incommunicado overnight……………143

Figure 6.6 The learning curve (time/ sec) (+/- SE) taken to access and forage on syrup) for 9 workers of Bombus impatiens allowed to contact with their nest-mates B1 (i.e. were watching the experienced bees from colony A)………………………………………………………………144

xvi

CHAPTER 1

INTRODUCTION

1.1 Introduction

The environment consists of different elements that could affect the behaviour of all surrounding organisms, including animals. If the animal lives in a stable environment, it needs to rely on the information that was gathered at the beginning of its life, while exhibiting an innate behavioural response to environmental cues. However, in an environment that is rapidly changing, the animal must gather and obtain new information continuously in order to survive.

Therefore, the processes of learning and cognitive abilities are vital components of an animal’s life.

Learning is the faculty of acquiring new (or modifying and reinforcing existing) knowledge and changes in behaviour. It also refers to the gaining of new skills, the elevating of surrounding values, and the creation of strategies or preferences, and may involve combining different types of information. Moreover, learning does not necessarily appear suddenly, but it is constructed and built upon training and is shaped by the environment and existing knowledge.

Learning also can be viewed as the process of applying the knowledge, rather than just a collection of factual and procedural knowledge. Learning thus generates relatively permanent changes in the organism (Schacter et al 2011) and the progress in learning over time has a tendency to follow learning curves.

Although the ability to learn is possessed by humans, animals and some machines, learning studies often concentrate on higher species such as humans, other primates and cetaceans. However, studying learning and its relevant aspect to all animals’ lives could contribute to a better understanding of the learning process, which involves vital functions for

1 the surviving animals, including insects and, specifically, the bumblebee (Bombus impatiens) in this study. Learning cannot be observed directly as it is different from other biological mechanisms and has to be analyzed through comparative studies and from behavioural responses to specific settings of natural or experimental conditions.

In reviewing the literature, there is a growing body of publications, both in quantity and theoretical coverage, focused on animal cognition and learning, and many of these have combined approaches from the regulations of psychology and behavioural ecology (e.g., Marler

& Terrace 1984; Johnston 1982; Heinrich1984; Johnston 1985; Roitblat 1987; Real 1994; Dukas

1998; Chittka & Thomson 2001; Papaj & Lewis 1993; Friederici & Menzel 1999; Shettleworth

2010; West-Eberhard 2003; Bouton 2007). Moreover, the theory of learning is a science developing to produce better understandings of how the learning process works. This is established primarily from the fundamental principles of many different branches of scientific inquiry.

Animals interact and respond to their environment using everything from simple, innate behaviour to the more non-associative and associative cognitively complex processes of learning and memory. This chapter provides definitions and short reviews of innate and non-associative learning. Additionally, it focuses on all other types of learning, including the types that were investigated in this dissertation. Each chapter, which refers to particular sets of experiments, includes a more specific introduction, thus this chapter is brief general overview.

1.1.1 Learning definitions

Learning in general can be defined as a change in behaviour that is a result of experience.

Psychologists have defined learning as "A relatively permanent change in behaviour due to

2 experience.” The meaning is the same as a person saying "I learned to do" as part of everyday life. (Domjan 2003; Raygor 2005).

1.1.2 Definitions of learning found through a review of the literature

“Behavioural changes as a result of individual’s experience.” Thorpe 1956.

" any process in which, during normal, species-typical ontogeny, the organization of an animal's behaviour is in part determined by some specific prior experience." Johnston 1982, p.70

"Learning, on the other hand, was a process whereby "nurture" (experience) could supplement the provisions of instinct, permitting an animal to deal with unusual or unexpected circumstances by acquiring the necessary behavioural skills." Johnston 1982, p.67

"From an ecological perspective, learning may be defined as the modification or maintenance of the behavioural relationships between an animal and its environment as a result of individual experience." Johnston 1985, p.6

"Learning, the acquisition of new informational relationships as a result of functional experience, is probably a capability of most neuronal networks, whether they are located in the brain, peripheral nervous system or tissue culture." Pickard 1991, p.253

“A change in behaviour due to experience.” Kamil 1994; Johnson 1991

“Developmental growth occurring within-lifetime that is the result of interaction between the individual and its environment.” Belew & Mitchell 1996

"Learning is a type of phenotypic plasticity." West-Eberhard 1989; Stephens 1991; Moran

1992; Scheiner 1993; Dukas 1998

"Learning may be defined as the acquisition of or change in memory that allows a subject to alter its subsequent responses to certain stimuli." Dukas 1998, p.133

"Learning how refers to the acquisition of behavioural traits involved in search, pursuit, and handling of the prey." Giraldeau & Caraco 2000, p.254

"Learning about is concerned with the collection of information pertinent to the estimation and assessment of alternatives." Giraldeau & Caraco 2000, p.253

" A long lasting change in behaviour that results from experience" Pearce 2008

"Learning is a cognitive ability defined as the acquisition of neuronal representations of new information." Dukas 2009

1.1.3 The benefits and the cost of learning

Through learning, animals have the opportunity to acquire information and are able to adapt to their environment at a local scale, these adaptations effect on their evolutionary history.

(Wcislo 1989; Dukas 1998; West-Eberhard 2003). The benefits of learning must have been enormous for evolution to have overcome those costs, (Kawecki 2010) and for many animals, learning mainly offers a benefit in finding food or a mate. Basically, if the benefit of obtaining certain information exceeded the cost of having it, then the animal would learn and use it.

However, if the cost of acquiring the knowledge exceeded the benefit, then the individual would exhibit non-learning behaviour ( Mery and Kawecki 2003).

Animals explore their environment for nesting (home) and food, and the learning function is related to the detection, acquiring, and tracking of changes in resources in changeable environments. Thus, the evolutionary benefit of learning will have an effect on an animal’s fitness. Moreover, learning generates variation in the type of food used by individuals between species and within the same species (West-Eberhard 2003). Specialization is known in bumble bees Bombus Spp (Heinrich 1976a and b, 1979a), honeybees Apis mellifera (Wells & Wells

1986), wasps (Sphex ichneumoneus (Sphecidae)) (Brockmann 1985), and birds, song birds

(Tinbergen 1960), mixed-species flocks of tropical seed-eating finches (Giraldeau 1984).

The mechanisms of acquiring and managing the information are increasingly complicated. These are referred to as the cost of cognition and they limit the abilities of learning of the organism. One such cost in the process involves maintaining the consistency of the information stored (Johnston 1982; Dukas 1999). Relying on information that’s exclusive to a time and place is beneficial, for instance, some bee species feed on a single flower species.

However, other bees are adapted to many different flowers, each with a different shape and a different flowering time, learning may be a better strategy in such cases. Darwin hypothesised that flower handling methods learned on one plant species interfere with previously learned handling methods of other plant species. This interference increase their handling time and handling errors; testing bumble bees (Bombus fervidus) to measure the constancy; However, when the bees foraging on simple flowers, they switched between species without increasing of handling times or handling errors, but they exhibited strong constancy when visiting more complex flowers with increasing of handling times and error when switching between flowers

(Laverty 1994, Dukas 1999). Another cost of cognition that must be considered is the rapid loss of value in recently acquired information when the ratio of environmental change/learning rate departs from an optimal value (Stephens 1989; Kerr & Feldman 2003). Thus, as the use of memory become more perfect and complete the cost of cognition become higher. Forming neuron connections may cause harmful side effects; it is also possible that genes that allow learning to develop faster and last longer may cause other changes. These could cost the nervous tissue (Atwell & Laughlin 2001), and are demonstrated by the plasticity shown by vertebrates in the size of the hippocampus in their brains, which grows larger when spatial memory is needed,

5 but reduced in size when it is not (Clayton & Krebs 1994; Jacobs 1996). There is growing evidence indicated that the growth of mushroom bodies (MBs) of the insect brain affected by olfactory learning (Menzel et al. 1974; Erber et al. 1980; Heisenberg et al. 1985; Davis

1993; Debelle and Heisenberg 1994; Meller and Davis 1996), and tasks requiring visual learning

(Mizunami et al. 1993). These have been exhibited volume plasticity in adult honeybees’ workers, queens, and males (Withers et al. 1993; Durst et al. 1994; Fahrbach et al. 1995, 1997), carpenter ant workers (Gronenberg et al. 1996), and fruit flies (Heisenberg et al. 1995; Barth et al. 1997). Kawecki (2010) indicated that learning had harmful side effects, on study of fruit flies, learning affect their cycle life by 15 percent less living by the fast-learning flies.

1.2 Innate behaviour

With a steady environment, information that has gathered regarding the unchangeable in the surroundings will generate innate behaviour throughout the animal’s life. One of the innate learning behaviours that appear in the animal is the fixed action pattern, unchangeable sequences of behaviour that should be performed to completion once started. The second innate learning behaviour is imprinting, the term used in both psychology and ethology, and demonstrated by certain species. It is innate behaviour that occurs during a critical time at an early stage in life and describes any kind of phase-sensitive learning (Shettleworth 2010). It was first used to describe situations in which an animal or person learns the characteristics of some stimulus.

1.3 Learning

1.3.1 Non Associative learning

Although non-associative learning is not one of the learning processes that were explored in this study, it should be mentioned for clarity. It refers to a permanent change in the strength of response to a single stimulus due to repeated exposure to that stimulus (Pearce 2008;

Shettleworth 2010). An example of non-associative learning, habituation is the reduction of responsiveness during continual exposure to unimportant stimuli. The learning underlying habituation is a fundamental or basic process of biological systems and does not require conscious motivation or awareness to occur. Indeed, without habituation I would be unable to distinguish meaningful information from the background, unchanging information. Animals exhibit some response to a stimulus, but as the stimulus continues to be presented and is unrewarded or harmful, these animals reduce subsequent responses. Habituation allows the nervous system to optimize sensory-motor processing by eliminating unnecessary responses. It allows us to adapt to the familiar in order to preserve my ability to react rapidly and appropriately to the new (Wood 1988; Pearce 2008; Shettleworth 2010). Another example of non-associative learning is sensitisation; the animal’s responsiveness increases following the repeated exposure of a stimulus (Bell et al., 1995). Sensitisation is considered to be beneath both adaptation and maladaptation processes in the animal.

1.3.2 Associative learning

Chapter Two includes some associative conditioning (i.e. operant conditionings), discriminative stimuli and extinction. Associative learning is learning or conditioning in which there is an association between two different events, stimuli (anything that comes in through your senses) or a behaviour and a stimulus that occurs or happens together; therefore, the association must occur for the learning producer (Drickamer at el. 2002; Raygor 2005). For instance, could a dog salivate at the sound of a bell if it never makes the link between the bell and the reward of food? Or why would a rat learn to press a lever if it never makes the connection between pressing the lever and getting a reward? Honeybees (A. mellifera) have been

7 shown associative learning by the proboscis extension reaction (Takeda 1961; Bitterman et al.

1983).

Associative learning includes two forms, classical and operant conditioning. Classical conditioning (also known as Pavlovian conditioning or respondent conditioning) (Pavlov 1927) is the kind of learning that occurs when a conditioned stimulus (CS) (e.g. sound or colour) that does not normally evoke the conditioned response (CR) is paired with an unconditioned stimulus

(US), (e.g. a reward or food), eliciting a reflexive unconditioned response (UR). After several pairings of both stimuli, the animal displays a conditioned response (CR) to the CS even in the absence of US. Although the CR is similar to the UR, it must be gained through experience

(Drickamer at el. 2002; Raygor 2005; Bouton 2007).

The paradigm of the proboscis extension by honeybees (A. mellifera) provides a good example of classical conditioning (Bitterman et al. 1983). In conditioning restrained honeybees

(A. mellifera) with classical conditioning of the proboscis extension response (Bitterman et al.

1983). Niggebrügge and colleagues (2009) found that faster colour discrimination learning was correlated with reduced colour similarity between stimuli.

The basic principle of operant conditioning is that some behaviour can be modified by their antecedents and consequences, and that the behaviour may change in form, frequency, or strength (Thorndike 1901; Skinner 1938). Behaviours that have positive consequences tend to occur more often. Those with negative consequences tend to occur less often (Drickamer at el. 2002; Raygor 2005).

Operant conditioning is distinguished from classical conditioning (or respondent conditioning). In classical conditioning, a previously neutral stimulus is repeatedly presented together with a reflex eliciting stimuli until eventually the neutral stimulus elicits a response on

8 its own. In operant conditioning, certain behaviour is either reinforced or punished, resulting in an altered probability that the behaviour can happen again (Raygor 2005). Operant conditioning deals with reinforcement and punishment to change behaviour. Although, operant behaviours develop and change because of what happens after they occur, they depend on their consequences. In contrast, classical conditioning involves putting together two stimuli before the behaviour occurs. The behaviours are learned and the conditioning procedure is different

(Domjan 2003; Raygor 2005).

Discrimination stimuli provide conditioning that allows for testing the discrimination abilities of animals and babies to differentiate between similar stimuli. If the animal is able to discriminate between the two stimuli, and responds differently to one stimulus (with the US eliciting different responses either to another CR or no CR at all), the response should grow stronger. At the same time, its response to the other stimulus should go through extinction because I am presenting the CS without the US (Bernstein and Nash 2008).

Chapter Two describes studies in which the bees were challenged to discriminate between two coloured stimuli by presenting two different artificial flowers. The upper disc of the artificial flower was rotatable, and both stimuli were rewarded, with two rotation direction (45o counter clockwise or clockwise). The test was designed to examine the bees’ ability to associate the colour of the upper disc with the rotation direction and whether or not extinction (i.e. The response or the behaviour gradually stops (Miltenberger 2012)) occurs when the artificial flowers

(CS) were presented to the bees unrewarded (US is absence) after a few trials. After showing extinction, animals eventually return to pre-training levels and spontaneous recovery occurs, though the prior conditioning was not completely eliminated by the extinction procedure

(Drickamer at el. 2002; Bouton 2007; Bernstein and Nash 2008).

1.3.3 Cognitive learning

Cognition is the ability of an animal to be aware of and make decisions about its environment. Cognitive learning usually refers to all learning that does not occur through traditional classical or operant conditioning. What is learned is usually more complex than a simple response to a stimulus.

1.3.3.1 Insight learning

First developed by Wolfgang Kohler (1925), insight learning is in a sense the “highest form” of learning observed. It is the ability to problem solve or perform a correct or appropriate behaviour the first time the animal is exposed to a situation. For example: a chimpanzee (Pan troglodytes) may stack boxes to obtain a food object hung out of its reach without ever having seen this solution to the problem before. There is no gradual shaping or trial and error involved, but rather internal organizational processes cause new behaviour. Moreover, insight learning does not directly involve using past experiences to solve a problem and is not restricted to primates but other animals could display insight learning (e.g. ravens and other birds).

(Shettleworth 2012)

1.3.3.2 Problem Solving

Problem solving is the process that involves discovering, analyzing and solving problems. It is the ultimate goal to come up with a solution, move obstacles and solve the problem (Mayer 1992). Solving problems requires a series of steps to be followed, which include recognizing the problem, developing strategies, and finally, organizing and applying these strategies (Schooler et al 1993). In the third chapter, studies are described that used trials and errors approach as a technique to problem solving by applying the concept of scaffolding and shaping.

Scaffold learning and cognition are the supporting information given throughout the learning procedure, modified to the requirements of the learners, for the purpose of helping them to achieve learning goals (Sawyer 2006). The notion of scaffold learning is mostly used in human psychology to refer to the increasing sophistication and complexity of a set of tasks. The task is simple when first introduced, but learners are then taught through increasing sophistication such subjects as reading, writing, mathematics, music, use of tools, and social interaction. Scaffold learning assumes that individuals may spend longer times or give up if, through inexperience, they are given tasks that seem too complicated but can be solved when presented through a gradually increasing set of difficulties. Thus, my studies advocate that the paradigm to demonstrate scaffold learning include tests in which the naïve subject animals give up, or, at least, show great difficulty, when presented with complex tasks in naiveté.

Vygotsky (1987) proposed that scaffolding is important for learners to obtain the skills to reach goals that would have been difficult or impossible to attain in naiveté. The Vygotskian theoretical perspective claims that learners can reach a high level of performance (development)

(Zone of Proximal Development ZPD) when they are scaffolded and supported. Vygotsky proposed that there are two characteristics of learning development: the “actual development as determined by independent problem solving” and the “potential development as determined through problem solving under (adult) guidance” (Vygotsky 1987). Furthermore, the ZPD is the area between independent achievement and achievement gained by support and guidance

(scaffolding) from more knowledgeable individuals. Therefore, learners’ ZPDs can be enhanced by the scaffolds provided in their surrounding environment.

One aspect of scaffold learning is the step-wise nature of support for problem-based learning (PBL). In PBL, scaffolding is gradually added, then modified, and finally removed as

11 the learners gain more knowledge to solve problems, develop skills, and determine and assess solutions (Hoffman and Ritchie 1997). The learners finally invent specific techniques to solve their problems and to use those in additional situations.

The instructional paradigm of scaffolding is intended to promote deeper learning than would be possible otherwise. It also provides sufficient information to support and encourage learners when they encounter new concepts or tasks and then develop skills that allow them to achieve their goals. These instructional facets of scaffold learning may include resources for information, compelling tasks, modeling and guides, guidance on the development of cognitive and social skills, and coaching. The eventual goal is that instructional scaffolding is no longer needed (practice makes perfect). With scaffolding (support), teachers choose the appropriate time to apply support, choose the right method and tools to follow, and decide when the support scaffold can be removed (Lajoie 2005).

Shaping is defined as reinforcing successive approximations to the desired behaviour. In considering shaping to be effective, two important practical facts should be remembered. First, the desired behaviour will be reinforcing relatively poor performance. Second, to shape any behaviour, gradual steps should be conducted, breaking the tasks into small steps, starting with the easiest task and finishing with the most complex. The shaping cannot be successful if the tasks are too complicated.

In Skinnerian shaping (Skinner 1938), animals are trained to perform tasks of increasing complexity, from rewarding positive responses to simple tasks in naiveté to more and more complex responses when exposed to more and more complex tasks. Thus, the formation of an existing response is gradually augmented through successive trials towards complex target behaviour and rewarding segments of that behaviour. As Skinner (1953) noted “the original

12 probability of the response in its final form is very low; in some cases it may even be zero.”

Thus, via shaping, animals can learn to perform the most difficult tasks that would never otherwise appear as part of their natural behaviour.

1.3.3.3 Conditional discrimination and response chaining

Conditional discrimination is shown when an animal responds differently to the same stimulus presented in differing contexts (Mostofsky 1965). In conditional problems, animals discriminate between different conditioning with the same background or between a condition with different background, (e.g. dicriminate between colors on the basis of a common odor or between odors on the basis of a common color). Compound–component discrimination has been known since the early work of Pavlov (1927), however, later in (1938) Lashley called it

“Conditional discrimination”, which the compound component were made identically in different combination. Lashley (1938) conditioned rats to differentiate between upright and inverted triangles targets on black or striated backgrounds, his results were proved by North and his colleagues (1958). Although there is an extensive literature on conditional discrimination learning by vertebrates using compounds of visual and auditory such as in rabbits (Saavedra,

1975), goldfish (Bitterman, 1984) and pigeons (Carter and Werner1972; Schrier and Thompson

1980; Thomas et al. 1988), also utilizing specific combinations of color and form as stimuli for pigeons (Born et al 1969). However, few studied has been published to demonstrate conditional discrimination by invertebrate. Specifically by honeybees (Apis mellijera) (Couvillon and

Bitterman 1988, 1989, 1991; Funayama et al. 1995), they trained honeybees to two different conditional discrimination problems, which required the bees to discriminate between two combination of components, two coloured objects scented with the same on the odour, or that

13 required them to discriminate between two differently scented targets on the basis of same coloured objects.

As early as 1899, James noted that the appreciation of sequential events must be important in the behaviours of animals, but as Weisman et al. (1980) discuss in their introductory review, little attention was paid to that aspect of learning and cognition. Response chaining is shown when an animal responds to a fixed sequence of objects or events in order to gain reward

(Pearce 2008). Response chain learning, also called serial recognition (Pearce 2008) and linking

(Taylor et al 2010), involves the subject acquiring skills to perform a series of tasks in order, so that one correct response provides the cue for next and it is the last correct response that produces a reinforcer (Skinner 2005). Chained responses may produce, or alter, some of the variables which control other responses (Skinner 2005) (as in the studies of Balleine et al. (1995) with rats that pressed a lever and then pulled a chain to obtain the reinforcer), but that situation may have been complicated by “chunking” (Terrance 1987, 1991) by which combinations of stimuli presented simultaneously (simultaneous chaining) make for improved cognition vs. single stimuli alone. Although previous research indicates conditional discrimination occurs in invertebrate behaviour (see below).

Sequence learning is shown by a subject’s having acquired, through experience, the ability to recognize different objects or events as they are encountered one after another in a particular order (Dehaene 1999). Sun and Giles (2001) suggest sequence learning could be the most important and prevalent kind of cognitive learning. The analysis of sequential behaviours plays an essential role in the study of classical and instrumental conditioning, and of problem solving and reasoning.

There are four interconnected problems involved in sequence learning. The subject must react by experience to: 1) sequence prediction, 2) sequence generation, 3) sequence recognition, and 4) sequential decision making (Sun 2001; Sun and Giles 2001). Sequence learning might be dependent upon an animal’s ability to organize learned behaviours hierarchically into behavioural chains with goals and sub-goals (Byrne & Byrne 1993; Byrne & Russon 1998). I consider that an animal’s tackling of those four interconnected problems requires that it has some understanding of where it is in the sequence; i.e. has some means of counting its accomplishments as it progresses.

Sequential decision making involves all of the above and can be thought of as having three non-mutually exclusive variations: a) goal-oriented, b) trajectory-oriented, and c) reinforcement-maximizing. These “problems” and “variations” show how sequences are formulated, the patterns of sequences are followed, and how different phases of sequence learning are related to each other (Shettleworth 2010; Sun 2001; Pearce 2008).

Chapter Four, experiments and their results, tested the ability of bees to conditionally discriminate patterns or count elements in a serial (conditional discrimination) or fixed order

(response chaining). The ability to count has been proven in some animals (Shettleworth 1998;

Pearce 2008) and conditioned as a way of measuring an animal’s cognitive capacity (Hauser

2000). Counting is the estimation of numbers of elements, events or tasks and can be accomplished through subitizing (Kauffman et al. 1949; Davis and Pérusse 1988), pattern recognition (as figure numeral recognition by honeybees [A. mellifera]) (Leppik 1953; Chen et al. 2003), and true counting, whereby the subject arrives at the number of objects it encounters. It has been observed in some vertebrates (Shettleworth 1998; Pearce 2008) (see Chapter 4), but is not well documented in invertebrates. To date, few studies imply numerical, true counting in

15 insects (Pahl et al. 2013) and mostly focused on social insects because they, as central place foragers, face navigational problems and may particularly benefit from a sense of number (Pahl et al. 2013). These insects count landmarks (Chittka and Geiger 1995; Dacke and Srinivasan,

2008), as well as possible step-counting in ants (Reznikova and Ryabko 1996, 2011; Wittlinger et al. 2006), and bees (Seeley 1977).

Response chain is the learning of two or more events or tasks presented in a fixed sequence. These events or tasks range from simple navigation to complex and manipulative tasks. The complexity of chain learning includes problem solving in which tasks may include barriers and obstacles that must be solved in order to complete the task, as shown in Kohler

(1925) pigeons (Columba livia), New Caledonian crows (Corvus moneduloides) (Taylor et al.

2010), chimpanzees (P. troglodytes) (Dohl 1968), and problem solving in bumblebees (Bombus impatiens) (Mirwan and Kevan 2014). Bumblebees have been shown to forage from mixed arrays of three flower types differing in complexity presented simultaneously (Gegear and

Laverty 1995, 1998). However, response chaining, coupled with problem solving, has not been tested in invertebrates until now, as far as I am aware.

1.3.3.4 Navigation and maze navigation

Navigation is the ability of many animals to orient themselves in their environment precisely without maps or instruction. Thus the abilities of animals to acquire, store, and recall information is necessary as they navigate. Many animals navigate effectively over short distances, but migrating animals navigate for thousands of kilometers.

Most animals are guided by olfactory and visual landmarks when navigating paths and returning to their home. However, in 1873, Charles Darwin suggested dead reckoning, also called path integration (Mittelstaedt and Mittelstaedt 1980), as a possible mechanism of

16 navigating by which the animal uses only the information of its speed and direction without the utilization of visual or other external landmarks.

Many means have been suggested for animal navigation, including navigation cues ranging from the sun, the polarization pattern of the blue sky, the night sky, the earth's magnetic field, and olfaction and vision; some of these have been proven (Mittelstaedt and Mittelstaedt

1980; Pearce 2008; Shettleworth 2010). Animals can use different cues to navigate efficiently, forage successfully and return home, and find nesting sites or mates. For instance, birds and insects integrate cues of landmarks with direction, though some animals can navigate in the dark without using any visual cues. Frisch (1967) found that honeybees (A. mellifera) as a central- place forager rely on the sun to navigate, depending on the polarization pattern of the blue sky and the earth's magnetic field.

Animals usually form an internal map using visual cues, but other senses including olfaction and echolocation could also be utilized. This kind of internal representation of the real world is called a cognitive map (i.e. a mental image of the environment restored and retrieved as necessary to guide an animal’s movements). The faster the animal navigates, the sooner it receives the reinforcers that were waiting, improving its cognitive map (Pearce 2008; Collett and

Graham 2004). Gould's (1986) experiments (in agreement with a number of earlier studies on other Hymenoptera e.g., Fabre 1915; Thorpe 1950; Chmurzynski 1964) indicate that bees possess more than route maps; his conclusion from these experiments is that bees may have

"cognitive maps" of their terrain (O’Keefe and Nadel 1978; Cartwright and Collett 1987), however Dyer (1991) concluded that bees used visual cues to locate the food source rather than using a cognitive map.

Using mazes to study the mechanism of navigation focused on vertebrates (Dale 1988;

Anderson 2000; Fu and Anderson 2006), Tolman & Honzik (1930) designed an experiment to examine the latent learning (i.e. learning that takes place before reinforcers are introduced) in rats. They concluded that the rats had developed a cognitive map of the maze, even when the reward was not available. They again used their cognitive map when food became introduced, implying that accurate cognitive maps have survival value.

Most invertebrate studies have used simple, single bifurcation choice chambers (T- or Y- mazes) as multiple-turn mazes (Honzik 1936) have rarely been used to examine the navigational capacities of invertebrates. The learning of complex mazes has been demonstrated in ants and bees, mostly coupled with sensory (visual, olfactory, textural) cues provided at choice points

(Schneirla 1929; Weiss 1953; Collett et al. 1993; Collett & Baron, 1995; Zhang et al. 1996;

Zhang et al. 1998, 1999; Chameron et al. 1988). Zhang et al. (1996, 2000) studied flying honeybees’ (A. mellifera) ability to navigate complex labyrinths containing many ‘‘dead ends’’ and four other types of mazes, which involved making a correct choice at each turn, coupled with visual cues or sequence turns to achieve the reward. Chapter Five focuses on the ability of bumblebees to navigate multi-turn mazes and re-navigate them after several days of not using them.

1.3.3.5 Social learning

Social learning is the concept of learning from others, mostly involving observation and modeling, and imitation and communication to transmit the learned behaviour. It requires attention, memory, and motivation (Whiten and Ham 1992; Shettleworth 1998). Although research on social learning has focused largely on vertebrates and within species (Lefebvre 1995;

Heyes and Galef 1996), researchers have recently shown that bees and other invertebrates can

18 learn through acquisition of information by social transmission from conspecifics. Moreover, social learning could occur between species (Leadbeater and Chittka 2008; Baude 2008).

Information transmission between two individuals can pass in four channels: acoustic, visual, chemical, and electrical. An individual animal may require information from two or more channels simultaneously before responding appropriately to the reception of a signal.

Furthermore, a stimulus may elicit a response under one circumstance, but be ignored in a different context.

In observational learning or modeling, the animal learns by watching other animals perform a task or solve a problem and replicates the administrator’s behaviour. It should copy three types of information all together: the demonstrator's goals, actions, and environmental outcomes. This manner of learning through observation can be displayed by a wide range of animal species. Such learning may occur in nature via the use and creation of tools by animals for the acquisition of food and water, grooming, defence, etc. (Hansell 2005, Pearce2008).

Through observation, the stripping technique used by black rats in the pine forest to extract the seeds from pine-cones is socially transmitted from mothers to their offspring (Zohar and Terkel

1996, Shetlleworth 2010).

Bumblebees can learn directly, through observing other conspecific foragers (within species), about the various rewarding flower species (Worden and Papaj 2005), but they are attracted to occupied flowers (Leadbeater and Chittka 2007). Moreover, they switch more quickly to other rewarding flower species if they observe bees already foraging there. The presence of other individuals (conspecifics) decreased the time spent discovering food resources; therefore, the utilization of social learning would also be advantageous when the flower species availability changes responsiveness due to seasoning sequences (Kawaguchi et al. 2006).

Nectar robbing “theft” by bumblebees (usually short-tongued) also spreads through social transmission. It occurs when individuals watching other nectar robbers either bite and make a hole at the base of a flower (primary robbing) or use holes that others robbers have already made

(secondary robbing) (Leadbeater and Chittka 2008).

Animal communication is a way of information transmission through such means as sound, vision, taste or smell, tactile or electrical impulses, or a combination of these.

Communication is generated by one member of a species to affect the behaviour of another in the same species (intraspecific communication) and this can be found in almost all animals

(Bradbury and Vehrencamp 2011). Honeybees Apis spp and stingless bees, Meliponini,

(Hymenoptera, Apidae) have evolved sophisticated communication systems (Frisch 1967; Nieh

2004), with honeybees possessing very advanced patterns of communication (i.e. the dance language decoded by Karl von Frisch in 1967). This serves as the means of information transmission for specific coordinates to abundant sources for potential foragers. Moreover, the food’s scent plays a major role in delivering the information. Some stingless bee species exhibit a form of dance which somehow resembles the honey bee waggle dance (Chittka 2009), but the communication of stingless bees is diverse and varies depending on the species differences. It may also range from odour trails to the potential encoding of food location (Kerr 1969; Roubik

1992; Dyer 2002; Nieh 2004). By contrast, bumblebees do not display dance communication, but rather a random movement by the returning bees. Successful foragers ran around the nest faster for a longer time, emitting a pheromone and delivering food to stimulate their nest-mates

(Dornhaus and Chittka 1999, 2001). It has been hypothesized, that the buzzing sounds made by bees’ wings provides a means of communication for bumblebees (Dornhaus and Chittka 2001).

However, there has not been sufficient study to test this hypothesis.

Additionally, both honey bees and bumblebees (Bombus spp) can avoid revisiting flowers that were previously visited and depleted by conspecifics (Goulson et al. 1998). Bees are able to detect recently exhausted flowers through scent marks of the oily residue that was left behind

(Saleh et al. 2007). For example, honey bees Apis mellifera use these short-lived scent marks as a repellent and this saves time that would otherwise be spent in empty flowers (Giurfa and Nfifiez

1992). It has been found that bumblebee species also exhibit this behaviour of avoidance. That said, bumblebee species Bombus terrestris and Bombus pascuorum have the ability to distinguish and reject flowers that not only were recently visited by conspecifics but also by other Bombus species (Goulson et al 1998). Chapter Six presents studies on the notion of social learning and the ability of bumblebees to transfer information through observation, modeling and imitation, revealing information transmission through communication.

1.3.4 Objectives

1. To determine bumblebees’ ability for Operant learning

2. To assess bumblebees’ capability for Problem solving through shaping and scaffold

learning

3. To test bumblebees’ ability for Conditional discrimination & manipulation of series

of obstacles with Response chains

4. To examine bumblebees’ capability for Maze navigation

5. To determine bumblebee’ ability for Social learning through observation and

communication

1.3.5 Overarching Hypothesis

Bumblebees have well-developed capabilities for learning complex tasks (e.g. manipulating objects, solving problems under various conditions and maze-navigation, some can involve social learning).

1.4 Glossary

This glossary is presented in Alphabetical (Raygor 2005).

Classical conditioning: Learning that results from the pairing of an unconditioned and a conditioned stimulus.

Cognitive map: A mental image of an animal’s environment that the animal can use to guide its movements.

Conditioned response (CR): A response elicited by a conditioned stimulus.

Conditioned stimulus (CS): A stimulus that elicits a response after being paired with an unconditioned stimulus.

Discrimination: The tendency to respond differently to two or more stimuli.

Discrimination training: Training an animal to respond differently to two different stimuli.

Discriminative stimulus: A stimulus that serves as a signal that a response will be followed by a reinforcer.

Extinction: Eliminating a previously conditioned response, the disappearance of a previously learned behavior when the behavior is not reinforced.

Habituation: A decrease in response to a repeated stimulus.

Latent learning: Learning that occurs in the absence of obvious reinforcers and only appears after reinforcement is introduced.

Law of effect: Thorndike’s (1901) term for the principle that responses that are followed by desirable consequences occur more often and responses that are not followed by desirable consequences occur less often.

Learning: A relatively permanent change in behaviour due to experience.

Modeling: The tendency to imitate the behaviour of another person.

Negative reinforcement: Increasing the frequency of a response by removing an aversive stimulus after the response occurs.

Operant extinction: When behaviour no longer produces predictable consequences, its return to the level of occurrence it had before operant conditioning.

Operant conditioning: A process of conditioning in which behaviour change is caused by the consequences of the behaviour.

Positive reinforcement Increasing the frequency of a response by presenting a reinforcing stimulus after the response occurs.

Punishment: A stimulus that decreases the frequency of the behaviour it follows.

Reinforcer: A stimulus that increases the frequency of the behaviour it follows.

Response: Any activity of the muscles or other identifiable behaviour.

Serial processes: Two or more mental processes that are carried out in order, one after the other.

Shaping: Reinforcing successive approximations to the desired behaviour.

Social-learning theory: The learning theory that stresses the role of observation and the imitation of behaviours observed in others.

Stimulus: Information perceived through the senses.

Stimulus control: A response is said to be under stimulus control when a particular stimulus controls the occurrence or form of the response.

Unconditioned stimulus (US): A stimulus that elicits a response before any conditioning has occurred.

Unconditioned response (UR): A response elicited by an unconditioned stimulus.

Chapter 2

Complex Operant Learning by Worker Bumblebees (Bombus impatiens Cresson (Hymenoptera: Apidae))

2.1 ABSTRACT

Animals react to salient stimuli via unconditioned responses, Pavlovian conditioning, conditioned manipulation of objects in simple and complex ways (operant learning), insight learning, and tool use. Bumblebees are known to learn to associate rewards (e.g. sugar syrup) with many stimuli. They will also manipulate natural and artificial complex flowers to obtain rewards. Even though those tasks involve operant learning, all have been rather natural

(resembling activities that bees would perform living free in the wild), and none have involved detour behaviour; thus none have required subject bees to approach a completely hidden, covered reward nor to move objects through multiple body lengths. I trained bumblebees with two types of operant behaviours that I believed represented challenges unlike any they have evolved to respond to. These were to drag variously sized caps aside, and to push and rotate discs through varying arcs (and by up to three body lengths) from the entrances of artificial flowers, in order to access a reward hidden beneath. Further, I successfully trained bees to use disc colour as a discriminative stimulus predicting whether rotating discs clockwise or counter clockwise would reveal the reward. The operant learning demonstrated for bumblebees indicates a level of hitherto unexpected ability, in that the subject bees learned novel behavioural sequences, manipulating and moving items in ways that seem far from any natural task that they would encounter. I suggest that this complex operant learning capacity demonstrated in bumblebees reflects a vertebrate-like degree of behavioural plasticity.

2.2 INTRODUCTION

Associative learning can be placed into two major categories: a) classical or Pavlovian, and b) operant or instrumental. The former involves learned associations between formerly neutral cues and unconditional stimuli from reinforcers (Stimulus-Stimulus learning), but the latter involves learning operations or manipulations that yield access to reinforcers (Stimulus-

Response learning). Operant learning may involve tasks such as passing a barrier or pushing a lever (Skinner 1938; Staddon and Cerutti 2003), or more complex ones such in as tool use

(D'Amato and Colombo 1988; D'Amato 1991; Shumaker et al. 2011; Sanz et al. 2013 ). For bees passing barriers, the seminal research of Laverty (Laverty 1980; Laverty and Plowright 1988;

Gegear and Laverty 1995), showed that they learned to manipulate flowers to obtain enclosed nectar, pushing past barriers in complex flowers as they move directly towards the reward. I investigated a form of operant learning that requires the bees to deviate from a direct path to the goal, whereby it performs a series of tasks that expose the reinforcer but that take the subject away from the reinforcer, so requiring it to return to the access the reward. Various mammals, birds and Crustacea have been shown to deviate from a direct path to obtain a reinforcer by complex strategies such as by pressing levers or operating switches that then expose reinforcers, mostly food, in operant conditioning chambers (Skinner 1938). However, in performing those tasks, the subjects did not go first to the precise site of access to the reinforcer where they had to move a barrier several body lengths to the side and then, secondarily, return to the access to obtain the reward. These experiments tested the capacity of worker bumblebees to solve the problems of manipulating and moving barriers coupled with their having to travel to (by flight), away from (by walking), and back (by walking) to the site of the reinforcer.

Although studies have been made on operant learning in invertebrates (Rubadeau and

Conrad 1963; Brembs 2003; Tomina, and Takahata 2010), including insects as diverse as flies

(Diptera) (Drosophila (Wustmann et al. 1996; Gong et al 1998; Brembs 2000), Phormia

(Sokolowski and Abramson 2010) and bees (Hymenoptera) (Kisch and Erber 1999; Mirwan and

Kevan 2014), most have not challenged their subjects to manipulate objects. An exception is the work of Pessotti (1972) who was able to train an individual forager of Melipona rufiventris

Lepeletier (Meliponini: Apidae) to discriminate between coloured stimuli (lights) and, by activating a lever at the training light, obtain the reinforcer. As far as I am aware, insects have not been demonstrated to be capable of operant learning involving deviation, once at the site of the reinforcer, from the direct path to the reinforcer. I, therefore, trained bumblebees (Bombus impatiens Cresson (Hymenoptera: Apidae)) to orient to, and discriminate between, artificial flowers (coloured paper or acetate discs) and then to solve several complex operations to obtain hidden reward, 50% sugar syrup held in 0.5 ml or 1. 5 ml centrifuge tube as the reservoir

(Gegear and Laverty 1995; 2005).

To extend the research results of others, and to go beyond the sort of tasks bumblebee workers might encounter in nature, I hypothesised that worker bumblebees could be trained to show a high capacity for learning to manipulate items greater in complexity than natural objects would present. I approached the hypothesis by using unnatural circumstances that avoid as much as possible, potential innate reactions as could be involved with flower visitation.

2.3 MATERIAL AND METHODS

Foragers of bumblebees (Bombus impatiens (Cresson, 1863) (Hymenoptera: Apidae)) from queen-right colonies of 30–40 workers/colony (supplied by BioBest Biological Systems

Canada [Leamington, Ontario]) were used in the experiments. When not being tested, colonies

27 were provided with a constant supply of pollen and sugar syrup. Four different colonies were used in the experiments. I used standard methods (see e.g. Laverty and Plowright 1988; Chittka

1988) to train the experimental subjects to forage on artificial flowers in a screened flight cage

(Figure 1). The specific details of the training regimes for each series of tasks are described below.

2.3.1 Experimental set up

Experiments were made indoors, in screened flight cages (2.15 m long × 1.20 m wide ×

1.80 m tall) with grey floors. A moveable screen on one side of the cages allowed experimenter access. Four colonies were used per experiment. Each was connected to a small, outer cage

(30×23×20cm) attached to the main flight cage by a gated, wire-mesh tunnel that allowed experimental control of the bees’ entry to and egress from the flight cage. Colonies, when not being tested, had constant pollen supplies and their diets were supplemented with sugar syrup.

Individually, foragers were marked on the thoracic dorsal surface with uniquely numbered and coloured Opalith tags (Plättchen, Christian Graze KG, Germany).

The experimental arena, from where the bees entered and exited (Fig. 1), comprised a green styrofoam plate 45x35x5cm, that was placed in the flight cage 165 cm, with 8 holes to hold 8 artificial flowers. The artificial flowers were made from centrifuge tubes (1.5 ml) and coloured surrounds (e.g. 9 cm coloured paper discs or 7 cm coloured acetate discs (see Figs 2 -

3). The tubes, hidden from the bees, were supplied with a reinforcer reward of 50% sucrose solution (syrup) (w:w), the amount of syrup was not controlled, but was replenished as soon as it was depleted to the extent that 3/4 of the foraging bee’s body-length was inside the centrifuge tube.

Bumblebee hive

Flight cage (testing arena) Flower Mesh patch tunnels Holding area

White cardboard gates

Figure 2.1: A plan view of the experiment setup with hive, holding area, flight cage testing arena, patch of artificial flowers and mesh tube routes with gates by which the bees were allowed to enter and exit the flight cage. The bees, in training or as trained, exited from the hive and could take only one route through the holding area to the testing arena in the main flight cage. The exiting bees were not allowed to use the diagonal route because its gate was kept closed. The gates after the holding area were opened and closed to allow single bees to enter the testing arena during testing. The bees returned to their hive from the testing area via the diagonal mesh tube route, the gate of which was opened as necessary to let the tested bee enter her hive. Eight artificial flowers arranged in two rows were used in each experiment.

2.3.1.1 Experimental procedure

The first step was to allow naïve bees to encounter artificial flowers. Once they were accustomed for a week to ten days to foraging at the simple artificial flowers, consisting of a centrifuge tube with surrounded blue disc, they were marked individually and challenged with an operant conditioning task as described for each experiment below. Each operant condition started with the simplest flower type (a) and one marked bee that was allowed to forage at it. The time that the bee spent to get the syrup reinforcer was recorded for 10 foraging trials (1 trial was 1 visit to the flower patch followed by the bee’s return to the hive). I quantified the rate of learning by 1) timing how long each bee took to manipulate the artificial flowers (from landing at the point of access to the reinforcer to its actual starting to imbibe the reinforcer) presented in each increasingly complex task and 2) by comparing the rates of learning for different tasks and levels of experience (numbers of trials completed). I noticed, as the experiments progressed, that individual bees behaved differently in manipulating the artificial flowers. Although the diverse techniques that bees used to get access to the syrup could not be quantified, they were recorded descriptively for Operant task 2. All marked bees which succeeded in manipulating the simplest- flower (Type a in each experiment), so passing the initial test, were challenged in subsequent experimental stages, but those that failed were eliminated. The subsequent stages involved operant tasks whereby the bees had to deviate from moving straight to the reinforcer syrup in the centrifuge tubes by dragging obstacles aside from the tubes’ openings.

2.3.2 Operant task 1: Sliding single and “multi” caps (Fig. 2)

This task required the bee to slide a cap or multiple caps covering the syrup laterally either to the right or the left for one cap, and to the right for “multi” caps. One bumblebee colony was used, and 16 individually marked bees.

2.3.2.1 Artificial flowers: Artificial flowers were made from centrifuge tubes (1.5 ml) and 9 cm diameter blue paper disc surrounds (Fig. 2). Four types of artificial flowers were used; the tube’s cap was cut off and attached to the stick of a cotton swab making the barrier to the tube

4.8 cm long (3 cm of the swab stick + 1.8 cm of the cap of the centrifuge tube). The cotton end was pinned down as the anchor/hinge at the edge of the flower to allow the cap to be swung to the left or right. The four successive operant tasks thus involved: Type a) Cap (inverted to allow it to slide off the tube’s opening) had to be moved in one of two directions (right or left), because two bent pins, at 1 cm distant either side, crossed the swab stick: Type b) was similar to Type a) of Fig. 1 but with the pin bent near the cotton swab to prevent the cap to be moved to the left:

Type c) was similar to Type b) (Fig 2) but with two inverted caps, cut along a cord about 1/8 diameter in from the circumference and then mounted beside each other with sticky tape; this involved greater movement by the bees: and Type d) was similar to Type c) directly above, but three caps were mounted linearly, required the bees to push them yet further to reach the sugar.

2.3.2.2 Experimental procedure

This operant condition started with presenting 8 artificial flowers of the simplest flower type (a) that includes one cap and could be slid either to the right or to the left. One marked bee was allowed to forage on them. To succeed, the bees had to fly to the artificial flower and then move the cap aside, away from the entrance to the centrifuge tube. Each bee that succeeded was allowed to advance to the next test of moving the cap or the series of connected caps aside to the right, by first 1, then 2, then 3 cm (i.e. about 2.5 x the bees’ body lengths) to access the reinforcer syrup. The time that the bee took to get the syrup reinforcer was recorded for 10 foraging trials.

A Centrifuge tube beneath the cap

Centrifuge tubes B * *

* * * Paper disc Centrifuge tube caps a) Slide to right or left b) Slide to right Centrifuge tube caps * * * *

Pendent pins Cotton c) Slide to right d) Slide to right Swab

Figure 2.2: Operant condition 1: Multi capped artificial flowers. The tube’s cap was cut off and attached to the stick of a cotton swab (3 cm long + 1.8cm cap of the centrifuge tube). The cotton end was pinned down as the anchor/hinge at the edge of the flower to allow the cap to be swung to the left or right. The reinforcer syrup was available in the centrifuge tube below the centre (*) of the artificial flower. Type a) Cap (inverted to allow it to slide off the tube’s opening) had to be moved in one of two directions (right or left), because two bent pins, at 1 cm distant either side, crossed the swab stick: Type b) was similar to type a) but with the pin bent near the cotton swab to prevent the cap to be moved to the left: Type c) was similar to type b) but with two inverted caps, cut along a cord about 1/8 diameter in from the circumference and then mounted beside each other with sticky tape: and Type d) was similar to type c), but three caps were mounted linearly. A: shows the artificial flowers with caps in place as they would be before manipulation by the subject bees. B: shows the positions of the caps after they would have been moved by the subject bees.

2.3.3 Operant task 2: Rotating an upper disc

In this task, the challenges for the bees were to rotate the Eva foam upper disc either counter clockwise or clockwise over different arcs depending on the experiment. This operant condition started with the presentation of 8 artificial flowers of the simplest artificial flower type

(Figs. 2.3a and 4 a) with the centrifuge tubes uncovered, then the centrifuge tubes half covered to train the bees to become accustomed with how to rotate the upper Eva foam discs. Once a subject bee had rotated the upper disc enough, access to the reinforcer was immediately available without her having to return to the original point where she pushed the upper disc with her head

(cf. Experiment 1).

2.3.3.1 Operant conditioning 2A: Rotatable an upper blue discs (Fig. 2.3)

This task required the bee to rotate a blue disc covering the syrup counter clockwise over various arcs (25˚, 45˚, 90˚, and 180˚), or clockwise one arc (25˚). One bumblebee colony was used, beginning with 17 individually marked bees and end up with 14 (some individuals either died or did not show up during the test).

2.3.3.1.1 Artificial flowers

Artificial flowers were made similarly to those described above with centrifuge tubes (1.5 ml), but with 7cm blue acetate discs instead of paper. The centrifuge tubes were attached eccentrically and close to the edge of the disc. Then, another 7cm blue disc was cut from Eva foam (Fuzhou Vlin Plastic Products Co., Ltd. Fujian, China) made of metallocene POE elastomer blended with EVA (Ethylene Vinyl Acetate)). One edge of this second disc was cut to make a truncated sector (wedge) with its maximum dimension of quarter of the circumference. A small (2 x 5 mm) white Eva foam tag was attached to the Eva foam disc near the edge of the removed wedge sector; the tag was made to provide the subject bees with a grip and a visible

33 signal. The second cut disc was attached to the first with a central pin to allow the former’s rotation by the bees pushing the white tag.

2.3.3.1.2 Experimental procedure

In this task, the challenges for the bees were to rotate the Eva foam, upper, disc. This operant condition started with presenting of 8 artificial flowers of the simplest flower (Fig 2.3a), with the centrifuge tubes half covered to train the bees to become accustomed with how to rotate the upper Eva foam discs. One marked bee was allowed to forage on the artificial flowers at a time. Once the 17 subject bees had shown that they were accustomed to the experimental set-up by rotating the upper disc slightly (10 degrees) to allow full access to the reinforcer, the centrifuge tubes were covered and each trained subject bee had to rotated the upper disc 25˚ counter clockwise , Type b, to access the reinforcer. Bees, which succeeded in manipulating

Type b flowers by landing on the flower and rotating the upper disc in the correct direction to access the entrance to the centrifuge tube, were advanced to allow for further tasks, but those which failed were excluded from further consideration. The set of tasks of advancing complexity was for the subject bees successively move the upper disc 25˚, 45˚, 90˚, and then 180˚ counter clockwise, followed by 25˚ clockwise) (respectively 1.2, 2.4, 4.7, 9.4 and 1.2 cm of deviation from a direct line to the occluded reinforcer reservoir tube) to uncover the hidden entrance to the centrifuge tube and feed. The time that the bee took to get the syrup reinforcer was recorded for

10 foraging trials for each bee.

2.3.3.2 Operant condition 2B: Rotation discrimination: Colour as a discriminative stimulus for rotation direction: (Fig. 2.4)

This third and final task required each bee to rotate an upper disc (either yellow or white) covering the syrup either 45° counter clockwise or 45° clockwise. This final test was also repeated in extinction to ensure that operant conditioning, rather than unconditioned approach

34 responses to any sugar odour, had occurred. Artificial flowers were made in the same way as described in the previous experiment 2A (Fig. 2.3) but with white or yellow discs of Eva foam sheet over the blue lower disc. One edge of the white or the yellow disc was cut to make a truncated sector (wedge) with the dimension 1/8 of the circumference. The upper disc was attached to the lower with a central pin to allow the former’s rotation by the bees moving it in the correct direction according to colour to obtain syrup (Fig. 2.4 a).

2.3.3.2.1 Experiment 2B 1: Subjects were tested for their abilities to use the colour of the upper disc as a discriminative stimulus and rotate it 45° in the correct direction. All flowers contained syrup. Two colonies were used: for colony (I) the 18 individually marked bees were challenged to rotate the upper (white) disc clockwise and (yellow) counter clockwise (Fig. 2.4 b), while for colony (II) 16 individually marked bees were challenged to rotate the upper (white) disc counter clockwise and (yellow) clockwise (Fig. 2.4 c).

In the training period, individual bees were trained to associate yellow and white flowers with the particular direction of rotation by giving them uncovered rewarded flower patches. For the initial training, the flowers were arranged in two rows of four flowers, one row yellow and the other white. At first, the bees had to learn to rotate the upper disc (which blocked half the opening to the reinforcer tube so that the bees could not feed until they moved it) 10 degrees (i.e. about half the bee’s body length) in the direction associated with the colour. Thus, the subject bees pushed more or less directly into the opening of the artificial flower to feed while displacing the disc 10 degrees.

On the first day of the experiment, individual bees were tested to associated yellow and white flowers with the particular direction of rotation, (e.g. yellow = counter-clockwise; white = clockwise for colony I and yellow = clockwise; white = counter-clockwise for colony II). The

35 artificial flowers were segregated so that for any specific trial the subject bees from colony I encountered only a patch A) comprised four yellow flowers at the first trial and a patch B) comprised four white flowers for the next trial (Fig. 2.4 b). For trials with colony II patch A) comprised four white flowers and patch B) comprised four yellow flowers (Fig. 2.4 c), thus the presentation sequence was reversed. For the actual tests, the rotation displacement of the upper disc necessary for the subjects to feed was increased to 45° (i.e. about three body lengths). Each bee was tested for 10 trials during which the presentation of the four flowers was alternated between four yellow and four white flowers (the flower colours were not presented together).

2.3.3.2.2 Experiment 2B 2: This experiment allowed us to test if the bees were associating the direction of rotation of the upper disc on the basis of experience with gustatory cues from the artificial flowers themselves). Freshly prepared 50% sucrose syrup is unscented

(Cameron 1981) so by using syrup vs. deionised water, the bees continued to forage in expectation of finding and tasting the syrup. Those bees from colony I (14 individuals were used) that were successful in consistently moving the white discs clockwise and the yellow discs counter clockwise in Experiment 2B 1 were challenged on the day after with choices of four coloured flowers in mixed-colour patches that were either rewarding (two flowers, one white and one yellow, contained syrup) or were not (two flowers of the other colours contained only water) as in Fig 2.4d: Patches A and B. Thus, an individual bee had to first rotate the coloured disc in the correct direction according to colour to expose the centrifuge tube, and then would encounter either a syrup reward or non-rewarding water. In patch A, the flowers were arranged (i.e.

Y=water=counter clockwise; W=syrup=clockwise; Y=syrup=counter clockwise;

W=water=clockwise) and in patch B (W=syrup=clockwise; Y=water=counter clockwise;

W=water=clockwise; Y=syrup=counter clockwise). Each bee was given 10 trials, the patches A

36 and B were alternated and the durations of the bees’ manipulations of each flower were recorded along with their sequence of choices and the directions they rotated the upper discs.

2.3.3.2.3 Experiment 2B 3: Re-testing trained subjects in extinction: Once 14 of the

16 tested bees from Colony I had succeeded in associating the colour of the upper disc with rotating it in the direction required to expose the syrup, they were then tested in extinction (i.e. the subject bees would, when tested with unrewarded (empty) artificial flowers, continue to rotate correctly for a number of trails) the upper disc. I divided the 14 surviving tested subjects

(two died) to two groups, 7 bees to be tested by presenting white stimulus to rotate it clockwise

45˚, and 7 bees by presenting yellow stimulus to be rotated counter clockwise 45˚. This was done by presenting one clean flower for each individual, the first time with an empty centrifuge tube, and second without tube. Thus, each bee encountered two unrewarded artificial flowers.

2.3.4 Data analysis

One-way ANOVA repeated measures were used. Following that, relevant pair-wise comparisons were made by Multiple Comparison Procedures (Holm-Sidak method) and

Multiple Comparison Procedures (Tukey Test) to refine the comparisons while maintaining an overall significance level of 0.05. A t- test was used to compare the accurate time between the two colonies in the time they took to manipulate the artificial flowers with colour and direction discriminative. To compare between the choices of the subject for the correct rotation direction, a chi-square test of equal probability of choice (Ho) was used. The statistical program was used

Sigma Plot (Systat software INC (V. 12.2.43f)).

Pin White tag Upper disc

Centrifuge Lower disc tube

9.4cm

1.2cm 4.7cm 2.4cm A B

Figure 2.3: Operant condition 2, 2A: Rotatable disc artificial flowers. Artificial flowers were made with centrifuge tubes (1.5 ml), and 7cm blue acetate disc. The centrifuge tubes were attached eccentrically and close to the edge of the disc. Then, another 7cm blue disc was cut from Eva foam sheets. One edge of the second disc was cut to make a truncated sector (wedge) with its maximum dimension of quarter of the circumference. A small (2 x 5 mm) white Eva foam tag was attached to the Eva foam disc near the edge of the removed wedge sector. The second cut disc was attached to the first with a central pin to allow the former’s rotation by the bees’ pushing the white tag. The challenges were to rotating the upper Eva foam disc over any one of 5 arcs. Type-a): for training period, followed sequentially by type-b): 25˚, type-c): 45˚, type-d): 90˚, type-e): 180˚ counter clockwise and type-f): 25˚ clockwise)

A: shows the artificial flowers with the upper discs in place as they would be before manipulation by the subject bees. B: shows the positions of the upper discs and where they needed to be moved by the subject bees.

Plastic lower Eva foam upper discs coloured disc

b) A B

a) Centrifuge tubes

c) A B d)

Figure 2.4: Operant condition 2B. Artificial flowers of rotatable white and yellow upper discs, operant condition 2B (similar to those shown in Figure 2.3) with white and yellow Eva foam upper discs a) For the training period all the artificial flowers contained syrup and were open to visits by the subject bees. b) Experiment 2B 1: Upper discs could be rotated 45˚ clockwise for white flowers and 45˚counter clockwise for yellow flowers, each bee was tested for 10 trials during which the presentation of colours was alternated, all flowers filled with syrup. Patch A: 4 yellow flowers, Patch B: 4 white flowers. c) Experiment 2B 1: Upper discs could be rotated 45˚ counter-clockwise for the white and 45˚ clockwise for the yellow, each bee was tested for 10 trials during which the presentation of colours was alternated, all flowers filled with syrup. Patch A: 4 white flowers, Patch B: 4 yellow flowers. d) Experiment 2B 2: As b) with arranging of the yellow and the white flowers. Patch A: the 2 flowers in the middle filled with syrup (S) and the 2 at the ends filled with water (W). Patch B: reversed, each bee was tested for 10 trials during which the presentation of the flower patch was alternated.

2.4 RESULTS

In every case, the bees’ behaviour generated the typical learning curves described by

Laverty (1980), Shettleworth (2010) and Pearce (2008), with subjects becoming increasingly adept at the task presented before reaching an asymptote. Naïve bees were first trained to simple flowers and then learned to manipulate the single occluding cap (Fig. 2.5a). The bees were then tasked with moving a single cap only to the right which they did immediately. This probably represents a simple unconditioned response to the closure of the artificial flower (Fig. 2.5b). In contrast, when tasked with two cap occlusions, they needed about 6 experiences to become adept

(i.e. the bees did not improve their performance after more experiences: the learning curve reached its asymptote, see Fig 2.5c). With three caps the bees’ learning curve shows its asymptote at about the same number of experiences (i.e. 6 - 8) but they took a longer time to complete the task (Fig. 2.5d). (N. B. the learning curve does not remove the short time about 2 –

3 s, the bees spent in walking away, i.e. to the side of the entrance to the centrifuge tube by a distance of about 3 cm and then walking back to the now uncovered entrance to feed). There was significant difference between the tested subjects in the amounts of time they took to manipulate the artificial flowers depending on the type of the flower (F2,9=18; P < 0.0001), and between visits to flowers (trials) (F2,9= 9.4; P < 0.001).

80 80

70 70

60 60 50 50 40 40 y = 7.9163x-0.116 y = 11.962x-0.355 30 30 R² = 0.5331

R² = 0.6679 Mean Mean time/sec Mean Mean time/sec 20 20 10 10 0 0 a) 1 2 3 4 5 6 7 8 9 10 b) 1 2 3 4 5 6 7 8 9 10 Trial Trial

80 80

70 70

60 60 50 50 y = 55.621x-0.369 40 y = 67.643x-0.787 40 R² = 0.6669

30 R² = 0.8989 30 Mean Mean time/sec 20 time/sec Mean 20 10 10 0 0 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 c) d) Trial Trial

Figure 2.5: Operant condition 1 sliding multi caps. The mean (±SE) of the durations (seconds) taken by the bees, after landing at the centre of the disc (time = 0), to slide capped or multi capped flowers and start feeding, for 10 trials, each trial being a single flowers’ patch visit (N=16). a) Single caps could be moved right or left. b) Single caps could be moved to right. c) Double caps could be moved to right. d) Triple caps could be moved right.

The most seemingly (to human beings) complex task that I set the bees is shown in (Fig.

2.6), for Operant condition 2A with a rotatable blue disc, i.e. required the bee to rotate a blue upper disc covering the syrup counter clockwise over various arcs (25˚, 45˚, 90˚, and 180˚), or clockwise one arc of 25˚. At first the bees had open access to the reinforcer syrup (the tube was open). Then the same individuals were tasked to move the rotating lid further and further around

(either to the left or the right) to access the tube’s opening and the reinforcement. The bees rose to each subsequent task, but initially they had to become familiar with it. The task that involved rotating disc the furthest (9.4 cm, through a detour of 180°) took the longest time to learn. There was significant difference between the tested subjects in their spending time depending on the type of the flower (F4,9 = 3.1; P < 0.021), and between trials (F4,9 = 17; P < 0.0001).

Also, bees were rapidly able to associate the rotational direction with the coloured stimuli

(Fig 2.7), as in operant condition 2B. In Experiment 2B 1, there was no difference between the two colonies in the time they took to manipulate the artificial flowers, I compared the mean time of each trial from each colony by using a t test (t = 1.0; df = 9; P = 0.34), they seemed to reach asymptote performance in 6 trials. In Experiment 2B 2, when the bees were challenged with coloured flowers in mixed-colour patches that were either rewarding (contained syrup) or were not (contained water), there was no difference between the bees’ in the times they took to manipulate the artificial flowers (F9,13= 0.61; P = 0.85) (Fig. 2.8). However, there were significant difference between trials (F9,13= 24; p < 0.0001) in that the bees took less and less time as they gained experience, as expected. In addition, there was no effect in the choice of flowers, rewarded or not.

40 40

30 30

20 20 Mean Mean time/sec Mean Mean time/sec 10 10

0 0 a) 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 Trial b) Trial

40 40

30 30

20 20

10 Mean time/sec 10 Mean time/ Mean time/ sec 0 0 c) 1 2 3 4 5 6 7 8 9 10 d) 1 2 3 4 5 6 7 8 9 10 Trial Trial

Mean Mean time/sec 10

0 1 2 3 4 5 6 7 8 9 10 e) Trial

Figure 2.6: Operant condition 2 experiment 2A. The mean (±SE) of the durations (seconds) taken by the bees, after landing at on the lower disc (time = 0), to rotate the Eva foam, upper, disc over 5 arcs for 10 trials. a) Rotating the upper disc counter clockwise about 25˚ (N=17). b) Rotating the upper disc counter clockwise about 45˚ (N=16). c) Rotating the upper disc counter clockwise about 90˚ (N=16). d) Rotating the upper disc counter clockwise about 180˚ (N=14). e) Rotating the upper disc clockwise about 25˚ (N=13).

The patch included two flowers of each colour filled with syrup and two flowers filled

2 with water (Fig. 2.4d), which were visited equally often regardless of colour or reward (χ 1 =

2 0.71; P = 0.4 for colour and χ 1 = 0.014; P = 0.9 for water vs. syrup: total number of visits by 14 bees was 140 with 75 to white and 65 to Yellow; 71 to syrup and 69 to water). It was expected that the bees, having learned to associate colour and rotational direction the previous day, and having landed on a flower for the first time that day, would rotate the disc in the correct direction. That was not the case. The bees rotated the discs in the wrong direction as often as they did in the correct direction; regardless of colour (7 bees of the 14 tested used correct rotations for the appropriate colour to expose the syrup or the water and 7 did not). However, after the bees visited their 2nd 3rd and 4th flowers in sequence their performance improved and remained high at better than about 10:4 correct rotations: incorrect rotations (Binomial test probabilities ranged from < 0.01 (4th and 10th trial) through 0.02 (5th, 7th and 8th trial) to 0.06 (6th and 9th trial) and 0.12 for trial 3) (Fig. 2.9). Testing in extinction further confirmed that bees responded correctly when there was an empty tube or no tube at all. All (14) the bees rotated the disc in the direction that was appropriate for each individual bee and the task encountered.

However, over successive trials in extinction, there was a sharp decline in appropriate behaviour; indeed, most individuals did not come back for a third trial, (3 individual bees came back, but did not forage at the flowers.)

Techniques used by the bees

I did not observe idiosyncratic behaviour by the bees simply moving the caps aside

(Operant Task 1). However, in Operant Task 2, each individual bee often developed its own technique to solve the problems of rotating the upper Eva foam disc to uncover the hidden reinforcer. The following techniques were observed:

80 w

50 W White & yellow 40 Y Y Yellow & white

Mean time/Sec Mean 30 Y W W 20 Y Y W Y 10 W Y W Y W Y W Y 0 W 1 2 3 4 5 6 7 8 9 10 Trial

Figure 2.7: Operant condition 2B, experiment 2B1. The mean (±SE) of the durations (seconds) taken by the bees, after landing, to rotate coloured discs 45° in their correct direction. The data are for 10 trials by between 16 and 18 individual bumblebees tested. (n x 10) trials: yellow discs (Y) counter clockwise and white discs (W) clockwise for colony I (n = 18: yellow then white) and white discs clockwise and yellow discs counter clockwise for colony II (n= 16: white then yellow). (W= white and Y= yellow). The coloured discs (four flowers of each) were presented in alternation.

140 White and yellow

120

100

B 80

60 Mean time/ Sec time/ Mean B 40 A B A A B 20 A B

0 1 2 3 4 5 6 7 8 9 10 Trial

Figure 2.8: Operant condition 2B, Experiment 2B2. The mean (±SE) of the durations (seconds) taken by the bees, after landing, to rotate coloured discs 45° in their correct direction, white discs clockwise and yellow discs counter clockwise. In patch A the flowers were arranged, for example, as follows Y=water=counter clockwise; W=syrup=clockwise; Y=syrup=counter clockwise; W=water=clockwise and in patch B, for example W=syrup=clockwise; Y=water=counter clockwise; W=water=clockwise; Y=syrup=counter clockwise. The patches were alternated in each trial (n= 14).

0.8

0.6

0.4

0.2

rotations according rotations to colour Proportio of correctof Proportiodirectional

0 1 2 3 4 5 6 7 8 9 10 Trial

Figure 2.9. Operant condition 2B, Experiment 2B2. The proportion of correct choices the experimental bees made in rotating discs on either yellow (= rotate right) or white (= rotate left) in array (patch) of both artificial flower colours (four of each) presented simultaneously the day after they had been initially shaped in arenas of four flowers of single colours (yellow or white).

47 a) The individual landed on the upper disc from above, inserted its proboscis beneath it and then it put its head between its mid-legs on the flower then pushed against the base of the flower to rotate the upper disc (some subjects used this technique with rotating the disc about 25° and 45°), b) landing on the lower disc, inserting the proboscis underneath the upper disc then pushing and rotate with the head, and c) same as b) landing on the lower disc, inserting the proboscis and head underneath the upper disc and also holding the upper disc between the head and both middle legs, then rotating with the pressure of the hind legs and the body.

2.5 DISCUSSION

That bumblebees are capable of operant learning is well established (Laverty and

Plowright 1988; Gegear and Laverty 1995; 2005; Chittka and Thomson 2001) and they can also show manipulative problem solving (Mirwan and Kevan 2014, see Chapter Three). They can learn to manipulate structures (natural or artificial flowers) by pushing forward directly to a reinforcer of nectar or syrup. Bumblebees were shaped to accomplish more difficult tasks by making the tasks unlike (to the best of my knowledge) anything encountered in nature, specifically by requiring the bumblebees to deviate from a direct approach and move barriers sideways by two to five body lengths before returning to the now accessible reinforcer.

Bumblebees thus successfully moved variously sized combinations of caps aside, and also rotated discs through varying arcs, up to 180°, from the entrances of artificial flowers to access the reinforcer hidden beneath. For the sliding caps task (Operant 1), this required their deviating from, and then returning to, the very spot that they first had to visit to start to move the caps aside. These series of actions are more removed from responses that have been previously demonstrated, i.e. pushing past barriers directly to access a reinforcer (Laverty 1980; Laverty and

Plowright 1988; Gegear and Laverty 1995), or “detour behaviour” (e.g. Wells 1964; Collet et al.

1993; Regolin et al. 1995; Zucca et al. 2005), but not so involved as detour planning (as in Portia labiata (Aranea: Salticidea) (Tarsitano and Jackson 1994; Tarsitano 2006)). The behaviour described is not the same as detour behaviour in which the subjects learn to avoid, or navigate around, barriers on their routes to the reinforcer (Collet et al. 1993; Tarsitano and Jackson 1994;

Regolin et al. 1995; Zucca et al. 2005; Tarsitano 2006), even to the point of detour communication in honeybees (Lindauer 1971; Collett and Collett 2000).

The experiments have thus demonstrated detour behaviour by worker bumblebees in that they learned how to reach the goal even when they had no clear view of it even though they did not move around interposed obstacles (as per Zucca et al. 2005). I also showed that the bumblebees were rapidly able to associate the directionality of rotating discs with colour cues

(white versus yellow), with great consistency. In all the tasks, subjects seemed to reach asymptote in about six to eight trials. Furthermore, I was able to rule out the unlikely possibility that subjects were merely pushing towards the odour of sugar (unlikely because sucrose has no vapour pressure), by demonstrating that pre-trained subjects rotated discs in the previously correct direction even in extinction, i.e. even when sucrose was absent.

When the bees were re-challenged the next day with a task that they had learned to accommodate successfully the day before, they had apparently forgotten it (Fig. 2.9). In a previously reported study (Mirwan and Kevan submitted, see Chapter Five) with a multiple-turn maze, I also noted that bees navigating the maze on the day after having learned to do so successfully, also had to retrieve their skills, especially how to enter and turn in the correct direction.

Finally, as additional evidence that the bumblebees were successfully learning an arbitrary task, rather than merely modifying an unconditioned approach response, in Operant

Task 2, the subjects developed a variable range of idiosyncratic behaviours, that they also seemed to modify reduce the times it took them to accomplish the task.

Thus overall, despite their small brains, I found that bumblebees can become skilled at rotational and lateral manipulations in moving objects from sources of food. The learning curves show that the bumblebees tested a) accomplishing the tasks of obtaining hidden rewards, and b) that they reached asymptotic performance rapidly, in under 10 trials. This novel use of tasks seemingly outside the context of nature, and the bees’ success in tackling them, now suggests that it would be both feasible and worthwhile to subject bumblebees to yet more cognitively challenging tests. For example, are they able to learn true detour tasks without being shaped through successive approximations? True detour tasks involve learning to move away from a reinforcer in order to gain it, and some stages of development, taxonomic groups, and animals with certain brain lesions find this extremely difficult, (e.g. Wells 1964; Collet et al. 1993;

Regolin et al. 1995; Zucca et al. 2005). If yes, are they also capable of the type of detour planning shown by some spiders (Portia labiata; Aranea: Salticidea) (Tarsitano and Jackson

1994; Tarsitano 2006)), indicating a representation of the goal even when it is not directly detectable? Finally, if pre-exposed to non-rewarded manipulanda, are they able to learn the properties of physical objects in such a way as to allow them to use these to obtain sugar in a single trial, when faced with novel problems that these manipulanda could solve (cf. literature on tool-use in birds: D'Amato and Colombo 1988; D'Amato 1991; Pearce 2008, Shumaker 2011;

Snaz 2013).

50 results of these experiments indicate an unexpected capacity for learning and individualism in bumblebees in that the subjects learned to manipulate items that seem far from any natural objects or ‘contrivances’ (Darwin 1877) that they would encounter in the nature. Adaptability, whether through behavioural plasticity, physiological adaptation or evolutionary change, is crucial to species’ survival. I suggest that the complex learning capacity demonstrated in bumblebees reflects hitherto unsuspected natural capacities, perhaps akin to those observed in vertebrates, to assemble sophisticated sequences of behavioural in response to novel challenges.

As such, the bumblebee model of operant learning may be useful in understanding general and evolutionary principles of learning and cognition.

Chapter 3

Problem solving by worker bumblebees Bombus impatiens (Hymenoptera: Apoidea)

3.1 ABSTRACT

During foraging, worker bumblebees are challenged by tasks ranging from simple to complex. The study’s goal was to determine whether bumblebees could successfully accomplish tasks that are more complex than those they would naturally encounter. Once the initial training to manipulate a simple, artificial flower was successfully completed, the bees were either challenged with a series of increasingly difficult tasks or with the most difficult task without the opportunity for prior learning. The first experiment demonstrated that the bees learned to slide or lift caps that prevented their access to the reinforcer sugar solution through a series of tasks with increasing complexity: moving one cap either to the right or to the left, or lifting it up. The second experiment demonstrated that the bees learned to push balls of escalating masses (diameters 1 and 1.27 cm) from the access to the hidden reinforcer (sugar syrup) reservoir of artificial flowers. In both experiments, when bees with experience with only the simplest task (i.e. an artificial flower without a barrier to the reinforcer) were presented next with the most complex or difficult task, they failed.

Only by proceeding through the series of increasingly difficult tasks were they able to succeed at the most difficult. I also noted idiosyncratic behaviours by individual bees in learning to succeed. The study’s results can be interpreted within the context of Skinnerian shaping and possibly scaffold learning.

3.2 INTRODUCTION

Research on cognition and learning in bees (Hymenoptera: Apidae) has centred extensively on sensory discrimination of colours, colour patterns, shapes, sizes, scents, textures and combinations thereof. Another aspect of cognition and learning in bees has embraced navigation and communication (von Frisch 1967; Chittka and Thomson 2001;

Kevan and Manzel 2012). Chittka and Thomson (2001) reviewed many aspects of learning and problem solving in bees, mostly with respect to sensory discrimination and responses to various cues. Little previous work has addressed gradual versus sudden exposure to complex tasks. Several researchers have tested honeybees (Apis mellifera L.) in mazes of varying complexity (Collett et al. 1993; Zhang et al. 1996, 1998, 2000) and found that increasing the complexity of sensory cues and navigation could be learned but within the constraints of some level of confusion does occur. There has been little research examining similar abilities of dealing with gradual and sudden exposure to complex tasks in bumblebees. Laverty and co-workers (Laverty 1980, 1994; Laverty and Plowright 1988;

Woodward and Laverty 1992; Gegear and Laverty 1998) challenged various species of bumblebees with simple to complex flowers. Flower complexity reflects how well the rewards (e.g. nectar or pollen) are hidden within the floral structure (Heinrich 1976, 1979;

Laverty 1980). However, it can also be defined by the learning investment ( i.e. development of motor skills) required in floral handling (Laverty 1994). In general, it has been shown that when faced with simple flowers, bees were immediately adept at foraging from them, but when faced with complex flowers, they required time and experience to learn how to manipulate them. Further, experience with simple flowers increases bees’ handling efficiency and reduces the subsequent learning times on other simple flowers;

53 however, it did not decrease handling times on more complex flowers when they were first encountered.

Problem solving, such as that demonstrated by maze-learning and by manipulation of simple to complex (problematic) flowers, can be invoked as part of how bees learn by sensory discrimination, navigation and communication (von Frisch 1967; Barth 1985;

Chittka and Thomson 2001; Kevan and Manzel 2012; Mirwan and Kevan 2013, see Chapter six). Age and size differences have been found to have little to no effect on learning in bumblebees (Raine et al. 2006; Muller and Chittka 2012; Raine and Chittka 2012). In the present experiments, I take the approach of challenging worker bumblebees, regardless of age and size, with increasingly difficult mechanical tasks, but outside the realm of floral structure, to test the bees’ abilities to solve problems in unnaturally complex circumstances.

I avoided, as much as possible, problems that could be interpreted as variations on innate foraging activities. I hypothesis that complex flower handling can be acquired by bumblebees as a result of experience with simpler flowers.

3.3 MATERIAL AND METHODS

3.3.1 General Methods

I used standard methods (e.g. Laverty and Plowright 1988; Chittka 1998) to train the experimental subjects, adult worker bees of bumblebees (Bombus impatiens (Cresson, 1863)

(Hymenoptera: Apoidea)), to forage on artificial flowers in a screened flight cage (Fig. 3.1).

The specific details of the training regimes for each series of tasks are described below.

Bumblebee hive

Flight cage Flower testing arena patch Mesh Holding tunnels area

White cardboard gates

Figure 3.1: A plane view of the experiment set-up with hive, holding area, flight cage testing arena, patch of artificial flowers and mesh tube routes with gates by which the bees were allowed to enter and exit the flight cage. The bees, in training or as trained, exited from the hive and could take only one route through the holding area to the testing arena in the main flight cage. The exiting bees were not allowed to use the diagonal route because its gate kept closed. The gates after the holding area were opened and closed to allow single bees to enter the testing arena during testing. The bees returned to their hive from the testing area via the diagonal mesh tube route, the gate of which was opened as necessary to let the tested bee enter her hive. Eight artificial flowers arranged in two rows were used in each experiment

3.3.1.1 Experimental set-up

Three experiments were carried out in indoor screened flight cages (2.15 m long × 1.20 m wide × 1.80 m tall) with grey floors. The bees used were foragers from queen-right colonies of 30–40 workers/colony, supplied by BioBest Biological Systems

Canada, Leamington, Ontario. Moveable screens on one side of the cages allowed experimenter access. Two colonies were used per experiment (six colonies in total). Each was connected to a small, outer cage (30 × 23 × 20 cm) (holding area) attached to the main flight cage (testing arena) by gated, wire-mesh tunnels that allowed experimental control of the bees’ entry to and egress from the flight cage. Colonies, when not being tested, had constant supplies of pollen, and their diets were supplemented with sugar syrup (50 % sucrose in water). Each forager was marked on the thoracic dorsal surface by numbers and coloured Opalith tags (Plättchen, Christian Graze KG, Germany).

The experimental flower patch (Fig. 3.1) was placed in the flight cage 165 cm from where the bees entered and exited. It consisted of a green Styrofoam plate 45 × 35 × 5 cm with eight holes to hold artificial flowers made from centrifuge tubes (either 1.5 or 0.5 ml) and coloured surrounds (9 cm blue paper discs). The tubes, hidden from the bees, were supplied with 50 % sucrose solution (syrup) (1:1 w/w sucrose: water) as the reinforcer. The amount of syrup taken by the bees was not controlled, but it was replenished as soon as it was exhausted; eight artificial flowers of the same type were presented to the tested bees.

During the experiments, the flowers were refilled and the barriers (caps or balls, see below) replaced.

3.3.1.2 Artificial flowers

Artificial flowers A (Fig. 3.2 and Table 3.1). Six different types of artificial flowers were made from centrifuge tubes (1.5 ml) and 9 cm diameter blue paper disc surrounds. Each of the six types required different, but similar, handling by the bees for them to gain access to the syrup. In all six, the tubes’ caps were cut off and attached to the sticks of cotton swabs, making the barrier to the tube 4.8 cm long (3 cm of the swab stick + 1.8 cm of the cap of the centrifuge tube). The cotton end of the swab was pinned down as the anchor/hinge at the edge of the blue paper disc allows the cap to be lifted or swung to the left or right. The ways the bees had to perform the six different tasks were as follows: (a) cap (inverted to allow it to slide off the tube’s opening) had to be moved in one of the three directions (right, left or up); (b) cap (inverted) had to be moved in one of the two directions (left or up: a pin prevented movement to the right); (c) cap (inverted) had to be moved in one of the two directions (right or up: a pin prevented movement to the left);

(d) cap (not inverted) had to be moved only up (two pins on the left and right prevented lateral movement), but there was a gap between the lip of the cap and the rim of the tube; (e) cap (inverted) had to be moved only up (two pins on the left and right prevented lateral movement), but the gap between the lip of the cap and the rim of the tube was eliminated; and (f) cap was cut into halves, the parts inverted and arranged on two swab sticks, each one or both of the half caps had to be moved in one of the two ways (left—right, or up: a pin prevented the two half caps from being moved across each other).

Artificial flowers B (Fig. 3.3 and Table 3.2). These were made as above, but from

0.5-ml centrifuge tube with the same design of 7 cm blue paper surrounds. Precision plastic balls of six different masses and two different diameters were used to cover the entrances of

57 the tubes: (a) 0.375 in. (1 cm) diameter, polystyrene ball at 0.47 g; (b) 0.375 in. (1 cm) diameter, acrylic ball, solid, clear at 0.54 g; (c) 0.375 in. (1 cm) diameter, Delrin ball acetal

POM, at 0.62 g; (d) 0.375 in. (1 cm) diameter, Teflon ball PTFE acetal homopolymer, solid, natural at 0.97 g; (e) 0.5 in. (1.27 cm) diameter, polystyrene ball polypropylene, solid, natural at 1.13 g; (f) 0.5 in. (1.27 cm) diameter, acrylic ball, solid, clear at 1.27 g; and (g)

0.5 in. (1.27 cm) diameter, Teflon acetal, solid, homopolymer, natural white at 2.32 g. The balls were obtained from K-mac plastic (K-mac Plastics & Distribution, Wyoming,

Michigan).

Blue paper Inverted centrifuge cap disc

Cotton swab Centrifuge tube

Not inverted cap Pins

Figure 3.2: Artificial flowers A and their increasing complexity. Sliding caps. Type a cap (inverted) had to be moved in one of the three directions (right, left, or up). Type b cap (inverted) had to be moved in one of the two directions (left or up: a pin prevented movement to the right). Type c cap (inverted) had to be moved in one of the two directions (right or up). Type d cap (not inverted) had to be moved only up. Type e cap (inverted) had to be moved only up. Type f cap was cut into halves, and the inverted parts had to be moved in one of the two ways (left–right, or up)

Figure 3.3: Artificial flowers B and their increasing complexity. Balls of different massed occluding the syrup reservoir. Type a 1 cm diameter, polystyrene ball, white grey at 0.47 g. Type b 1 cm diameter, acrylic ball, clear at 0.54 g. Type c 1 cm diameter, Delrin ball white creamy at 0.62 g. Type d 1 cm diameter, Teflon ball white at 0.97 g. Type e 1.27 cm diameter, polystyrene ball white grey at 1.13 g. Type f 1.27 cm diameter, acrylic ball, clear at 1.27 g. Type g artificial flower with Teflon, white 1.27 cm diameter 2.32 g looked the same as type d but larger and is not depicted

3.3.1.3 Experimental procedure

The first step was to allow naïve bees to encounter simple artificial flowers (with no occluding ball or barrier). Once they were accustomed to foraging at those artificial flowers for a week to 10 days, they were marked individually and then challenged with learning tasks as described for each experiment (below). Each experiment started with the simplest flower Type (a) at which single marked bees were allowed to forage. The time that the each bee spent to get the syrup was recorded for ten of its foraging bouts in sequence (i.e. 10 trials). The time to complete the task was measured from when the bee landed on the flower until it inserted its proboscis into the sucrose solution using a stop watch. The techniques that bees used to get access to the syrup were recorded descriptively. All marked bees which succeeded at taking syrup from the initial, simplest flower were challenged in subsequent experimental stages, but those that failed were eliminated.

3.3.2 Experiments

3.3.2.1 Experiment 1: sliding and lifting caps (Fig. 3.2)

The first experiment involved the bees learning to solve a relatively complex problem after first being confronted with a simple, but related, problem using capped centrifuge tubes in artificial flowers A (see Fig. 3.2). After experience with single caps which the bees could displace in any direction (left, right or up), the bees were challenged with single caps that they had to displace in either of two directions (including up) and then in only one direction (up). To succeed in obtaining the reinforcer syrup, the bees had to displace the cap that occluded the opening to the centrifuge tube. The experiment involved trials with bees that had never encountered the challenge of Type a) flowers, the simplest.

Then, each bee was challenged through the series of flower types to Type f), the most

61 complex (Type e) of the series in which the occluding cap needed to be moved upwards.

The times the bees took to obtain the reinforcer syrup were recorded. To compare the abilities of bees that had learned to manipulate a series of artificial flowers of increasing complexity (above) and ‘inexperienced’ bees that had learned to forage at only the simplest artificial flowers (centrifuge tube without occluding cap), inexperienced bees were confronted with Type e) flowers (Fig. 3.2), the most complex flowers.

3.3.2.2 Experiment 2: pushing balls (Fig. 3.3)

The next experiment paralleled Experiment 1, but the task given to the bees was to displace occluding balls from the opening of the centrifuge tubes in the centres of artificial flowers B shown in Fig.3.3. Bees were allowed to learn the task through being confronted with a series of increasingly massive balls (Experiments 2a and b). In Experiment 2b, I also challenged bees that had learned to forage from flowers with open centrifuge tubes immediately with the most massive ball (Type g of Fig. 3.3).

Experiment 2a: increasing of the mass of the ball (six balls of increasing masses)

To succeed in obtaining the reinforcer syrup, the bees had to remove the ball. The experiment involved trials with bees that had never encountered the challenge of Type a) with the lightest ball. Then, each bee was challenged through the series of types to Type f) with the heaviest ball blocking the entrance to the tubes. The times the bees took to obtain the reinforcer syrup were recorded. Those that succeeded with the lightest ball were then challenged with the heavier balls (including, eventually, the heaviest). To compare inexperienced and experienced bees, inexperienced bees were given Type f) flowers

(Fig. 3.3), the heaviest ball, after they had learnt to forage only at the simplest (no occluding ball or barrier) flowers. The length of time each subject took to remove the ball from the

62 artificial flowers and start to forage was recorded for ten trials on each artificial flower’s type. The subjects that gave up were observed, and recorded how many of them gave up and in which trial.

Experiment 2b: Increasing of the mass of the ball (three balls of increasing masses)

In this experiment, the artificial flowers used as the artificial flowers in Experiment

2a but with flowers of Type a), Type e) and an additional flower type (Type g) with the heaviest ball available to us at twice the mass of Type e), i.e. Teflon acetal, solid, homopolymer, natural white 1.27 cm diameter, 2.32 g in weight.

The experiment started by training bees (already familiar with artificial flowers without any barriers to the syrup reservoir) to forage from artificial flowers with entrances to the reinforcer syrup reservoir occluded by the lightest ball (Type a), the time of how long it took the bees to remove the ball from the artificial flower and start to forage was recorded.

The subject bees that succeeded in feeding on the syrup were used in further tests, but the ones which gave up were eliminated. The successful subject bees then were tested for pushing ability with Type e) flowers and then Type g) flowers (with the heaviest balls). The length of time each subject took to remove the ball from the artificial flowers and start to forage was recorded for ten trials on each artificial flower type. The subjects that gave up were observed, and recorded how many of them gave up and in which trial.

3.3.3 Data analysis

The behavioural hypothesis that worker bumblebees can solve complex problems especially if experienced is tested through analyzing the mean durations taken by the bees, after landing, to slide or lift the caps or to remove the ball from the artificial flower and start to forage differed between flower complexity types. One-way ANOVA for repeated

63 measures was used. Following that, relevant pair-wise comparisons were made by Multiple

Comparison Procedures (Holm-Sidak method) to refine the comparisons while maintaining an overall significance level of 0.05. Fisher’s exact test was used to analyse the number of subjects that succeeded in manipulating the flower and obtaining the reinforcer versus the number that failed. The states program was used Sigma Plot (Systat software INC (V.

12.2.43f)).

3.4 RESULTS

All the bees that had experience with sliding and lifting caps with increasing complexity or with pushing balls of increasing masses from the artificial flowers performed the tasks up to the most complex capped-flowers (Table 3.1) or the heaviest balls (Table 3.2). However, many (Tables 3.1, Experiment 1 and Table 3.B, Experiment 2a), or all (Table 3.2,

Experiment 2b), of the inexperienced bees presented with the most complex tasks gave up.

Throughout these experiments, the time it took the experienced bees to perform the tasks did not increase as the complexity increased (Table 3.1; Fig. 3.4 and Table 3.2; Figs. 3.5, 3.6).

Table 3.1: Types of sliding and lifting artificial flowers occluded by caps, as used in Experiment 1. As the letters increase, the flower complexity increases to reach the most complex of type e) flowers, (type f) was easy to manipulate). Subject bees were inexperienced except for foraging successfully on simple artificial flowers without barriers to the reinforcer (INE), or experienced (E), the mean time of the first trial ±SE.

Experienced Number of Mean time Flower (E) or subjects tested: (seconds ± SE) for type Flower description inexperienced number that bees to succeed on Fig. 3.2 (INE) succeeded their first trial

Cap (inverted) had to INE Type a) be moved (right, left, 13:12 43.9 ± 4 or up)

Cap (inverted) had to E Type b) 12:12 11.9 ± 1.6 be moved (left or up)

Cap (inverted) had to E Type c) be moved (right or 12:12 21.9 ± 3.2 up)

Cap (not inverted) had E Type d) 12:12 20.1 ± 2.5 to be moved only up

Cap (inverted) had to E Type e) 12:12 33.8 ± 2.4 be moved only up

Cap was cut into halves, the parts E Type f) inverted had to be 10:10 6.1 ± 0.5 moved (left—right, or up)

Cap (inverted) had to INE Type e) 14:4 450 ± 52.8 be moved only up

Table 3.2: The types and diameters of balls used with their increasing masses and the numbers of bees tested vs. those that succeeded. Subject bees were inexperienced, except for foraging successfully on simple artificial flowers without barriers to the reinforcer (INE) or experienced (E) for Experiment 2a and Experiment 2b.

Test Flower Ball Diameter Mass, Number of Mean time to Subjects type Material in, cm gram Subjects tested finish the first vs. Succeeded trial, secs ±SE

Experiment 2a

INE Type a) Polystyrene 3/8, 1 0.47 15:13 113±25

E Type b) Acrylic 3/8, 1 0.54 13:13 32±5

E Type c) Delrin 3/8, 1 0.62 13:13 33±6

E Type d) Teflon 3/8, 1 0.97 12:12 24±7

E Type e) Polystyrene 0.5, 1.27 1.13 12:12 19±4

E Type f) Acrylic 0.5, 1.27 1.27 9:9 43±15

INE Type f) Acrylic 0.5, 1.27 1.27 15:6 239±19

Experiment 2b

INE Type a) Polystyrene 3/8,1 0.47 19: 15 407±59

E Type d) Polystyrene 0.5, 1.27 1.13 14: 14 29±8

E Type g) Teflon 0.5, 1.27 2.32 14: 10 103±35

INE Type g) Teflon 0.5, 1.27 2.32 14: 0 ∞

60 Type a) INE Type b) E Type c) E Type d) E Type e) E Type f) E

20 Mean time/ time/ MeanSec

0 1 2 3 4 5 6 7 8 9 10 Trial A)

700 Type e) E Type e) INE 600

500

400

300

Mean time/ time/ MeanSec 200

100

0 1 2 3 4 5 6 7 8 9 10 Trial

Figure. 3.4: Sliding and lifting caps. A) Shows the mean time of sliding and lifting caps ±(SE) for ten trials on artificial flowers (Fig. 3.2) Type a) on which caps could be moved three ways (right, left and up) IEN: inexperienced bees (N = 12). Type b) caps could be moved two ways (left and up) E: experienced bees (N = 12). Type c) caps could be moved two ways (right and up) E: experienced bees (N = 12). Type d) caps in right position and could be lift E: experienced bees (N = 12). Type e) caps upside down and could be lifted but with difficulty because of the lack of anywhere to gain purchase E: experienced bees (N = 12). Type f) caps divided and could be slid to left or right or lifted E: experienced bees (N = 10). B) Shows the mean time of lifting caps ±(SE) for ten trials on Type e) artificial flowers IEN inexperienced bees (N = 4) and caps upside down and could be lifted but with difficulty because of the lack of anywhere to gain purchase E: experienced bees (N = 12)

260 240 Type a).INE 0.47g Type b) E 0.54g 220 Type c) E 0.62g Type d) E 0.97g 200 Type e) E 1.13g Type f) E 1.27g

180 Type f) INE 1.27g 160 140 120 100

Mean time/ time/ MeanSec 80 60 40 20 0 1 2 3 4 5 Trial 6 7 8 9 10 A)

260 240 220 200

180 160 Type f) E 1.27g 140 120 Type f) INE 1.27g 100

Mean time/ time/ MeanSec 80 60 40 20 0 1 2 3 4 5 Trial 6 7 8 9 10 B)

Figure 3.5: The mean (±SE) of the duration (seconds) taken by the bees, after landing, to push balls away from the centrifuge tube entry. The data are for ten trials by between 6 and 13 individual bumblebees tested, (N × 10) observations: A) Type a) polystyrene ball 1 cm diameter at 0.47 g, INE (inexperienced bees) (N = 13). Type b) acrylic ball 1 cm diameter at 0.54 g, E (experience with the previous ball) (N = 13). Type c) Delrin ball 1 cm diameter at 0.62 g, E (experience with the previous balls) (N = 13). Type d) Teflon ball 1 cm diameter at 0.97 g, E (experience with the previous balls) (N = 12). Type e) polystyrene ball 1.27 cm diameter at 1.13 g, E (experience with the previous ball) (N = 12). Type f) acrylic ball 1.27 cm diameter at 1.27 g, E (experience with the previous balls) (N = 9). Type f) acrylic ball 1.27 cm diameter at 1.27 g, INE (inexperienced bees) (N = 6). B) Type f) acrylic ball 1.27 cm diameter at 1.27 g, E (experience with the previous balls) (N = 9). Type f) acrylic ball 1.27 cm diameter at 1.27 g, INE (inexperienced bees) (N = 6)

600 Type a) INE Type e) E Type g) INE

500

400

300 Mean time/ time/ MeanSec 200

100

0 1 2 3 4 5 6 7 8 9 10 Trial Figure 3.6: The mean (±SE) of the durations (seconds) taken by the bees, after landing, to push balls away from the centrifuge tube entry. The data are for first trial by between 6 and 13 individual bumblebees tested; the maximum allowable giving up time was set at 20 min. Labels on the horizontal axis are explained as follows, for (10) observations: Type a) polystyrene ball 1 cm diameter at 0.47 g, INE: inexperienced bees (N = 15). Type e) polystyrene ball 1.27 cm diameter at 1.13 g, E (experience with the previous ball) (N = 14). Type g) Teflon ball 1.27 cm diameter at 2.32 g, E (experienced bees) (N = 10)

A pair-wise comparison of the data indicates that the first encounter with the task

(Type a flower) took significantly longer than subsequent encounters with the same or similar tasks (Type b to Type d, and Type f flowers, Fig. 3.2) and that the bees working with

Type e flowers always took statistically longer to accomplish their task (Fig. 3.4a, b). In the sliding and lifting caps experiment, bees that did not have the opportunity for previous experience of sliding and lifting caps, only some succeeded (4 of 14 bees vs. 12 of 13 experienced bees; P < 0.01 by Fisher’s exact test), they took statistically longer times to succeed that those bees that had experience (<34 vs. 471 s (Table 3.1; Fig. 3.4 A,B)

(F2,9 = 5.48; P = 0.02).

In the experiment with pushing balls, pair-wise comparisons of the data indicate that the first encounter with the task (type a flower) took significantly longer than subsequent encounters with the same or similar (Type b to Type e, Fig. 3.3; Table 3.2 Experiment 2a) or

(Type e, Fig. 3.3; Table 3.2 Experiment 2b) and that the bees working with Type f or Type g flowers always took statistically longer to accomplish their task (Figs. 3.5 A,B and 3.6). As a result of this, 6 of 15 bees without experience in Experiment 2a and without shaping experience succeeded in obtaining the syrup with the less massive of the two most massive balls (vs. 9 of 9 that succeeded with shaping experience: a statistically significant difference, P < 0.01 by Fisher’s exact test) (Table 3.2 Experiment 2a). Even so, those bees that succeeded took much longer to succeed than did those bees that had the opportunity for learning (18 vs. 69 s (Table 3.2 Experiment 2a) (F2,9 = 7.64; P = 0.004); some struggled for as long as 20 min and gave up. None the bees that succeeded (6 of 15) showed any observed consistency of technique in how to push these balls, unlike the experienced bees, which had developed stable idiosyncratic techniques throughout their training (see below). In

Experiment 2b (pushing balls), some bees did give up even though they had previous experience, even so they did not give up at the first or second trial, only in later trials (3rd and 5th), after some learning (Table 3.2, Experiment 2b). The bees without experience in

Experiment 2b all gave up when challenged with the most massive ball, versus 10 of 14 experienced bees that succeeded (P < 0.01 by Fisher’s exact test; Table 3.2 Experiment 2b).

As these experiments proceeded, I noted that individual bees developed their own techniques to solve the problems of pushing the ball from the entrance of the artificial flowers that they encountered. I did not attempt to quantify those idiosyncrasies. The following techniques were observed: (a) from above, the individual landed on the ball, inserted its proboscis beneath it and then it put its head between its mid-legs on the flower then pushed against the base of the flower to lever the ball off; (b) same as above (a), but the bee used both middle legs without putting its head on the flower; (c) same as above (b) but the bee used both middle legs and did not insert its proboscis beneath the ball; (d) same as above (a), but the bee used one middle leg and used one hind leg to push the ball aside;

(e) as (d) but without inserting the proboscis; (f) landing on the flower, inserting the proboscis underneath the ball then pushing with both middle legs; and (g) landing on the flower, using the head to butt the ball while pushing with both middle legs. Once an individual bee had developed its own technique, it did not seem to change thereafter.

3.5 DISCUSSION

It is well known that bumblebees quickly learn to manipulate simple natural and artificial flowers, but take longer to handle more complex ones (Laverty 1980; Laverty and

Plowright 1988; Gegear and Laverty 1995, 2005). Thus, increasing speed and accuracy in handling a given complex task indicate that experience and memory are important

71 components of worker bumblebees’ learning to manipulate objects that they encounter in their natural lives. Although Laverty (1980, 1994) tested experienced and inexperienced bees for their abilities to manipulate complex flowers, he did not increase complexity in serial order. Moreover, he compared different species of bees that differed in size and tongue length and differed in behaviour from generalists to specialists. These experiments show that those abilities of increasing speed and accuracy in complex tasks extend to the bees’ abilities to solve problems in unnaturally complex circumstances where innate reactions would seem unlikely to apply (as could be involved with flower visitation).

Although Laverty (1978) reports that the behaviour of experienced bees was relatively consistent, when inexperienced bees landed on complex flowers, they needed to search for the floral access to the rewards. He concluded that even small changes in floral form had large effects on the bees’ learning to manipulate them (Laverty 1978).

Worker bumblebees learn through experience and solving problems. That learning is presumably an outcome of shaping and scaffold learning. In shaping (Skinner 1953, pp 92–

93), animals can be, for example, trained to perform tasks of increasing difficulty. Those range from rewarding naïve animals’ positive responses to simple tasks to more and more involved responses in more and more difficult tasks, just as I have shown in these experiments, especially in the experiments with the bees’ removing the increasingly massive balls occluding access to the reinforcer in which the task is the same, but becomes more difficult as the balls’ masses increase. The related notion of scaffold learning is used in human psychology to refer to increasing sophistication and complexity of a set of tasks that at first are simple when the task is introduced, and then are used to teach learners through increasing sophistication in such subjects as reading, writing, mathematics, music, use of

72 tools and social interactions (Sawyer 2006; Olson and Platt 2000). It assumes that the subject may spend longer or give up if first given overly complex tasks they can solve well if instead presented with gradually increasing degrees of difficulties. Vygotsky (1987) proposed a means of quantifying learning rates through the Zone of Proximal Development

(ZPD) which is the difference between (a) the actual development as determined by independent problem solving and (b) the potential development as determined through problem solving under guidance or incremental experience including problem solving (My addition to the concept originally restricted to scaffolding). The experiments with the bees sliding and lifting caps to access the reinforcer sugar syrup could be considered scaffolding in which the experimenters coached the bees through a series of incrementally complex tasks. The tasks presented involved the complexity of moving caps to the left or right or above to allow access to the reinforcer.

The ZPD values I obtained from these experiments are as follows: Experiment 1 with floral complexity from Type a to Type e flowers (Fig. 3.2), ZPD = 113.1–29.1 = 84 s;

Experiment 2a with difficulty from lightest to most massive ball used (Fig. 3.3)

ZPD = 69.1–17.9 = 51.2 s; and Experiment 2b with difficulty from the lightest to the most massive ball available (Fig. 3.3) ZPD = ∞–103 = ∞ s (the inexperienced bees never pushed the most massive ball from the artificial flower). The ZPD values indicate that skill acquisition in bumblebees can be measured as it is for human learning (Bransford et al. 2000).

Thus, through experience, worker bumblebees become able to solve new problems they encounter, rather just giving up as bees that have had no previous experience do.

Experiences provide bumblebees advantages in both quantitative (time) (Figs. 3.4, 3.5, 3.6)

73 and qualitative (strategy) modes. Through solving problems, learners may invent specific, sometimes idiosyncratic, techniques to solve their problems. Additional studies are required to quantify and analyse the kinds of different behaviours I noted, as Thorndike did with domestic cats (Felis catus) (see Chance1999). Other, more refined studies using my techniques and concepts may elucidate the issues of age and individual differences in learning in bees, including honeybees (A. mellifera) (Ray and Ferneyhough 1997; Pankiw and Page 1999; Raine et al. 2006; Muller and Chittka 2012) and between species (Ray and

Ferneyhough 1997; Laverty and Plowright 1988; Amaya-Marquez et al. 2008; Amaya

Marquez 2009). Also, neurotoxins, such as some insecticides in sublethal doses, change the capacities of bees to perform tasks (Thompson 2003; Alston et al. 2007; Gill et al. 2012), and my approach could be applicable to such research.

Chapter 4

Conditional discrimination and response chains by worker bumblebees (Bombus impatiens Cresson (Hymenoptera: Apidae))

4.1 ABSTRACT

I have investigated the concept that recognizing the number of elements in a pattern is important to, and part of, learning by conditioning worker bumblebees to discriminate the number of artificial nectaries (one, two, or three centrifuge tubes inserted into artificial flowers) from which they could forage in association with their location in a three-compartmental maze, the mean visitation of each compartment was 17 visits for compartment 1, 31 visits for compartment 2 and 25 visits for compartment 3, visiting flowers with one, two and three nectaries respectively. Additionally, I challenged the bees to learn to accomplish three different tasks in a fixed sequence during foraging. They had to slide open doors to be able to proceed to an artificial flower patch where they had to lift covers to the artificial nectaries from which they then fed. Then, the bees had to return to the entrance way to their hive, but to actually enter, were challenged to rotate a vertically oriented disc to expose the entry hole. The mean time of their first trial was 116 sec, 22 sec and 52 sec for sliding doors, lifting caps and rotating the disc respectively. The bees were adept at associating the number of nectaries with their position in the compartmental maze, taking about 6 trials to arrive at error-free foraging. The other bees became adept, after three days of shaping, at more or less error free sequential task performances. Thus, they had learned where they were in the three-task chain sequence. The results of these experiments suggest that the tested bees may have developed a sense of sub-goals that needed to be achieved by recognizing the number of elements in a pattern and possibly chain response in order to achieve the ultimate goal of successful foraging and return to their colony. These results

75 also indicate that the bees had organized their learning hierarchically, as evidenced by their proceeding to completion of the ultimate goal without reversing their foraging path to return to the colony without food.

4.2 INTRODUCTION

In solving several problems during a single period of activity animals may use various strategies. In general, subjects are able to attain their goals by having acquired through experience the ability to respond to different objects or events as they are encountered. Upon encountering any of the particular objects they respond by conditional discrimination and learned responses. Lashley (1938), for example, conditioned rats to differentiate between upright and inverted triangles targets on black or striated backgrounds, and his results were replicated by

North and his colleagues (1958). Conditional discrimination has been demonstrated also in rabbits (Saavedra, 1975), goldfish (Bitterman, 1984) and pigeons (Carter and Werner197; Schrier and Thompson 1980; Thomas et al. 1988) trained with compounds of visual and auditory stimuli.

Pigeons also conditioned to specific combinations of color and form (Born et al 1969). Although there is an extensive literature on conditional discrimination learning by other species, however, few studied has been published to demonstrate conditional discrimination by invertebrate.

Conditional discrimination has been observed in honeybees (Apis mellijera) (Couvillon and

Bitterman 1988, 1989, 1991; Funayama et al. 1995), which were trained on two different conditional problems that required to discriminate between two differently colored objects on the basis of an odour, or that required to discriminate between two differently scented targets on the basis of same coloured objects.

As early as 1899, James noted that the appreciation of sequential events must be important in the behaviours of animals, but as Weisman et al. (1980) discussed in their

76 introductory review, little attention was paid to that aspect of learning and cognition. Response chain learning, also called serial recognition (Pearce 2008) and linking (Taylor et al 2010), involves the subject acquiring skills to perform a series of tasks in order, so that one correct response provides the cue for next and it is the last correct response that produces a reinforcer

(Skinner 2005). Chained responses may produce, or alter, some of the variables which control other responses (Skinner 2005) (as in the studies of Balleine et al. (1995) with rats that pressed a lever and then pulled a chain to obtain the reinforcer), but that situation may have been complicated by “chunking” (Terrance 1987, 1991) by which combinations of stimuli presented simultaneously (simultaneous chaining) make for improved cognition vs. single stimuli alone.

Although previous research indicates conditional discrimination occurs in invertebrate behaviour

(see below).

Conditional discrimination is shown when an animal responds differently to the same stimulus presented in differing contexts (Mostofsky 1965), and response chaining is shown when an animal responds to a fixed sequence of objects or events in order to gain a primary reinforcer

(Pearce 2008). However, sequence learning (or chaining) also involves the subjects responding by conditional discrimination and learned responses so as to continue, but is shown by the subjects having acquired, through experience, the ability to recognize different objects or events as they are encountered one after another in a particular, fixed, order (Dehaene 1999). Sun and

Giles (2001) suggest that sequence learning could be the most important and prevalent kind of learning. I consider that an animal’s ability to tackle those interconnected problems requires that the animal has some abilities to discriminate where it is in the sequence, i.e. have some means of counting its accomplishments as it progresses.

In two different experiments, both requiring learning an array of complex tasks with delayed rewards and different rewards, I trained worker bumblebees in Experiment 1 to discriminate the pattern one, two, or three artificial nectaries presented in arrays in three different compartments (recognizing a numerical pattern of objects, and also which compartment in sequence (the conditional context) they were in during the same task). In Experiment 2, the bees were trained to solve an array of complex manipulative problems with goals and sub-goals

(conditioned reinforcers). Their ultimate goal (primary reinforcer) was to exit the hive, forage and return. The sub-goals were first, opening a sliding door to gain entry to the foraging arena, second pushing up occluding caps over the artificial flowers to obtain the syrup reward reinforcer, then third rotating a disc to re-enter the hive (which, in this case, was the ultimate goal (primary- reinforcer)). Obtaining a favourable outcome (e.g., the door opens) reinforces the response (i.e., pushing the door), and reveals the operandum (e.g., the arm of the maze) that allows the animal to make the next response (i.e., enter the door and run in the arm) that will be eventually reinforced by another favourable outcome (e.g. the chain drops in), that allows another response leading to the primary reinforcer (i.e., food). In this sense, animals are performing chained behaviours.

4.3 MATERIAL AND METHODS

4.3.1 General methods

Bumblebee foragers (Bombus impatiens Cresson, (Hymenoptera: Apidae)) from queen- right colonies of 30–40 workers/colony (supplied by BioBest Biological Systems Canada

[Leamington, Ontario]) were used in the experiments. When not being tested, colonies were provided with a constant supply of pollen and sugar syrup. Four different colonies were used.

Experiments were conducted in indoor screened flight cages (2.15 m long × 1.20 m wide

× 1.80 m tall) with grey floors. A moveable screen on one side of each cage allowed access. One bumblebee colony was connected to a small, outer cage (30×23×20 cm) that served as a holding area. The holding area was attached to the main flight cage (testing arena) by gated, wire-mesh tunnels that allowed control of the bees’ entry to the flight cage and the maze. During experiments, bees exiting the hive could take only one route through the holding area to the testing arena (the gate on the diagonal route being kept closed). The gates between the holding area and the testing arena were manipulated as needed to allow single bees to enter the testing arena. Once in the testing arena, bees had to access a feeding area that was located 165 cm from entrance and exit points. The feeding area consisted of a green Styrofoam plate 45 x 35 x 5 cm with 8 holes that held centrifuge tubes (1.5 ml). The tubes, hidden from the bees, were filled with

50% sucrose solution (syrup) as the reinforcer reward. The amount of syrup was not controlled, but was replenished as soon as it was exhausted, when 3/4 of the bee’s body disappear inside the centrifuge tube of the artificial flower. After foraging in the feeding area, bees were allowed to return to the hive via the diagonal route. The first step was to allow bees to forage for syrup (a

50% w:w solution of sucrose in water) in the feeding area, at the far end of the testing arena from the colony’s access, where eight centrifuge tubes mounted in a styrofoam base. Once the bees were accustomed to foraging at those feeders for a week to ten days during this time, they were marked individually on their thoracic dorsal surfaces with uniquely numbered and coloured tags

(Opalith Plättchen, Christian Graze KG, Germany), and then challenged with learning tasks as described for each experiment (below).

4.3.2 Experiment 1: Conditional discrimination: Testing bumblebees’ abilities to recognize elements in a numerical array

Here I tested bees’ abilities to recognize three types of flowers, first with one nectary

(tube), second with two nectaries (tubes) and third with three nectaries (tubes). Each tube

(nectary) had its own entrance.

4.3.2.1 Artificial flowers

Three types of flowers (with one, two or three nectaries) were used in this experiment (Fig. 1a) Type a) flower: A simple flower was made of 7 cm of blue plastic disc. A single 0.5 ml

centrifuge tube (nectary) was attached to the disc at its centre.

Type b) flower: Same as type a) with two centrifuge tubes (nectaries) were attached to the

disc across the centre.

Type c) flower as a) with three centrifuge tubes (nectaries) were attached to the disc at the

centre arranged in a triangle.

4.3.2.2 Experimental set up

I used two cages (described above) each with one bumblebee hive attached (Fig. 4.1 a).

Two partitions were used to make a three compartment maze in each cage. The partition closest to the point at which the bees entered was a white opaque plastic sheet. The second partition was cardboard. To travel between the compartments, the bees had to fly through a square hole (15 cm2) cut through each partition. The hole closest to the point at which the bees entered the cage was 55 cm above the floor and 5 cm from the right wall. The second hole (in the cardboard) was

100 cm above the floor and 55 cm from the right wall. Thus, the subject bees had to fly a convoluted trajectory through the maze from entering the cage to reaching the back wall of the last compartment (Fig. 4.1 a) and back. At the back of each compartment an array (patch) of

80 artificial flowers was displayed at an angle of 110°. [The angle of presentation was enough to stop the sugar syrup within the tubes from running out] Each patch included four flowers from each type (1, 2, or 3 nectaries (tubes)) randomly arranged (Fig. 4.1 b).

4.3.2.3 Experimental Procedure

I started by training the bees to forage from their hives into the three compartments with flowers and their nectaries filled to the brim with 50% fresh sucrose solution. On the first day of training, I used fully loaded flowers in each compartment. The bees were allowed to forage freely. The bees emptied the flowers in the first compartment and so had to find their way into the second and forage there. Once the flowers in Compartment 2 were emptied, the bees had to progress to the third. Thus, the bees had become familiar with the flower patches, the nectaries in each flower, and the tri-compartmental arena in which to forage. On Day 2 and 3, in

Compartment 1, those flowers with a single nectary were filled with syrup, but the flowers with two and three nectaries were filled with water. At the same time, the flowers in Compartment 2 were filled, those with two nectaries with syrup and those with one and three nectaries with water. In Compartment 3, I filled the flowers so that only those with three nectaries had syrup and the rest had water. The bees were then allowed to forage freely. Thus, the bees had become familiar with the flower patches, and the numbers of rewarding nectaries in each flower in association with the compartmental sequence, all rewarding flowers were refilled after becoming depleted.

After the bees had learned that the rewarding flowers differed according to number of nectaries and in which compartment they were located, the experiment was started on Day 4.

After three days of training, I chose 24 bees (of 30 that started) that were successful in learning to travel throughout the cage and associate the number of nectaries with the rewarding

81 flowers in each compartment for further testing. For further testing I removed syrup reward, filling them with water, from all the flowers in Compartments 1 and 2. Only those flowers with three nectaries in Compartment 3 remained rewarding. After each bee had foraged successfully, the patch of flowers was rotated 180o. Thus, each bee had to enter the cage, test the flowers and the nectaries in Compartment 1, then finding only water provided, had to repeat the exercise in rewardless Compartment 2, and then had to enter Compartment 3 to obtain syrup from only those flowers with 3 nectaries. During the experiment, I recorded the bees’ choices of compartments and flowers for each of 10 trials.

Mesh tunnels

Figure 4.1: a) An elevation view of the experiment setup with hive, holding area and flight cage, testing arena, divided into a three compartment maze. The compartments were connected by holes cut into the dividers. An array of artificial flowers was mounted on Styrofoam base in each compartment. Mesh tube routes with gates allowed the bees to enter and exit the flight cage. The bees, in training or as trained, exited from the hive and could take only one route through the Holding area to the Testing arena in the main Flight cage. The trajectories of the bees show outgoing forager stopping on the floral array in the first compartment (rewarded during training but not when counting objects and accomplishments were tested), going on to the second and stopping (again rewarded during training but not during testing), and then proceeding to the third (always rewarded). On returning to their hive, the bees’ trajectories did not have stops at the floral arrays. The bees entering the flight cage were not allowed to use the diagonal mesh tube because the gate in it was kept closed. The gates after the holding area are opened and closed to allow single bees to enter the Testing arena during testing. The bees returned to their hive from the Testing area via the diagonal mesh tube route, the gate of which was opened as necessary. b) Flower patch A and B.

4.3.3 Experiment 2: Response chains: Bumble bees manipulating multiple obstacles

In this experiment, I investigated bumblebees’ abilities to operate different obstacles (Fig.

4.2 a, b and c) in serial order by presenting them to the subjects in fixed order: first, sliding doors

(squares); second lifting an obstacle to access the artificial flowers; finally rotating a disc to allow them to return to the hive.

4.3.3.1 Material

Sliding doors (Fig. 4.2 a): A black (interior and exterior) box (30 x 20 x 15cm ) with a front of plastic transparent was placed inside the testing cage and connected by mesh tube to the hive. Opposite the transparent wall the box had four holes, each one covered with a square sliding door (4 cm2) of pink Eva foam sheet which, when opened, allowed the bees to exit the box and enter the cage. These doors could be opened by being slid along horizontal tracks above and below.

Artificial flowers (Fig. 4.2 b): Eight artificial flowers were made with centrifuge tubes

1.5 ml, with the caps removed. The tubes were inserted through the edge of blue acetate discs (7 cm diameter). Each of the caps (1.5) cm was attached to a 4.5 cm long one-sided cotton swab, and the assembly anchored with a pin through the cotton so that the cap occluded the entry to the centrifuge tube. To gain access to the reinforcer syrup, the subjects had to lift the caps. A pin either side of the rod of the cotton swap prevented lateral movement. The arrays of flowers were mounted on a 45 x 35 x 5 cm rectangular Styrofoam plate.

Rotatable disc (Fig. 4.2 c): A white Eva foam disc (12cm) with one edge cut to make a truncated sector (wedge) with its maximum dimension of quarter of the circumference was equipped with a small (3 X 7 mm) black Eva foam tag attached near the short, radial, edge of the removed wedge. The tag provided subject bees with a visible signal and grip to rotate the disc.

This disc was placed vertically over the circular 1.5 cm hole providing exit from the testing cage into the mesh tube leading back to the hive.

4.3.3.2 Experimental set up

As in Experiment 1, two cages with attached bumblebee hives were used (Fig 4.2 d). The black box (Fig. 4.2 a), with a transparent plastic front (for observation), was placed 15 cm from cage’s solid wall through which the mesh tube from the hive was inserted, and close (5 cm) to the front screen wall of the cage. The square doors inside the box could to be slid to only the right to allow egress (the first reinforcer) to the main cage and to the artificial flowers. The patch of eight artificial flowers (Fig. 4.2 b) was presented at the opposite end of the cage from the hive.

On the artificial flowers, subject bees had to lift the movable caps to access the second reinforcer: sugar syrup. Finally, a rotatable white Eva foam disc (Fig. 4.2 c) hid the entrance of the mesh tube leading from the main cage to the hive. Subject bees have to rotate the disc upwards to uncover the entrance hole and gain access to their nest-mates (the third reinforcer).

4.3.3.3 Experimental procedure

After the bees learned to forage for seven days in the main cage without the door-box, with simple artificial flowers (no caps over the nectary) at the cage’s end, and no rotating disc at the hive-end of the cage, the bees were individually marked (as described above) and further training began. On the next day (Day 1), the black box was attached with all four holes open and the front covered with a black sheet to allow light to enter only from the four holes. Thus the bees learned their way through. In the middle of the same day, the black sheet was removed. On

Day 2, the doors were slid to the left to cover half of each hole. Then in the middle of the day, the doors were slid to cover almost the entire hole. The subject bees had to learn that the holes were hidden by the doors and how to slide them to the right to pass through. During Day 2, the

86 bees foraged from the simple artificial flowers. The 14 individuals that failed to slide the doors were removed from the box and eliminated from further consideration.

On Day 3, without the door-box doors being closed, the bees were trained to forage from the complex artificial flowers with caps that needed lifting (Fig. 4.2b). On Day 4, when the last task for the bees was presented with the door-box doors open and with simple artificial flowers, they had to learn how to rotate the white Eva foam disc (Fig. 4.2c), to access their hive after foraging. The disc was presented first so that it did not occlude the hole, but was then lowered to cover half of the hole, and finally to cover almost all the hole. At this stage, it was not dropped to completely occlude the hole’s location (as in the final trials). The three individuals that could not rotate the disc were excluded from further tests.

After the bees learned each of those independent tasks, on Day 5, they were tested with all three tasks: 1) door-box doors fully to the left covering the holes and requiring the sliding open of one door, 2) foraging from the complex artificial flowers with movable caps in place requiring lifting to access the syrup, and 3) to rotate the white disc, now completely covering the hive entry tube, to enter to their hive. The time that it took each individual bee to finish each task and the total obstacle course was recorded for 10 trials for each bee.

Hole Eva foam sliding square Pins Centrifuge tube Black Eva foam tag cap Black pin Blue disc

White 4.5 cm one Entry disc side cotton swab Entry hole a) Black box b) Centrifuge tube c) Tracks Cage wall

Figure 4.2: A plan view of the experiment setup to investigate bumble bees solving sequential obstacles: a) front view of the black box, b) pushing up cap artificial flower, c) Eva foam rotatable white disc mounted on the testing arena wall to block the entry, d) Experiment setup with hive, holding area, flight cage testing arena, an artificial flower patch, mesh tube routes with gates by which the bees were allowed to enter and exit the flight cage.

4.3.4 Data analysis

Experiment 1: Conditional discrimination: Recognizing the number of nectaries (tubes in artificial flowers (Fig. 4.1)). A one-way completely randomized ANOVA was used to compare bees’ choices of the correct flower (number of nectaries in the respective compartment) on each patch.

Experiment 2: Response chains: Manipulation of multiple obstacles (Fig. 4.2). To compare between bees and their individual durations to finish each obstacle and the total course, an one way repeated ANOVA measurement was used. To identify all pair-wise multiple differences, Tukey’s tests were used.

4.4 RESULTS

4.4.1 Conditional discrimination: Recognizing the number of nectaries (Experiment 1)

These results showed that bumble bees have the capability of recognizing the number of elements in an array of one, two and three. The subjects learnt to choose the rewarding flowers of each patch in direct accordance with the compartment they were in. The mean value of flower choice of each patch corresponding with previous experience was great enough to conclude that their choice was not random and their previous experience affected that selection.

In Compartment 1, the bees chose one nectary more than they did 2 or 3 nectaries (F2,9 =

46.42; P < 0.0001). In comparing the performances (choosing an artificial flower with a single nectary) of all the individual bees in Compartment 1, inter-individual difference in learning performance between tested bees was not significant (F2,9 = 2.78; P = 0.11) (Fig. 4.3 a).

In Compartment 2, the bees chose flowers with two nectaries more than they did one or three nectaries (F2,9 = 124; P < 0.0001). In comparing the performances (choosing an artificial flower with two nectaries) of all the individual bees in Compartment 2 inter-individual difference

89 in learning performance between tested bees was not significant (F2,9 = 1.25; P = 0.32) (Fig 4.3 b).

In Compartment 3, the bees chose three nectaries more than they did one or two nectaries

(F2,9 = 534; P < 0.0001). In comparing the performances (choosing an artificial flower with three nectaries) of all the individual bees in Compartment 3, inter-individual difference in learning performance between tested bees was not significant (F2,9 = 1.15; P = 0.37) (Fig 4.3 c).

Not only had the bees learned to associate the number of nectaries with the compartment in which they were foraging, they had learned that Compartments 1 and 2 were rewardless, even though they had single and double nectaries in the arrays of artificial flowers presented (Fig 4.3 d). Eight of the 24 bees in their final trial learned to traverse Compartments 1 and 2 without stopping. Some bees (7), even in their final trial, did stop in Compartment 1 and probe single nectary artificial flowers with only water in them, but then flew on. Also (11) bees in their final trail did not stop in Compartment 1 and then probed double nectary artificial flowers with only water in them in Compartment 2, but then flew on to forage correctly in Compartment 3. (Five) bees in their final trail did stop in Compartment 1 and probe single nectary artificial flowers with only water in them, and then stopped in Compartment 2 and probed double nectary artificial flowers with only water in them, but finally flew on to forage correctly in Compartment 3.

4.4.2 Response chains: Manipulating multiple obstacles

These results show that worker bumblebees can learn to operate three different and complex tasks in serial order. The mean durations among the bees to solve first obstacle, sliding doors, were not significantly different (F9,9 = 1.9; P = 0.18), however, as expected there was a statistically significant difference between trials as the bees learned (F9,9 = 20; P < 0.0001) (Fig.

4.4 a).

With second obstacle, lifting caps occluding access to the nectary of an artificial flower, again the bees did not differ in manipulation time amongst themselves (F9,9 = 1.5; P = 0.28), but there was a statistically significant difference between trials as the bees learned (F9,9 = 4; P =

0.026) (Fig. 4.4 b).

For the final obstacle, rotatable disc, these results show same pattern as with the other obstacles of no significant difference between the bees for their manipulation times (F9,9 = 1.2; P

= 0.40) but significant differences between trials (F9,9 = 33; P < 0.0001) (Fig. 4.4 c).

In assessing the bees’ performances in completing the whole “obstacle course” I also found no significant difference between the bees for their manipulation times (F9,9 = 2; P = 0.16) but significant differences between trials (F9,9 = 37; P < 0.0001) (Fig. 4.4 c).

As expected, the different tasks took different times to accomplish and were learned at different rates (Fig. 4.4). Nevertheless, in the final trials the bees took almost the same amounts of time to manipulate each obstacle (Fig. 4.4).

The learning rates were estimated by two statistics. I subtracted the mean time for completion of the task from the time it took on the first encounter for each of the 10 bees. I also fit the learning curves to a power function (see Figs. 4.4 a, b, c) (Ritter and Schooler 2002) and used the exponent as a measure of learning rate (Table 4.1).

As these experiments proceeded I noted that the bees employed similar techniques to sliding the door, lifting the cap, and rotating the disc. To slide doors to the right they used their right middle (mesothoracic) legs. To lift caps they inserted their proboscides beneath the cap and then lifted with their proboscides and mesothoracic legs. To rotate the disc preventing immediate entry to the hive, they pushed with their front and both mesothoracic legs.

Compartment 1 Compartment 2 30 50 45 25 40 20 35 30 15 25 20 10 15 5 10

5 Mean number of visitation of number Mean Mean number of visitation of number Mean 0 0

1 N 2 N 3 N 1 N 2 N 3 N Flowers with Flowers with a b

Compartment 3 40 Three Compartments 35 30 30 25 25 20 20 15 15 10 10 5 5

Mean number of visitation of number Mean 0 0 1 N 2 N 3 N Mean number visitation of of 10trials 1 N 2 N 3 N 1 N 2 N 3 N 1 N 2 N 3 N Flowers with Compart 1 Compart 2 Compart 3 Compartments c d

Figure 4.3: The mean number of visitation (N=24 bees in each of 10 trials) landing on one of the three training flower types. The flower type rewarded depended on the number of nectaries (1N, 2N and 3N (tubes)) and the compartment (Fig. 1) in which they were presented. a) number of bees landing on the three types of flowers on the patch in compartment 1, b) number of bees landing on the three types of flowers on the patch in Compartment 2, c) number of bees landing on the three types of flowers on the patch in Compartment 3, and d) the number of visitations (i.e. bee-flower visits) to each flower type in the three compartments when only those flowers with three nectaries in Compartment 3 were rewarding and all other flowers contained only distilled water (in Compartment 1, most visitations were to the flower type with 1 nectary (which had been rewarding during initial training) , in Compartment 2 flower type with two nectaries (which had been rewarding during initial training) were the most visited, and in Compartment 3 flowers with three rewarding nectaries (which had also been rewarding during initial training) were almost exclusively visited.)

50 140

120 40

100 30 y = 21.85x-0.445 80 R² = 0.7091 60 20 y = 116.68x-1.165 40 R² = 0.9322 time/Sec Mean Mean time/ time/ MeanSec 10 20 0 0 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 Trial Trial a) Sliding doors b) Lifting caps

70 Figure 4.4: The mean (±SE) of the durations

60 (seconds) taken by the bees, to manipulate

50 each obstacle (N=10) during 10 trials and 40 power function best fits to the curves of y = 38.222x-0.955 30 results. a) sliding doors to the right to enter R² = 0.9296

20 the flight cage and forage, b) pushing up Mean time/Sec Mean 10 caps to start feeding, c) rotating disc to get

0 access to their hive. Comparisons of the 1 2 3 4 5 6 7 8 9 10 times taken to solve each of the three tasks Trial (a, b, c) indicate that the times taken for the c) Rotating disc 1st trial for each of 10 bees for each obstacle differed significantly (F2,9 = 12; P=0.0002), but by the 10th trial for each of 10 bees for each obstacle the times taken were

statistically the same (F2,9 = 2; P = 0.19).

Table 4.1: Learning rates of worker bumblebees for the different obstacles to be manipulated as calculated by the difference in time between first and tenth encounters, and by the exponent of the power function for the learning curve over 10 trials.

Obstacle type Time on first Time on Learning difference Power function encounter 10th trial (first – last) exponent

Lifting caps 28.1 6.8 21.3 - 0.44

Rotating discs 53.5 4.6 48.9 - 0.95

Sliding doors 112.3 7.3 105 - 1.16

4.5 DISCUSSION

The primary question addressed was “can bumblebees learn to perform a sequence of complex tasks using conditional discrimination and a fixed order, in a response chain?” I consider that an animal’s tackling a sequence of tasks requires that it have some understanding of where in the sequence it is (the conditional context), i.e. it should have some means of counting its accomplishments as it progresses.

In the first experiment on conditional discrimination, I trained the bees to associate the number or pattern of nectaries (1, 2, and 3N) with the position that the bees found themselves in navigating a three compartment maze. The subject bees were able to make that association so that they learned to forage on artificial flowers with one nectary (but not those with two or three) in the first compartment, two nectaries (but not those with one or three) in the second, and, in the most complex challenge, they foraged at artificial flowers with three nectaries (but not those with one or two nectaries) in the third compartment of the maze. Others have shown counting abilities in a wide variety of animals (Pearce 2008), including insects (landmarks and dots by honeybees

(Chittka and Geiger 1995; Dacke and Srinivasan 2008; Gross et al 2009); nectaries (Bar-Shai et al. 2011a, b) and in cicadas, time intervals (Karban et al. 2000)). Also, visual decision-making based on number recognition has been invoked in honeybees by Leppik (1953) in experiments where pattern recognition may have been more important, and by Gross et al. (2009) who showed that honeybees could learn, in a Y-maze, to distinguish two vs. three vs. four randomly placed dots, but were unable to distinguish between four and more dots. These results, although not providing explanatory mechanisms, indicate that workers of B. impatiens learned to recognize the number of nectaries, possibly by pattern recognition or by counting, subitizing (

95 immediate recognition of small number of objects) or otherwise enumerate (Pearce 2008) and knew which compartment of the maze they had reached.

Having found that worker bumblebees of B. impatiens could recognize the numerical patterns of objects (as shown for B. terrestris by Bar-Shai et al., 2011a, b), I challenged them in a second experiment to learn to accomplish complex tasks in a fixed sequence. Although the experiment does not demonstrate chaining, these results correspond to sequence recognition and sequential decision making through several actions: goal-oriented, trajectory-oriented, and reinforcement-maximizing, all of them lead to the goal (in these experiments, obtaining sugar syrup and returning home) at the end (Sun 2001). These results show that the bees learnt a series of complex tasks 1) sliding doors to gain entry to their foraging arena, 2) orienting to an array of eight artificial flowers and lifting the caps covering the nectaries to obtain the hidden sugar syrup and 3) returning to the entry point to their hive where they had to rotate a disc to expose the tunnel to the hive. Thus they accomplished solving a multistep set of problems in a fixed order.

Previous studies have shown that many kinds of animals, including insects can navigate or overcome barriers (ants (Chameron et al. 2011), honeybees (reviewed by Collett et al. 1993), and bumblebees (Chittka and Thomson 1996; Mirwan and Kevan submitted)). Menzel (1990) suggested that honeybees do have, to some extent, an internal representation of sequential tasks.

Martin (1965) showed that honeybees were able to react to a series of four scents, of which only one sequence of odours lead to the reward. Gegear and Laverty’s (1995, 1998) research indicates inter-task interference constrains solution of several tasks presented simultaneously, not in sequence. These experiment is different because it involves three rather different obstacles, each presented in sequence, that need manipulation skills (sliding, lifting, and rotating) coupled with navigation and numerical pattern recognition (perhaps even with counting).

These experiments and their results are different from associative chaining or linking theory (e.g., in which each response becomes the stimulus for the next) (Spiegel and McLaren

2006) and numeracy because I have linked the relationship between recognizing numerical arrays of objects while knowing where in a sequence of tasks an animal might be. That combination relates to problem solving, navigation and orientation, and cognition. Moreover, I have presented to bumblebee workers a fixed series of tasks that may demonstrate that they organize learned behaviours hierarchically into behavioural chains with goals and subgoals

(Byrne and Byrne 1993; Byrne and Russon 1998). After passing through the sliding door box, the bees have a choice of returning to their hive or proceeding to the floral array, lifting the occluding caps to forage before returning home. Clearly, their priority was to proceed and forage.

The second, and more complex type of response chains includes problems in which tasks may include barriers which require operation by the test subjects, and all obstacles must successfully negotiated in order to complete the task, as shown for pigeons (Columba livia;

Kohler 1925; Birch 1945; Epstein et al. 1984; Epstein 1987), New Caledonian crows (Corvus moneduloides; Taylor et al. 2010), chimpanzees (Pan troglodytes; Döhl 1968) and bumblebees

(Bombus impatiens; Mirwan and Kevan 2014). The size and type of obstacles vary according to the tasks, but response chaining coupled with problem solving has not been tested in invertebrates until now, as far as I know.

I cannot yet invoke chaining (sequence learning) as the learning paradigm of my test bees because I did not vary, or provide choice, in the sequence of tasks. I am sure that such experiments would yield interesting results.

Pattern recognition, counting and solving complex problems by response chaining are part of the foraging strategies of bees. These results show that the capacity of worker bumblebees to solve different complex problems through conditional discrimination and in a particular order involves their knowing where in a sequence of tasks they are (counting and prioritizing the tasks) and being able to recognize a numerical array of objects as they perform some tasks. Gegear and Laverty (1998) found that the capacities of bees to perform different complex tasks when presented simultaneously was constrained, but these experiments show that such constraints are not as severe when different complex tasks are presented sequentially. Floral fidelity/constancy may be constrained by bees being unable to perform several complex manipulative tasks within a mixed patch of flowers, but are likely less constrained when moving between patches of different flowers, even on single foraging trips. Their abilities to perform different complex tasks in sequence may be an important part of bees’ capacities to sample different flowers as they forage, and explain how they “major” and “minor” (Heinrich 1979) when tracking changing resource availabilities over time and space.

Chapter 5

Maze learning and route memorization by worker bumblebees (Bombus impatiens

(Cresson) (Hymenoptera: Apidae)

5.1 ABSTRACT

Bumblebees can learn to navigate by walking (as they do in nature within topographically complex spaces containing their nests), rather than flying, through complex mazes in the absence of specific visual, chemical or textural cues and retain their memories of the route for most of their adult lives (at least up to 15 days). Bumblebees can navigate through complex mazes by memorizing the entire sequence of appropriate turns, and their choice of correct first turn. Thus, their observed proficiencies indicated that the individual bumblebees had each memorised the maze by learning motor sequences which were not linked to visual, chemical or textural stimuli, and that their memories were triggered by contextual cues associated with the bees' positions in a sequence. Their performances over 5 to 15 days of not having gone through the maze to forage remained unaltered after 5 days, but after 10 and 15 days the bees’ performances worsened and because they did not retrieve their memories to former levels of performance, it seems some level of confusion can be inferred.

5.2 INTRODUCTION

In orienting themselves in their environments, animals, especially central-place foragers, memorize landmarks (visual, olfactory, tactile) and other information and retrieve stored information as necessary as they navigate. It is well known that ants, bees and wasps are guided by olfactory and visual landmarks when following paths and travelling from and returning to

99 their nests (Baerends 1941; Thorpe 1950; Janzen 1971; Rosengren 1971; Heinrich 1976; Collett

1992; Collett et al. 1992; Wehner 1992; Wehner et al. 1996; Thomson et al. 1997; Ohashi and

Thomson 2012; Lihoreau et al. 2012).

Animal navigation and the cues involved have been studied under controlled environments in nature and in enclosed spaces as large as planetaria and as small as laboratory bench top choice chambers and mazes. Such environments are useful tools to study route navigation and to relate it to cognitive map building in the brains of subject animals (Anderson

2000; Munn 1950; Olton 1977; Gallistel 1990; Collett and Zeil 1998; Healy 1998; Chameron et al 1998; Anderson (ed.) 2000).

Bumblebees forage by flying to and from their nests and perform many tasks by walking.

Within the nest, the queen, workers and even drones (Cameron 1985) attend the brood, maintain and expand the nest (Sladen 1912; Free and Butler 1959; Goulson 2010). Bumblebees build their nests in piles of stones, underground in cavities, and in buildings. In such places, they must walk and navigate through topographically simple to complex spaces.

There is a long history of maze learning studies in vertebrates, primarily rats, mice, and pigeons (Dale 1988; Anderson 2000; Fu and Anderson 2006). In contrast, few studies have used mazes to study navigation in invertebrates. Most of these studies have used simple, single bifurcation choice chambers (T- or Y-mazes) to examine the sorts of cues used in navigation

(e.g. flat worms of the now-infamous Worm Runner’s Digest (see McConnell 1966)), crayfish

(Tierney and Andrews 2013), cuttlefish (Cartron et al. 2012), fruit fly Drosophila spp (Quinn et al. 1974; Hay 1975; Bicker and Spatz 1976; Tully 1984; Guo et al. 2000) and especially ants

(Hymenoptera; Formicidae) (Goetsch 1957; Schneirla 1941; Bernstein and Bernstein 1969;

Muller and Wehner 1988; Chameron et al. 1998; Wilson and Hölldobler 1990). These mazes

100 have also been used to examine cues (visual, olfactory, tactile) used by eusocial bees

(Hymenoptera: Apidae), mostly worker caste of western honeybees (Apis mellifera) (Menzel

1981; Kevan and Lane 1985; Zhang et al. 1998; 2000) and worker caste of bumblebees (Bombus spp.); (Chittka1998; Dyer et al. 2007; Dyer et al 2008; Han et al.2010). Slightly more complex are multilateral choice chambers (Colton and Samuelson 1976) in which the subject animals face an array of choices, such as of colours (as with Kevan’s (1972) multilateral visiometer), shapes

Lehrer et al. 1995), and scent (Martin 1965). Bumblebees placed in a 12-arm radial choice chamber showed an ability to distinguish radial vs. concentric visual patterns (Plowright et al.

2006; Seguin and Plowright 2007).

Multiple-turn mazes (e.g. stereotypical rat-mazes (Honzik 1936)) have been rarely used to examine navigational capacities by invertebrates. Learning of complex mazes has been demonstrated in ants and bees; they have proven to be skilled at reaching a food source by navigating mazes, mostly with sensory (visual, olfactory, textural) cues provided at choice points

(Schneirla 1929; Weiss 1953; Collett et al. 1993; Collett & Baron, 1995; Zhang et al. 1996;

Zhang et al. 1998, 1999; Chameron et al. 1988). Fruit flies (Drosophila melanogaster) have also been shown to navigate multiple bifurcation mazes with illuminated end points (Hay 1975), in completed darkness (Bicker and Spatz 1976), and with textural cues (Platt et al. 1980). Zhang et al. (1996) studied flying honeybees’ abilities to navigate complex labyrinths containing many

‘‘dead ends’’ that involved making a correct choice at each turn to achieve the reward. Zhang et al. (2000) explored worker honeybees’ capabilities to navigate in flight through four types of mazes: constant turn, zig-zag turns, irregular and variable turns. Their results showed that the bees could learn to navigate all four types of mazes but performed best in the constant-turn mazes and increasingly poorly in the zig-zag mazes, the irregular mazes, and the variable

101 irregular mazes. Although multi-turn maze learning and ambulatory navigation (such as I have used) may seem unnatural, it may be important in bumblebees’ lives within and immediately beyond the nest.

The study examines three questions:

1. Can worker bumblebees navigate complex, multiple-turn, irregular (but not variable)

mazes, consisting of several ‘‘dead ends’’ and involving making correct decisions to

accomplish the goal of passing through the mazes?

2. Can the memory of the navigational route through the maze deteriorate when the

study bees are prevented from using the maze for several days?

5.3 MATERIAL AND GENERAL METHODS

Foragers of bumblebees (Bombus impatiens (Cresson, 1863) (Hymenoptera: Apoide)) from queen-right colonies of 30–40 workers/colony (supplied by BioBest Biological Systems

Canada [Leamington, Ontario]) were used in the experiments. I started with 5 colonies, but used only the three that had the greatest similarities in learning abilities at the very start. I continued to monitor for inter-colony differences in navigation as these experiments progressed (as described below). When not being tested, colonies were provided with a constant supply of pollen prepared from honeybee collected pollen from the Honeybee Research Centre, University of Guelph and freshly made 50% sugar syrup.

5.3.1 Experimental set-up

Experiments were conducted in indoor screened flight cages (2.15 m long × 1.20 m wide

× 1.80 m tall) with grey floors. The flight cages were set up inside a flight room illuminated by

10, 60 Watt daylight fluorescent tubes set for 12 hours light (when the experiments were made) and 12 hours darkness. A moveable screen on one side of each cage allowed access. One

102 bumblebee colony was connected to a small, outer cage (30×23×20 cm) that served as a holding area (Fig. 5.1). The holding area was attached to the main flight cage (testing arena) by gated, wire-mesh tunnels that allowed control of the bees’ entry to the flight cage and the maze. During experiments, bees exiting the hive could take only one route through the holding area to the testing arena (the gate on the diagonal route was kept closed). The gates between the holding area and the testing arena were manipulated as needed to allow single bees to enter the testing arena. Once in the testing arena, bees had to navigate through the maze to access a feeding area that was located 165 cm from entrance and exit points. The feeding area consisted of a green

Styrofoam plate 45 x 35 x 5 cm with 8 holes that held centrifuge tubes (1.5 ml). The tubes, hidden from the bees, were filled with 50% sucrose solution (syrup) as the reward. The amount of syrup was not controlled, but was replenished as soon as it was exhausted. After foraging in the feeding area, bees were allowed to return to the hive via the diagonal route.

103

Bumblebee hive

Flight cage (testing arena) Feeding Mesh Holding area Mesh tunnel tunnels area Maze

White cardboard gates

Figure 5.1: A plan view of the experiment setup showing hive, holding area, flight cage testing arena, feeding area and mesh tube (tunnels) routes with gates by which the bees were allowed to enter and exit the flight cage. The bees, in training or as trained, exited from the hive and could take only one route through the Holding area to the Testing arena in the main Flight cage. The exiting bees were not allowed to use the diagonal route because the Gate in it was kept closed. The Gates after the Holding area are opened and closed to allow single bees to enter the Maze (Fig. 5.2) in the Testing arena during trials. The bees returned to their hive from the Testing area via the diagonal mesh tube route without having to navigate the Maze, the Gate of which was opened as necessary.

104

5.3.1.1 Description of the Maze (Fig. 5.2)

The maze was 13 x 13 x 2.1cm with nine rows. The path through the maze contains several dead ends and many turns to right and left. A prototype cardboard maze was made to make initial tests of experimental feasibility. The results were positive. I then had three identical mazes made to be used for the experiments reported in this paper were made of opaque white plastic (high density polyethylene) walls and floor and covered with a transparent Plexiglas lid.

In this maze, visual cues were limited as there was no colour difference between the walls and the floor, and no other cues were provided. There may have been slight textural differences resulting from the machining that left circular marks on the floor and vertical marks on the walls and perhaps shadows (imperceptible to the investigators) from the diffuse lighting in the flight room. The mazes were sufficiently small and deep that bees within them would be unable to see over the opaque channels. Directly overhead, diffused light from the fluorescent illumination entered the testing arena through the screen roof remained the same in all experiments. Because it is known that bumblebees can leave footprint pheromone as chemical signals for other bees

(Saleh et al. 2007), I used the prototype (non-washable) maze to check for possible trail-marking influences and used the same protocol thereafter for all experiments (see below).

During the experiment, the maze in use was connected to the holding cage by a mesh tunnel. That mesh tunnel provided a connection (with gates to control bees’ access to the flight cage) from the holding area (also part of the apparatus that allowed control of the bees’ access to the flight cage) to the bumblebees’ domicile (Fig. 5.1).

105

Exit

Entry

1.4cm

Figure 5.2: Multiple-turn maze with several dead ends made from a solid block of white high density polyethylene so that it had the same white colour for the walls and the floor.

106

Prior to the start of the experiments, naïve bees were allowed to forage for syrup in the feeding area for 7-10 days to accustom them to foraging in the experimental arena, but without the maze. During this period, bees had direct access from the hive to the flight cage (i.e through mesh tubes and holding area) and returning through the diagonal mesh tube. After this period of conditioning was complete, but prior to the start of the experiments, bees were marked individually on their thoracic dorsal surfaces with uniquely numbered and coloured tags (Opalith

Plättchen, Christian Graze KG, Germany). Colonies, when not being tested, had constant pollen supplies and their diets supplemented with sugar syrup.

5.3.1.2 Experiment procedure

First, the bees were restricted from flying directly to the feeder by having to crawl through a screen tube that opened into the maze. Thus, to forage, the subject had to pass through the maze to exit and then fly to the feeding area. Because some bees took a long time to accomplish the task, or failed to pass through the maze, a criterion of a maximum of 20 minutes to achieve success was established. Bees that failed to pass through the maze in 20 minutes were removed from the maze and eliminated from the experiment.

For the behavioural assay of the possible effects of trail-marking pheromone in the prototype cardboard maze, I timed how long it took for each of 15 bees (in the order of their entering the maze) tested to complete their passage through the maze and noted where in the order those bees that failed to navigate the maze came. I reasoned that the first few bees to encounter the maze would have the greatest difficulty and either fail or take a longer times that the last few bees which might have had the benefit of a pheromone trail left by previous bees. I tested that hypothesis with Spearman’s Rank Correlation. I followed this protocol for all experiments thereafter.

107

For each successful individual, and each time it passed through the maze, I recorded choices of turning left or right, traveling time (by stop watch, +/- 0.1 sec) to pass through the maze, and how many mistakes (mistakes measured as turning in the wrong direction and/or once having turned in the correct direction, changing direction) were made. I also recorded the number of times each bee navigated the maze before it did so without error. Following that, and to test for the extent of the bees’ capacities to remember how to navigate the maze, I restricted the bees access to the maze for several days (5, 10 and 15) while they had flight access from their domicile directly to the feeder (as at the very start), and then challenged them with the mazes again.

5.3.2 Experiments

5.3.2.1 Experiment 1: Maze Navigation

Each individual marked bee was allowed to enter the maze directly through the mesh tunnel leading to the maze in the flight cage. Again, each bee was left to learn its way through the maze without interference by the experimenter. For each individual bee, I recorded the first turn (left or right (correct)) and the amount time (as above) required to navigate through the maze was recorded for 20 passes through the maze (5 passes/ bee in the first day and 15 passes

/bee in the second day). I also noted the number of passes required before the bee could navigate the maze without error.

To assess that the three colonies tested comprised bees with similar in learning abilities

(see Raine and Chittka 2008), I compared 1) the amount of time it took 15 bees from each colony to pass through the maze on their first encounters with it, 2) the learning curves by repeated measures ANOVA, and 3) the ZPDs (Zone of Proximal Development) (after Vygotsky 1987) quantifies learning rates as the difference between a) the time taken to solve a problem at first

108 encounter and b) the time taken after the solution to the problem has been learned (Mirwan and

Kevan in press) for each bee according to its colony of origin.

5.3.2.2 Experiment 2: Maze re-navigation after several days

To examine the ability of the bees to remember how to navigate the maze, experienced bees were allowed to fly and forage freely in the flight cages for 5 (Colony 1), 10 (Colony 2), and 15 days (Colony 3) before being re-challenged in days 6, 11, and 16 respectively. The bees were then observed for 20 passes through the maze (10 passes in the first day and 10 in the second day), during which I recorded the first turn (left or right (correct)), length of time (as above) to pass through the maze and the number of mistakes made. Even though the preliminary tests with the prototype cardboard maze, and subsequent experiment on the plastic mazes indicated no influence of chemical signals (trail-marking pheromones) (see results below), for re- navigation I wanted to be doubly sure. Thus, the mazes were washed in 70% ETOH after each bee in each trial had been tested.

Results from other experiments suggested that social communication (Mirwan and Kevan

2013, see Chapter six) could result in lack of independence between experiments if groups of bees from the same colonies were used for more than one period of absence from the maze.

5.3.3 Data analysis

One and two way repeated measurement ANOVA was used to a compare the results of the experiments in SAS to analyze these findings. To compare between the choices of the subject

2 for the correct first turn, a chi-square test of equal probability of choice (Ho) was used.χ .

5.4 RESULTS

5.4.1 Possible Influence of Trail-marking

I found no correlation between the order in which the bees entered the prototype cardboard maze and the time it took the bees to navigate through (Spearman’s Rank Correlation

109

= -0.090; p = 0.75). In all subsequent experiments, I obtained similarly insignificant Spearman’s

Rank Correlations (for three experiments, analyses resulted in a Spearman’s Rank Correlation ranging from -0.37 to 0.31 (p= 0.17 to 0.39).

5.4.2 Maze Navigating

Worker bumblebees (B. impatiens) (from colonies 1, 2, and 3) could navigate through the maze without reference to visual cues. With increasing experience, the bees became better and better as they came to navigate the maze at their maximum speeds (Fig. 5.3). There were no significant difference between the three colonies in the times it took the bees to navigate the maze for the first time (F2,14 = 0.45; P = 0.65). There were no differences in the times it took the bees from the different colonies to navigate the maze over the 20 trials (ANOVA repeated measures F2,19 = 2.07; P = 0.14) (Fig. 5.3). The time for the first 5 trials (F2,4 = 1.3; P = 0.29) and last 5 trials (F2,4 = 2.33; P = 0.11). The ZPD values for the bees from each colony did not differ significantly either (F2,11 = 0.15 ; P = 0.86). As expected, there were significant differences between trials made during specific experiments as the bees became increasingly adept (F2,19 =

90.74; P < 0.0001).

The same similarities between colonies can be seen in Fig. 5.4 for the numbers of mistakes (about 7 mistakes for inexperienced bees on their very first encounter with the maze) vs. less than 1 for experienced bees with perfection (no mistakes) after several trials. The number of mistakes that the bees made declined so that after 2 days and 17 trials they navigated the maze without errors, there were no significant differences between the 3 groups (colonies) of bees on the first day of testing in the numbers of mistakes the bees made (F2,19 = 0.77; P = 0.47) but, as noted above, there were significant differences between the trials (F2,19 = 68.38; P < 0.0001) (Fig.

5.4).

110

The maze had the entry hole in the middle of its front wall. The bees had to choose first of all upon entering the maze between turning left or right (the correct direction for continued navigation through the maze). As expected, the inexperienced bees, regardless of colony, made random choices in their very first turns on entering the maze (expected errors by random choice

50%) (χ2 = 0.04; P = 0.84) (Fig. 5.5; Table 5.1, cell A). Once the bees became experienced their error rate on entering the maze was non-random and to the right (correct) (χ2 = 81.0; P < 0.0001)

(Fig. 5.5; Table 5.1, cell B).

111

800 800

700 700

600 600 500 500 400 400

300 300

Mean time/ time/ MeanSec Meantime/Sec

200 200 100 100 0 1 3 5 7 9 11 13 15 17 19 0 1 3 5 7 9 11 13 15 17 19 Trial Trial

a) Colony 1, N= 12 b) Colony 2, N= 14

800 Figure 5.3: The mean (±SE) of the durations 700 (seconds) taken by the bees from three different

600 colonies to navigate through the maze from their th 500 first encounter (trial 1) to their 20 trial. 400

300

Mean time/ time/ MeanSec 200 100 0 1 3 5 7 9 11 13 15 17 19 Trial

c) Colony 3, N= 12

112

10 10 9 9 8

7 7 6 6 5 5 4 4

Mean mistakes Mean 3 Mean mistakes Mean 3 2 2 1 1 0 0 T 1-4 T 5-8 T 9-12 T 13-16 T 17-20 T 1-4 T 5-8 T 9-12 T 13-16 T 17-20 Trial Trial

a) Colony 1, N=12 b) Colony 2, N=14

10 9 Figure 5.4: The mean (±SE) numbers of 8 mistakes made by bees from three different

7 colonies when navigating through the maze. The 6 findings are presented by groups of trials (trials 1 5 - 4 combined; 5 - 8 combined; to 17 – 20 4 combined).

Mean mistakes Mean 3 2 1 0 T 1-4 T 5-8 T 9-12 T 13-16 T 17-20 Trial

c) Colony 3, N=12

113

120 110 110 100

100 90 80 80 70 70 60 60 50 50 40 40 30

% of correctof % first turn % of correctof % first turn 20 20 10 10 0 0 T1-4 T5-8 T9-12 T13-16 T17-20 T1-4 T5-8 T9-12 T13-16 T17-20 Trial Trial

a) Colony 1, N=12 b) Colony 2, N=14

110 Figure 5.5: The percent (±SE) of bee-choices for 100 the correct (to the right) first turn by tested bees

90 entering the maze. The findings are presented by 80 groups of trials (trials 1 - 4 combined; 5 - 8 70 combined; to 17 – 20 combined). See also results 60 of statistical tests in Table 5.1, cells A and B. 50 40

30 % of correctof % first turn 20 10 0 T1-4 T5-8 T9-12 T13-16 T17-20 Trial

c) Colony 3, N=12

114

5.4.3 Maze re-navigation (remembering the maze) after several days

To examine the bees’ capacities to remember how to navigate the mazes after given several days off from having to navigate the maze, the bees from colony 1, were given 5 days off, bees from colony 2 were given 10 days off, and those from colony 3 were given 15 days off.

The bees retained their memories of how to navigate the mazes after periods during which they were not exposed to the challenge of navigating a maze. Figure 5.6 shows that inexperienced bees, regardless of colony, took on average 10 minutes to navigate the maze, but after 9 trials took less than 0.5 minutes.

After not having had to traverse the mazes for 5 days for colony 1, 10 day for colony 2 and 15 days for colony 3, it is evident that the bees retained memories of how to navigate the maze (Fig. 5.6). I compared experienced bees that had had various times off from using the maze, (after 5, 10 and 15 days off) vs. inexperienced (day 1 test, for each colony respectively) and found all experienced bee were quicker and more accurate in navigating the maze than were inexperienced bees (Fig. 5.6). Bees from Colony 1, when re-tested on day 6, performed much faster than they did when inexperienced, and faster than the bees from colonies 2 and 3, when they were re-tested on days11 and 16 respectively. The differences between the times taken to navigate the maze differed between number of days off (t = 2.1, df = 19; p = 0.046). Bees that had only 5 days off performed almost as well as highly experienced bees that had no days off in after 10 trials, but after 10 days (t = 0.16, df = 19; p = 0.87) and 15 days (t = 0.40, df = 19; p =

0.69) off their memories seemed to have deteriorated so that they took longer to navigate the maze in their first 5 trials than did those bees than had only 5 days off.

115

800 700

700 Day1 Day6 Day1 Day11

600 600 500 500 400 400

300 300 Mean time/sec Mean 200 time/Sec Mean 200 100 100 0 0 1 3 5 7 9 11 13 15 17 19 1 3 5 7 9 11 13 15 17 19 Trials Trials a) Colony1 b) Colony2

800 Figure 5.6: The means (±SE) of the durations 700 Day1 Day16 (seconds) taken by groups of bees to navigate

600 through the maze. Day 1 refers to bees from 500 all three colonies that had no days off from 400 navigating the maze (Colony1, 2 and 3) Day 6 300 refers to bees from colony 1 that had 5 days

200 relief from navigating the maze and were re- Mean time/sec Mean 100 tested on day 6 (N = 10 bees); Day 11 refers 0 similarly to bees from colony 2 with 10 days 1 3 5 7 9 11 13 15 17 19 relief (N= 10 bees), and Day 16 to ten bees Trials from colony 3 with 15 days off (N= 10 bees). c) Colony3

116

The same effects can be seen for the numbers of mistakes the bees made in navigating the maze (experienced bees vs. inexperienced) (Fig. 5.7). There are significant difference between colonies 1, 2 and 3 in the effects of the days off.

Those bees from colony 1 that had only 5 days off from navigating the maze retained good memory of how to navigate the maze on the first few re-encounters with it (t = 3.0, df = 19; p = 0.007). Those bees that had 10 and 15 days off from navigating the maze started by showing the same amount of difficulty as those bees that were inexperienced with the maze (i.e. in their first day of testing). Although the bees that had 10 and 15 days off (t = 6.6, df = 19; p = 0.0001) and (t = 5.1, df = 19; p = 0.0001) from navigating the maze did improve their performances, the data indicate significant differences, however they never achieved the accuracy of the bees on the first day of testing or of the bees that had only 5 days off.

As noted above, making the correct first turn choice was always less accurate at first encounters with the maze but dropped with increasing experience (Table 5.1, cell A vs. B).

Those bees that had had 5 days off from navigating the maze made apparently random choices on entering the maze, as did those bees that had 10 and 15 days off (Fig. 5.8). Statistical analyses of the complete data set (summarized in Table 5.1) indicate that the effects of days off were not statistically significant after 5 days off (t = 1.94, df = 19; p = 0.066), but were statistically significant 10 days off (t = 2.15, df = 19; p = 0.005) and 15 days off (t = 2.2, df = 19; p = 0.043). Further statistical analyses by χ2 indicate that all the bees that had time off from navigating the maze improved the correctness of their choices as their experience increased (Fig.

5.8), but differed according to the number of days off. The bees that had 15 days off performed worse than those that had 10 days off, and those performed worse than those that had only 5 days

117 off, and the last-mentioned performed as well as those that had no days off (Table 5.1, cells C to

N).

Although the total times the bees took in the first trial to navigate the maze after 5, 10 and then 15 days of not using the maze (days off) for colonies 1, 2, and 3 were low compared to

(statistically less than) first trial test with inexperienced bees, they were high compared to those from the last trials of any bees of any colony (Fig. 5.9). The durations to complete the maze rose in parallel (Fig. 5.9 a). The same groups of individually tested bees on their last trials with the maze showed a markedly greater rate of making errors if they had 10 or 15 days off from having to navigate the maze, but not if they had 5 days or no days off (Fig. 5.9 a). Moreover, the number of mistakes followed the same trend of diminishing with experience but increasing after lack of use of the maze. The rate of mistakes within the maze that experienced bees made on their first new encounters with the maze increased, especially after 10 days off, almost reaching that of the inexperienced bees (Fig. 5.9 b).

118

9 9

8 Day1 Day 6 8

7 7 Day 1 Day 11 6 6 5 5 4 4

3 3 Mean mistakes Mean 2 mistakes Mean 2 1 1 0 0 T 1-4 T 5-8 T 9-12 T 12-16 T 17-20 T 1-4 T 5-8 T 9-12 T 12-16 T 17-20 Trials Trials a) Colony1 b) Colony2

10 9 Figure 5.7: The mean (±SE) numbers of Day 1 Day 16 8 mistakes, grouped by four trials (1 – 4; 5 – 8; 9 7 – 12; 13 – 16; 17 – 20) that the tested bees 6 made when navigating through the maze. Day 1 5 4 refers to bees from all three colonies that had 3 no days off from navigating the maze (Colony Mean mistakes Mean 2 1, 2 and 3); Day 6 refers to bees from colony 1 1 that had 5 days relief from navigating the maze 0 and were re-tested on day 6 (N=10 bees); Day T 1-4 T 5-8 T 9-12 T 12-16 T 17-20 11 refers similarly to bees from colony 2 with Trials 10 days relief (N=10 bees), and Day 16 to ten c) Colony3 bees from colony 3 with 15 days off (N=.10)

119

120 120

Day 1 Day 6 Day1 Day11 100 100

80 80

60 60

40 40 20

20 correctof % first turn % of correctof % first turn 0 0 T 1-4 T 5-8 T 9-12 T 13-16 T 17-20 T 1-4 T 5-8 T 9-12 T 13-16 T 17-20 Trials Trials a) Colony1 b) Colony2

120 Day1 Day16 Figure 5.8: The percent (±SE) of bees’ choices 100 of the correct (right-hand) first turn on entering 80 the maze. Day 1 refers to bees from all three

colonies that had no days off from navigating 60 the maze (Colony 1, 2 and 3); Day 6 refers to 40 bees from colony 1 that had 5 days relief from 20 navigating the maze and were re-tested on day 6 % of correctof % first turn (N = 10 bees); Day 11 refers similarly to bees 0 from colony 2 with 10 days relief (N = 10 bees), T 1-4 T 5-8 T 9-12 T 13-16 T 17-20 and Day 16 to ten bees from colony 3 with 15 Trials days off. c) Colony3

120

800 Trial 1 Trial 20 700

600

500

400

300 Mean time/ time/ MeanSec 200

100

0 C1 D1 C1 D6 C2 D1 C2 D11 C3 D1 C3 D16

a) Mean duration

16 Trial 1 Trial 20 14

6 Mean mistakes Mean 4

0 C1 D1 C1 D6 C2 D1 C2 D11 C3 D1 C3 D16 . b) Mean mistakes

Figure 5.9: The mean (±SE) durations (seconds) bumblebees spent to navigate the maze successfully and mean numbers of mistakes. C refers to (Colony) and D refers to (Day) bees from all three colonies that had no days off from navigating the maze (C1, C2 and C3); D6 refers to bees from colony 1 that had 5 days relief from navigating the maze and were re-tested on day 6 (N = 10 bees); D11 refers similarly to bees from colony 2 that had 10 days relief from navigating the maze and were re-tested on day (N = 10 bees), and Day 16 to ten bees from colony 3 with 15 days off (N = 10 bees). a) is the mean duration that bumblebees spent on the first trials and last trials, b) the mean number of mistakes they made on the first trials and last trials.

121

2 Table 5.1: Chi-square values (χ ) of first turn choice on entering the maze (H0: equal likelihood of turning left or right) for the first and last 5 trials of three colonies combined (cells A & B) and separate (cells C – N), for tested on day 1, and re-tested on days 6, 11 and 16.

Colonies A. Inexperienced bees on first 5 trials B. Last 5 trials (i.e. from 16 – 20) in 1, 2 & 3 the first day test in the first day test combined

2 = 0.04; P = 0.8415 = 81.0; P < 0.0001 χ 1

Colony 1 C. Inexperienced bees on first 5 trials D. Last 5 trials in the first day test

in the first day test

2 χ 1 =7.84, P = 0.005 = 84.64; P < 0.0001

E. Bees from previous test (cell C) F. Bees from previous test (cell C) after 5 days off on last 5 trials after 5 days off on first 5 trials

2 χ 1 = 0.16; P = 0.69 =70.56; P < 0.0001

Colony 2 G. Inexperienced bees on first 5 trials H. Last 5 trials in the first day test

in the first day test

2 χ 1 = 0.16, P = 0.69 = 77.44, P < 0.0001

I. Bees from previous test (cell G) J. Bees from previous test (cell G) after 10 days off on last 5 trials after 10 days off on first 5 trials

2 χ 1 = 4.00, P = 0.05 = 27.04, P < 0.0001

Colony 3 K. Inexperienced bees on first 5 trials L. Last 5 trials in the first day test

in the first day test

2 χ 1 = 5.76; P = 0.016 = 81.0; P < 0.0001

M. Bees from previous test (cell K) N. Bees from previous test (cell K) after 15 days off on last 5 trials after 15 days off on first 5 trials

2 χ 1 = 0.16; P = 0.69 = 1.44; P = 0.23

122

5.5 DISCUSSION

Although bumblebees forage by flying to and from their nests, they perform many tasks by walking. Within the nest, the queen, workers and even drones (Cameron 1985) attend the brood, maintain and expand the nest (Sladen 1912; Free and Butler 1959). The Amazonian bumblebee, B. transversalis (Olivier) even forages for nest-building thatch materials along its established walking trails (Cameron and Whitfield 1996). It is well known that bumblebees build their nests in piles of stones, underground in cavities, and in buildings where they must walk and navigate through topographically simple to complex spaces. Thus, although multi-turn maze learning and ambulatory navigation (such as I have used) may seem unnatural, it may be important in bumblebees’ lives. The practical application of ambulatory navigation is exemplified in the technology that uses foraging bumblebees to deliver biological control agents against crop pests and pathogens from special walk-through dispensers on the hives (Kevan et al.

2008; Kevan et al. 2014).

I have shown that worker bumblebees (Bombus impatiens) can learn to navigate by walking through complex mazes with multiple turns and several dead ends, as used in classical conditioning with vertebrates (Honzik 1936). These results also indicate that the bees did not use chemical signals from possible trail-marking pheromones (Corbet et al. 1984; Free 1987).

Throughout these experiments, the results indicate that the colonies did not differ in their capacities to learn to navigate.

In the first set of experiment, the results indicate that foragers of B. impatiens have the capacity to negotiate multi-turn maze in the absence of visual and chemical cues. Zhang et al.

(2000) suggested that honeybees learn by using a fixed motor program, possibly linked to some form of path integration, as these results also suggest. There are few studies with bumblebees

123

(Bombus spp.) and these studies used only simple bifurcation mazes (e.g. Chittka 1998; Chittka and Thomson 1996; 1997) to study cue discrimination (choice) and memory. I am not aware of complex maze-learning experiments made with bumblebees.

The bees in this experiment with a complex maze of 9 channels, each with only one successful exit, took a long time (between 10 and 15 minutes) at first to pass through the maze, and required numerous of attempts to pass through maze without making mistakes, but improved their performances markedly after a few trials (Fig. 5.4). They navigated the mazes without mistake after 15-17 trials and did so in about 30 secs. Once bees had learned to navigate the mazes, there were no differences in their subsequent performances over 20 trials as measured by the number of errors the bees made within the maze or the time they took to successfully traverse it (Figs. 5.4 and 5.5).

Having shown that workers of B. impatiens can learn to navigate a complex maze, I tested their capacity to remember the correct route. I prevented the bees from using the mazes for

5, 10 or 15 days, during which time the maze was absent from the test arena. These results show clearly that the bees remembered the routes through the mazes for up to16 days, although their performance in the maze deteriorated the longer they had been absent from it, especially after 10 days. Each bee’s memory of the maze was assessed by the length of time it took them to pass through the maze (Fig. 5.7), the numbers of errors made within the mazes (Fig. 5.8), and whether or not they made the correct first choice on entering the mazes (Fig. 5.9). I found that the bees’ durations and error rates increased the least from 5 days off and became the most protracted after

16 day off. Interestingly, no matter how many days off the bees had, their abilities to make the first correct turn decreased to about the same level (i.e. 1:1 correct: incorrect) as that of

124 completely inexperienced bees encountering the mazes for the very first times. They did improve in making correct first turns into the maze on subsequent trials.

Interestingly, the observed deterioration in performance after different times away from having to navigate the maze was different for different parts of the task. These results show that the deterioration was greatest when it came to making the correct initial turns (the number of mistakes rose to the expected random choice value of 1:1) (Fig. 5.9). Thus, it appears that the bees had forgotten how to enter the maze, but once they were in the maze, they had sufficient recall of how to traverse the maze such that total length of time to pass through the maze was not greatly affected. This suggests that the bees invoked some sort of memory, whereby after making a correct initial choice, they were able to recall more accurately subsequent correct choices (see also Chittka 1998; Chittka and Thomson 1997). Even so, the bees I tested differed in respect of the times to traverse the maze and the number of mistakes they made within it. The bees that had

10 or 15 days relief from using the maze, unlike those which had 5 days off, were unable to navigate the maze as accurately and speedily as they did when they first learned its operation.

I conclude that the bees I tested seemed to retrieve their memories of how to navigate the mazes successfully by simple recall of learned motor sequences (Anderson 2000; Zhang 1995) not associated with visual, chemical or textural stimuli. I suggest that use of the learned motor sequences were followed by association, recollection and recognition triggered by contextual cues associated with their position in the mazes (see also Collett et al. 1993) and by relearning.

That the bees’ simple recall of which way to turn upon entering the maze deteriorated most rapidly (but was rapidly re-learned) and that, after 10 or 15 days of not using the maze, the bees did not reach their earlier performance levels suggests that the bees experienced some confusion as a result of not having continual practice. Chittka (1998) found that workers of B. terrestris

125 were able to recall coloured stimuli for up to 28 days but their performances in doing so deteriorated. These results, similar to those of Chittka (1998), show that performance did not decline overnight but, when bees were tested after a delay of several days, the durations and number of mistakes made by the bees increased. I also found that bees that were given 5, 10 or

15 days off started at a better level of performance than during the initial training. Thus, information acquired during initial training was not fully forgotten. The decline in performance may have been caused passive decay of memory or to interference from other information acquired during days off. However, interference with information stored in memory is usually limited to similar tasks (Mackintosh, 1974; Chittka and Thomson, 1997; Anderson (ed.) 2000), suggesting that passive decay is involved.

It seems that memory retrieval in workers of B. impatiens appears to work in a manner similar to that in humans and other vertebrates that have been tested (Anderson 2000; Schacter

2001).

These results support Chittka’s (1998) view that bumblebees have a large capacity for long-term memory and retain both sensory and spatial information (Lindauer, 1963; Menzel,

1969; 1990) as well as durable motor memories (Laverty 1980, 1994; Laverty and Plowright

1988; Chittka and Thompson 1997; Raine and Chittka 2007; Mirwan and Kevan 2013, in press).

As Chittka et al. (1995, 1997) noted and these results indicate, bumblebee workers seem to show difficulty in handling memories efficiently and in the appropriate context if the challenge is absent for some days (e.g. between 6 and 10 days in this case). Thus, these results indicate that little information is completely lost, although over increasing lengths of time retrieval becomes slower and more error prone (Chittka and Thomson 1997). Further experiments on age

126 dependent effects (age of the workers) would determine if long term memory is affected differently by differing periods (5, 6, 7, 8, 9 days etc.) of lack of use of learned skills.

127

Chapter 6

Social Learning in Bumblebees (Bombus impatiens): Worker bumblebees learn to manipulate and forage at artificial flowers by observation and communication within the colony.

6.1 ABSTRACT

Social learning occurs when one individual learns from another, mainly conspecific, often by observation, imitation or communication. Using artificial flowers, I studied social learning by allowing test bumble bees to a) see dead bumble bees arranged in foraging positions or b) watch live bumble bees actually foraging or c) communicate with nest-mates within their colony without having seen foraging. Artificial flowers made from 1.5 mL centrifuge tubes with closed caps were inserted through the centres of blue 7 cm plastic discs as optical signals through which the bees could not forage. The reinforcer reward syrup was accessible only through holes in the sides of the tubes beneath the blue discs. Two colonies (A and B) were used in tandem along with control (C and D) colonies. No bee that was not exposed (i.e. from the control colonies (C and D)) to social learning discovered the access holes. Inside colony B, I imprisoned a group of bees that were prevented from seeing or watching. Bees that saw dead bumblebees in foraging positions, those that watched nest-mates or non-nest-mate foraging, and those that had only in-hive communication with successful foragers, all foraged successfully, with mean time of

15 minutes, 69 secs, 78 secs, and 195 secs respectively, for the first trial. The means of in-hive communication are not understood and warrant intense investigation.

128

6.2 INTRODUCTION

Social learning is defined (by ethologists) as any learning from conspecifics

(Shettleworth 1998); however, I note that social learning between species is known, and mostly involves observation, imitation by observing and replicating another's behavior, and modeling to transmit the learned behaviour from one individual to others (Whiten and Ham 1992). Social learning through individuals’ interactions with other animals or their products (Galef et al. 1984;

Heyes 1994) encompasses attention, memory, and motivation. Social theory calls social learning a bridge between behaviourism (i.e. learning based upon behaviour that is acquired through conditioning which occurs through interaction with the environment) and cognitive learning (i.e. learning by using reason, intuition and perception) (Bandura 1977,1986; Giraldeau and Caraco

2000; Giraldeau et al. 1994). Research on social learning has focused largely on vertebrates

(Lefebvre 1995; Heyes and Galef 1996). However, a growing number of researchers have shown recently that bees and other small brained animals can also learn through acquisition of information by social transmission (Leadbeater and Chittka 2008, 2011; Giurfa 2012; Baude et al. 2008). Nonetheless, the possibility that social learning might extend to motor skills (skills, i.e. manipulating complex flowers), in addition to simple declarative knowledge (facts), remains mostly untested in invertebrates (Leadbeater and Chittka 2008).

Insects, especially eusocial bees, show remarkably complex learning abilities (Giurfa

2003, 2012; Menzel et al. 2006; Kevan and Menzel 2012) and social information often leads to the relatively long-term changes in behaviour that constitute social learning. The dance communication of honeybees (Apis spp.) (von Frisch 1967; Punchihewa et al. 1985), sounds in Melipona costaricensis (Aguilar and Briceño 2002), and other means of communication in other bees (Nieh 2004), ants (Jackson and Ratnieks 2006), wasps (Jeanne 1981; Schöne and

129

Tengö 1981) and Octopus (Fiorito and Scotto 1992) serve as examples. As Giurfa’s short, but informative review, notes simple mechanisms based on elemental associations, either Pavlovian or operant, may account for social learning in animals with miniature brains so social learning should not be considered surprising or a highly cognitive ability (Giurfa 2012).

To assess the potential for social learning in bumblebees (Bombus impatiens), I investigated the spread of a foraging techniques from experienced bees to inexperienced bees in the same and different colonies. I explored three different paradigms: a) using a model

(positioned dead bees), b) observation with imitation (of foraging live bees), and c) intra-colony possible communication within the domicile, by using a novel task they could never have done before.

6.3 MATERIAL AND METHODS:

Experiments were made in indoor screened flight cages (2.15 m long × 1.20 m wide ×

1.80 m tall) with grey floors. The bees used were foragers of Bombus impatiens (Cresson, 1863)

(Hymenoptera: Apoidea) from queen-right colonies of 30–40 workers/colony (supplied by

BioBest Biological Systems Canada (Leamington, Ontario)).

6.3.1 Experimental Procedures

Moveable screens on one side of the cages allowed experimenter access. Four colonies were used in this experiment, Colony A was placed in cage I, Colony B was placed in cage II,

Colonies C and D (control) were placed in Cage III and IV. Each was connected to a small, outer cage (30×23×20cm ) (holding area) attached to the main flight cage (testing arena) by gated, wire-mesh tunnels that allowed experimental control of the bees’ entry to and egress from the flight cage. Colonies, when not being tested, had constant pollen supplies and their diets were supplemented with sugar syrup. Individually, foragers were marked on the thoracic dorsal

130 surface with uniquely numbered coloured tags (Opalith Plättchen, Christian Graze KG,

Germany).

The experimental arena of artificial flowers (Fig. 6.1) was placed in the flight cage 165 cm from where the bees entered and exited. It comprised a green Styrofoam base 45 x 35 x 5 cm with 8 artificial flowers. The first step was to allow naïve bees to encounter simple centrifuge tubes mounted in a green Styrofoam base (the tubes were hidden and the forager could access to the syrup only through the opening of the tube). Once they were accustomed to foraging at those tubes for a week to ten days, they were marked individually and then challenged with learning tasks as described for each experiment (below).

6.3.1.1 Artificial flowers:

Artificial flowers were made of 1.5mL centrifuge tubes inserted into the centres of blue plastic discs, 7cm in diameter. The centrifuge tube was capped so that the bees could not obtain the contained syrup (50% sucrose w:w as the reinforcer, the amount of syrup was not controlled, but was replenished as soon as it was exhausted) from the surface of the plastic disc. Instead, a small hole (0.5 cm in diameters) had been drilled into one side of each centrifuge tubes just below the lip (Fig. 6.2). The artificial flower was then attached to a yellow pipette tube mounted on 35 x 52cm Styrofoam base. Eight flowers in two rows of four flowers each were arranged with the bored holes facing the central aisle between the rows of flowers and so presented to the bees in each experiment. Thus, the bees could orientate to the blue disc of the artificial flower, but could not obtain syrup except by going under the disc to the hole in the tube’s wall. Test bees were assessed based on their abilities to learn and replicate foraging behaviours without actually having performed them.

131

Bumblebee hive

Flight cage. Flower Testing patch Holding arena area

White cardboard gates

Figure 6.1: A plan view of the experiment setup with hive, holding area, flight cage testing arena, patch of artificial flowers and mesh tube routes with gates by which the bees were allowed to enter and exit the flight cage. The bees, in training or as trained, exited from the hive and could take only one route through the holding area to the testing arena in the main flight cage. The exiting bees were not allowed to use the diagonal route because the gate in it was kept closed. The gates after the Holding area were opened and closed to allow single bees to enter the testing arena during testing. The bees returned to their hive from the Testing area via the diagonal mesh tube route, the gate of which was opened as necessary. Note that the main flight cage’s end wall, through which the mesh tunnels ran, was a wooden panel so the bees in the tunnels or in the holding area could not see the flower patch.

132

Capped centrifuge tube Blue disc Figure 6.2: Artificial flowers were made of 1.5mL centrifuge tubes inserted into the centres of blue plastic discs, 7cm in diameter. The centrifuge tube was capped so that the Hole bees could not obtain the contained syrup from the surface of the plastic disc. Instead, a small Yellow hole (0.5 cm in diameters) had been drilled into pipette one side of each centrifuge tubes just below the lip.

133

6.3.1.2 The experimental groups:

1. A group of foragers from Colony A (A2) were used from which bees were tested without

a model for 2 trials with 30 minutes of giving-up time, then the model (dead bee) was

introduced and the group allowed to forage alongside models pinned in the robbing

position.

2. A group of foragers from Colony A (A1) was provided with enough food (syrup and

pollen) while imprisoned so that they could observe group A2 foragers for 10 hours.

They were then released and allowed to forage alone.

3. A group of foragers from Colony B (B1) was provided with enough food (syrup and

pollen) while imprisoned and treated as foragers in group A1 except that they were able

to watch foragers from a different colony (colony A) rather than from their own colony.

4. A group of foragers from Colony B (B2) was kept contained inside the colony and not

allowed to forage from the artificial flowers. They had opportunity to interact (inside the

nest) with group B1for 24 hours and then were allowed to forage alone.

5. Foragers from Control colonies Colony C with 15 subject bees and Colony D with 10

subject bees) were challenged to forage through “the access holes” of the artificial

flowers without models nor opportunity to watch other foragers on the artificial flowers

nor opportunity to communicate with foragers that had successfully foraged at the

artificial flowers. They were allowed 30 minutes to succeed.

6.3.2 Experiments:

Over a period of several days, individually marked bees were trained to forage from simple centrifuge tubes (described above).

134

6.3.2.1 Experiment 1: Control experiment

Following the initial training, the bees from the control colonies, Colonies C and D, were tested by challenging them with the 8 artificial flowers described and arranged as above (Fig.

6.1), with 30 minutes of giving up time (allowing the bee to return inside the hive after 30 minutes, when the bee did not forage successfully). There was no need to replace the flowers in this experiment because no bee foraged successfully at them.

6.3.2.2 Experiment 2: Using a model dead bee and observing nest-mate

From test Colony A which was placed in Cage I, individually marked worker bees were segregated into two groups of bees (A1 with 12 bees and A2 with 14 bees). One group (A1) was removed from the colony and imprisoned in a mesh tube (20 cm long and 3 cm diameter with

0.4 x 0.4 mm mesh) kept out of sight of the experimental cage. These bees were provided with enough food (syrup and pollen). These bees were to be placed later in the aisle between the two rows of artificial flowers. Group A2 (14 individuals) were prevented from leaving the colony and foraging until testing could be started the next day. Once the A1 bees had been sequestered, bees in group A2 were used for testing one by one. Each bee from Group A2 was released and allowed to forage at the artificial flowers (without dead bees in place) for two trials with 30 minutes of giving up time. None of them were successful to forage. After two trials of giving up, a model (dead bee) was introduced. At this point, newly dead bees were placed on the artificial flowers with their heads at the access hole. The dead bees came from the same colony (A) and had been killed by freezing at -18oC one day before the experiment and allowed to thaw and warm to ambient air temperature for 3 hours before the experiment started. Each bee from Group

A2 was released and allowed to forage at the artificial flowers with dead bees in place. After each bee from group A2 had made three successful foraging visits to any one of the artificial

135 flowers, dead bees in place, I replaced the used artificial flowers with cleaned ones that did not have dead bees in place. That avoided the possibility that pheromone signals could influence the results. The visits of each of the A2 bees to the artificial flowers with (3 trials for each of 10 bees) and without dead bees (7 trials for each of the same 10 bees) were observed and timed for a total of 10 foraging bouts, access time was measured by using a stop watch, and the time started when the subject bee entered the testing arena and stopped when the subject bee started to probe for the syrup.

In the follow-up experiment, the cohort of 10 bees from the same colony (A1 bees) that had been imprisoned were placed in a mesh tube size (as described above) between the array of artificial flowers so that they could watch the successful experienced foragers (A2 bees) noted above. The A1 bees had the opportunity to watch the A2 bees at work for 10 daylight hours and were not allowed to return home until the next morning (at 8:00 am.), so preventing them from having communication with their nest mates (except for watching during the day), for 20 hours.

The A1 bees, upon release in the morning, voluntarily and immediately returned to their hive but within 5 minutes started to re-emerge from the domicile. They were then allowed to forage singly at the experimental array of new and clean artificial flowers without the experienced A2 bees present. The visits of each of these A1 bees to the artificial flowers were observed and timed for a total of 10 trials.

6.3.2.3 Experiment 3: Observing none nest-mate and communication possibilities

In a tandem experiment to test if the bees could communicate within the hive how to forage on the artificial flowers I used a completely different colony (B) which was placed in

Cage II. In colony B, I segregated two cohorts of 12 sister or half-sister worker bees each of individually marked bees (as above). The workers in Colony B (Cage II) were allowed to forage

136 freely from 8 centrifuge tubes not provided with artificial floral discs or holes in the walls. One cohort (B1) was later imprisoned in a mesh tube, these bees were provided with enough food ( syrup and pollen), (as described above) and transported to Cage I where they were placed in the array of artificial flowers (as described above for A1 bees) and allowed to watch foragers from colony A forage for 10 hours. The same protocol as for A1 bees was used to treat the imprisoned workers from Colony B, except that the mesh tube prison and its inmates were removed from

Cage I for the night to the bench supporting the cage. In the morning the “prison” and its inmates of B1 bees were returned to Cage II, where the inmates were released. As with the A1 bees as described above, the B1 bees voluntarily and immediately returned to their hive but within 5 minutes started to re-emerge from the domicile. At the same time, the second cohort

(B2) was allowed to forage freely at plain centrifuge tubes. Thus, the B2 bees had no opportunity to come into contact, or to see, the artificial flowers with the holes in the centrifuge walls. Group

B2 were prevented from leaving the colony and foraging until testing them after their nest- mates

B1 finished testing.

After the B1 bees had been returned to their home cage in the morning after being imprisoned in the mesh tube overnight, and had re-entered their home domicile, bees from both cohorts started to exit from their domicile but were denied access to the main cage. At this time, an array of 8 artificial flowers (with discs and holes in the walls) was placed into cage II. Then, only B1 bees were allowed to forage individually at that array and the B2 bees were denied entry into the main part of the cage. The B1 bees were each allowed to forage from the artificial flowers (newly cleaned for each trial and each bee) three times. After that, they were allowed to forage at the flowers 7 more times. Thus, the B2 bees still had had no opportunity to come into contact, or to see, the artificial flowers with the holes in the centrifuge tube walls, but they had

137 contact with experienced nest-mates, the B1 bees. The next day, B2 bees were allowed to forage at newly cleaned artificial flowers in the standard array. These bees were observed for 3 trials, followed by another 7 (as described above) and the durations (i.e. the time it took the bees to manipulate the flower) of the foraging bouts/trials were recorded. At this time, all bees from the first cohort (B1) were prevented from entering the main cage. At no time during the experiment were bees of both cohorts allowed to forage at the artificial flowers at the same time.

6.3.3 Data Analyses:

To compare between groups and trials, I used one way repeated measurement (using

Sigma Plot Statistic v12.0) and to isolate the group or groups that differed from the others I used a multiple comparison procedure. The duration for the manipulation of the artificial flowers on the first visit by foragers was used for inter-experimental comparisons both within and between colonies (groups). Level of significance performed test with alpha = 0.5. I used Multiple

Comparison Procedures (Holm-Sidak method): all pair-wise, overall significance level = 0.5.

For comparison between the two groups of learning through observation, I used t-test.

6.4 RESULTS

Bees from the control colonies (C and D) in Cage III and IV had no opportunity for social learning and all subject bees, which were observed proved incapable of foraging successfully at the artificial flowers(15 from Colony C and 10 from Colony D), with 30 minutes trial.

When the bees of group A2 (Colony A (in Cage I)) were challenged by presenting the artificial flowers without dead bees in place for two trials with 30 minutes of giving up time, none of them were successful to forage. 10 out of 14 (the rest gave up and did not show up for more testing) of the group (A2) were able to see dead bees at all 8 flowers as they foraged freely from their colony. When they foraged, they did so by climbing the artificial stem (pipette tube)

138 of the flower and, positioning themselves beside the dead bee under the disc and took syrup.

These bees were not at first fully adept at foraging beside the dead bees, but after about 3 trials they became adept at the task (Fig. 6.3). After having had that experience, and when the dead bee was absent, those same experienced bees foraged successfully from new and cleaned artificial flowers. However, they did not require familiarization with the dead bee-less artificial flowers and were fully adept on their first visit (Fig. 6.3).

In the follow-up experiment, a cohort of 10 different bees from the same colony (A1) that had been imprisoned in the mesh tube were placed between the array of artificial flowers so that they could watch successful experienced foragers for a day (the A2 bees) at first typically landed on the upper surface of the coloured disc of the artificial flower, then crawled under and down to access the reinforcer syrup through the holes in the sides of the centrifuge tubes. After about 3 visits, these A1 bees flew directly to the openings on the sides of the centrifuge tubes to forage

(Fig. 6.4).

To assess the importance of watching active foragers vs. the presence of the dead-bee model, I compared the time it took for the bees to manipulate (i.e. to land on the flowers, orient to their correct positions to forage and then to imbibe syrup) the flowers on their first visit (cf.

Figures 6.4 and 6.3). The difference in time is huge: with the model dead bees the initial visit to succeed at obtaining the reinforcer syrup was 900 +/- 174.8 secs (mean +/- SE; n = 10 bees), whereas after watching, the bees took only 69 +/- 17.4 secs (n = 9 bees) to forage successfully

(Student’s t = 4.53 df = 17; p = 0.0003.

Following experiments on the first colonies (Colonies A, C and D), I introduced to the experimental set-up, another colony (Colony B) in another cage (II).

139

The watcher worker bees from colony B, cohort 1 (B1 bees) showed the same behaviour as A1 bees (from colony A) when challenged with the artificial flowers (see Figures 6.4 and 6.5).

To assess the importance of watching active nest-mate foragers vs. non-nest-mate foragers, I compared the time it took for the bees to manipulate the flowers on their first visit (cf.

Figures 6.4 and 6.5). There is no statistical difference in times that either group to succeed at obtaining the reinforcer syrup was 69 +/- 17.42 secs (mean =/- SE; n = 9 bees) after watching nest-mates vs. 78 +/- 27.14 secs (n = 8 bees) after watching non-nest-mates (Student’s t = 0.3 df=15; p = 0.77), this test is important because after watching they return to the colony for five minutes and to avoid nest-mate contact.

The bees in Cohort B2 had no chance to see artificial flowers with or without dead bees in foraging positions, nor to observe their nest-mates or non-nest-mates foraging at the artificial flowers. Cohort B2 bees had only opportunity to communicate with their nest-mates while their nest-mates were foraging, with exclusive access, to the artificial flowers. Figure 6.6 presents the surprising results that B2 bees had somehow learned how to forage from the artificial flowers.

To assess the importance of communicating with active nest-mate foragers vs. no communication and vs. learning by observing a model (dead bees) or active foragers (nest-mates or not), I compared the time it took for the bees to manipulate the flowers on their first visit (cf.

Figures 6.3, 6.4, 6.5, 6.6). The bees that had the opportunity for in-nest communication only before foraging took longer time than the bees that had watched either nest-mates (Fig. 6.5) or non-nest-mates (Fig. 6.4) forage. However they were quicker than the bees that had learned by having only the dead bee models in place (Fig. 6.3). Statistical analysis by ANOVA supports those observations (F3,9 = 2.3 ; p = 0.046); durations to successfully obtaining the reinforcer syrup on the first experimental encounter rank in the following order watcher of nest mates (69

140 secs; Fig. 6.4) = watcher of non-nest-mates (78 secs; Fig. 6.5) < communicators (195 secs; Fig.

6.6) < observers of dead bees (906 secs; Fig. 6.3) < no clues provided (all 15 bees unsuccessful;

∞ secs).

I provide the detailed statistical tables for the results of one-way repeated measures for

ANOVA (Table 6.1).

141

1100 1000 900

800

700 600 500

400 Mean time/ time/ MeanSecs 300 200 100 0 1 2 3 4 5 6 7 8 9 10 Trial number

Figure 6.3: The learning curve (time/ sec) ( +/- SE) taken to access and forage on syrup) for 10 initially naive workers of Bombus impatiens allowed to forage freely, but only one at a time, at artificial flowers with and without dead bees present. After the bees had demonstrated their ability to forage at the flowers with dead bees present (i.e. after 3 trails), those flowers were replaced with cleaned ones without dead bees present. The activities of the foragers were recorded for a further 7 visits. H0 that the durations for successful foraging are independent of experience is rejected (F9,9 = 19.7; p < 0.001)

142

Mean time/ time/ MeanSecs 30

0 1 2 3 4 5 6 7 8 9 10 Trial

Figure 6.4: The learning curve (time/ sec) ( +/- SE) taken to access and forage on syrup) for 9 workers of Bombus impatiens allowed to watch experienced foragers at artificial flowers for 10 hours and held incommunicado overnight. In the morning these bees demonstrated their ability to forage at the flowers after 3 trials. The flowers were replaced with cleaned ones after each of the first three trials and for each individual bee tested. The activities of the foragers were recorded for a further 7 visits. H0 that the durations for successful foraging are independent of experience of having watched nest-mates forage is rejected (F8,9 = 10.7; p < 0.001).

143

110 100 90

70 60 50

40 Mean time/Secs Mean 30 20 10 0 1 2 3 4 5 6 7 8 9 10 Trial

Figure 6.5: The learning curve (time/ sec) ( +/- SE) taken to access and forage on syrup) for 9 workers of Bombus impatiens allowed to watch experienced foragers ( from different colony (Colony A))at artificial flowers for 10 hours and held incommunicado overnight. In the morning these bees demonstrated their ability to forage at the flowers after 3 trails. The flowers were replaced with cleaned ones after each of the first three trials and for each individual bee tested.

The activities of the foragers were recorded for a further 7 visits. H0 that the durations for successful foraging are independent of experience of having watched other, non-nest-mate, bees forage is rejected (F8,9 = 7.7; p < 0.001)

144

250

200

150

100 Mean time/ time/ MeanSecs

0 1 2 3 4 5 6 7 8 9 10 Trial

Figure 6.6: The learning curve (time/ sec) (+/- SE) taken to access and forage on syrup) for 9 workers of Bombus impatiens allowed to contact with their nest-mates B1 (i.e. were watching the experienced bees from Colony A). The B2 bees were kept inside the hive and then, after release to forage, had apparently learned to manipulate the artificial flowers through communication with their nest-mates. H0 that the durations for successful foraging are independent of experience of having communicated with their nest-mates is rejected (F8,9= 21.4; p < 0.001)

145

Table 6.1: Statistical values from repeated one-way Analysis of Variance of the findings from experiments in which dead bees were used as models to aid in the learning process for foraging by living bee, in which living bees were able to watch other living bees (nest-mates and non- nest-mates) forage to aid in the learning process, and in which living bees which had no opportunity to observe models or other living bees foraging learned to forage by within-colony communication.

Source of Variation Degrees of freedom Mean Squares F value Probability

Using dead bees in the foraging position on the artificial flowers (Figure 6.3) Between Bees 9 110263

Between Trials 9 80380 19.68 <0.001

Residual 81 40824

Total 99

Watching nest-mates (Figure 6.4) Between Bees 8 678.144

Between Trials 9 3427.832 10.74 <0.001

Residual 72 318.910

Total 89

Watching non-nest-mates (Figure 6.5) Between Bees 8 493.550

Between Trials 9 4223.784 7.69 <0.001

146

Residual 72 549.001

Total 89

Communication within the domicile (Figure 6.6) Between Bees 8 2577.375

Between Trials 9 31795 21.37 <0.001

Residual 72 1487.610

Total 89

The difference between four groups of tested bees. Between learning type 3 35003

Between Trial 9 36803 2.30 0.046

Residual 27 15990

Total 39

147

6.5 DISCUSSION

When naive bumble bee workers first encounter a flower from which they can obtain a reward (e.g. nectar or pollen), they must learn how to manipulate it. Laverty (1994) has shown that bumblebee workers (Bombus impatiens, B. fervidus, B. vagans, B. rufocinctus, B. consobrinus) become increasingly adept (i.e. by speed and accuracy of manipulation) with increasing experience. Moreover, Dornhaus and Chittka (2004, 2005) noted that returning foragers stimulated colony-level foraging activity. Baude et al. (2008) described the inter-colony facilitation in foraging by B. terrestris as the use of inadvertent social information (ISI) whereby foragers watched each other’s activities and learned from that. Leadbeater and Chittka (2009) showed that worker bumble bees (B. terrestris) learned to discriminate between two kinds of flowers depending on whether or not they contained nectar faster if conspecific foragers were present and foraging at the same time than if they were alone. They also noted that if dead bumble bee models were present in posed foraging positions, the effect was the same: the experimental bees learned faster than if no dead bee was present. Their results indicate that social learning at flowers can be a component of foraging efficiency. More recently Leadbeater &

Chittka (2009) have shown that nectar robbing can spread socially among bumble bees foraging at horizontally oriented tubular artificial flowers (Leadbeater and Chittka 2008; Sherry 2008)

[probably by watching other bees and encountering holes already made in the flowers].

Leadbeater and Chittka (2009) state that social learning within the nest is unlikely; however, these results indicate the possibilities of transmissions of the information inside the nest. It is possible that measures of rates of learning (e.g. Figs 6.3-6.6 in this study) also indicate effects of stimulation by experience or the presence of other foragers. Even though these results indicate that those bees that watched living foragers (i.e. nest-mate or no nest- mate) learned faster than

148 which could see dead bees (cf Figs 6.3, 6.4 and 6.5). I raise the idea that the difference could reflect that stimulation accelerates social learning. I also noted that Worden and Papaj (2005) used a stationary and moving model bees and found quicker responses of trained forager bees to the latter. It is also known that bumble bees, as other bees, communicate socially through pheromones (Wyatt 2003) and can discriminate between recently visited flowers and flowers which have not been visited from some time (Cameron 1981; Goulson et al. 1998; Wilms and

Eltz 2008). Renner and Nieh (2008) showed that foragers of B. impatiens can associate scentedness of rewarding food sources (flowers) and share this ability with their nest-mates. The same phenomenon has also been shown for other species, e.g. B. terrestris (Schmitt and Bertsch

1990; Schmitt et al. 1991; Goulson et al. 2000; Witjes and Eltz 2007).

Physical contact, especially antennal and body contact may be important in the transmission of information on the location, quality, quantity, and nature of floral resources in honeybees (Rohrseitz and Tautz 1999), and stingless bees (Hrncir et al. 2000; Schmidtet al.

2006). However, little is known about the role of physical contact in the lives of bumblebees.

Food exchange (trophallaxis) may be the most primeval form of social communication in eusocial bees, but not bumble bees (de Marco and Farina 2003; Hart and Ratnieks 2002) and may provide information about food quality and odour for some species. Bumble bees may be able to gain such information by sampling resources (nectar and pollen) once deposited in the colony. Observation and social learning strengthen a colony’s foraging efficiency both by intake of more resources by the same colony or by promoting a competitive stratagem by learning from rival colonies (Baude et al. 2011).

These experiments were designed to extend my understanding of the potential for social learning in bumblebees following from the work of Worden and Papaj (2005), Kawaguchi et al.

149

(2006) and Leadbeater and Chittka (2005, 2011). I controlled for external cues, such as scentedness of or pheromone residues on the artificial flowers (cleaned as used) and reinforcer syrup (sucrose in water has no vapour pressure and was always made fresh for each experiment).

The domiciles used were always in the same locations relative to the arrays of artificial flowers.

The visual signals were highly controlled such as to the colour of the artificial flowers, the dead bees were posed on them, and the active foragers that imprisoned bees could watch.

150

CHAPTER 7

CONCLUSION

It is fundamental to understand that studying and knowing the critical factors affecting pollinators in visiting flowers, and floral relation to the cognitive behaviours of pollinators is relevant to understanding their ecology.

Pollinators can learn to manipulate structures (natural or artificial flowers) to push forward directly to a reinforcer of nectar or syrup. In doing this, bumblebees were shaped to accomplish more difficult tasks by making these tasks unlike anything encountered in nature.

This required the bumblebees to manipulate with increasingly complex single or multiple tasks, deviating from a direct approach and moving barriers sideways by two to five body lengths before returning to the now accessible reinforcer. Some tasks required their deviating from, and then returning to, the very spot that they first had to visit to start to move the barriers aside. Thus in Chapter Two, in operant tasks (coupled with manipulative problem solving requiring the bumblebees to deviate from a direct approach and move barriers sideways by two to five body lengths before returning to the now accessible reinforcer), bumblebees successfully moved variously sized combinations of caps aside, and also rotated discs through varying arcs (up to

180°) from the entrances of artificial flowers to access the reinforcer hidden beneath. These experiments have demonstrated a kind of detour behaviour by worker bumblebees in that they learned how to reach the goal even when they had no clear view of it. They did not move around interposed obstacles (as per Zucca et al. 2005), but instead moved them aside. Additionally, the experiments showed that the bumblebees were rapidly able to associate the directionality of rotating discs with colour cues (white versus yellow) and did so with great consistency.

151

Furthermore, to be able to rule out the unlikely possibility that subjects were merely pushing towards the odour of sugar (sucrose has no vapour pressure (Cameron 1981)) by demonstrating that pre-trained subjects rotated discs in the previously corrected direction even in extinction (i.e. even when sucrose was absent). The results of these tasks achieved under controlled conditions rather than in the wild, demonstrate the value of giving bumblebees additional cognitive tests of increasing complexity.

One question remaining is: are they able to figure out and perform “detour tasks” without being influenced by successive approximations? Such tasks require that the bee would need to move away from a reinforcer in order to gain access to it, a decision making process that animals belonging to some taxonomic groups or possessing certain brain lesions would find quite difficult (e.g. Wells 1964; Collet et al. 1993; Regolin et al. 1995; Zucca et al. 2005). Additional future research could investigate whether, in cases where true detour behaviour is revealed, whether the bees demonstrate the detour planning capability demonstrated by certain types of spiders (Portia labiata; Aranea: Salticidea; Tarsitano and Jackson 1994; Tarsitano 2006)?

Furthermore, can they learn the properties of physical objects that would allow them to master these manipulanda in such a way as to obtain sugar in a single attempt (cf. literature on tool-use in birds (D'Amato and Colombo 1988; D'Amato 1991; Pearce 2008, Shumaker 2011; Snaz

2013).

Leading from the results of Chapter Two, more questions raised about bumble bees’ abilities to problem solve were then tackled. Bumblebees overcome such challenges through experience and problem solving, presumably an outcome of shaping and scaffold learning.

Thus, increasing speed and accuracy in handling a given complex task through shaping and scaffolding indicate that experience and memory are important components of worker

152 bumblebees’ learning to manipulate the objects that they encounter in their natural lives.

Even small changes in floral complexity have been shown by other researchers to have a large effect on the bees’ learning to manipulate them (Laverty 1978). These experiments, in my third chapter, reached the same conclusion, increasing the complexity in serial order, and assessed inexperienced bees for their ability to manipulate complex flowers at the most difficult level of the intricacy (i.e. the inverted single cap could be lifted or the heaviest ball).

The experiment in Chapter Three thus showed that the abilities of the bees to manipulate flowers’ complexity increased the speed and the accuracy in handling complex tasks, and explained their abilities to solve problems in naturally complex circumstances. In the experiment where the bees had to remove increasingly massive balls that occluded access to the reinforcer (Chapter Three), the task is the same but becomes more difficult as the balls’ masses increased. This presented an example of shaping where animals can be trained to perform tasks of increasing difficulty. These ranged from rewarding naïve animals’ positive responses to simple tasks to more and more involved responses in more and more difficult tasks.

The concept of scaffold learning in animal behaviour was brought about in the experiment the first experiment of Chapter Three with the bees sliding and lifting caps to access the reinforcer sugar syrup. This could be considered scaffolding in which the experimenter coached the bees. The tasks presented involved the complexity of moving caps to the left or right or above to allow access to the reinforcer. In addition, the theory of the

Zone of Proximal Development (ZPD) was introduced as a means of quantifying learning rates. The subject may spend longer or give up when first given overly complex tasks they

153 can solve well, provided the tasks are presented incrementally. The ZPD values indicate that skill acquisition in bumblebees can be measured as it is for human learning (Bransford et al. 2000). Experience provide bumblebees with advantages in both qualitative (time) and quantitative (strategy) modes, Thus, through experience, worker bumblebees become able to solve new problems they encounter rather just giving up, unlike bees that have had no previous experience.

Following the solving of a single problem, sequential-multiple problem solving experiments were conducted. Chapter Four went beyond even quite sophisticated versions of associative sequence learning (Spiegel and McLaren 2006) and numeracy, because I have linked the relationship between pattern recognition possibly by subtizing or counting objects and accomplishments as well as chaining or linking theory (e.g., in which each response becomes the stimulus for the next). The latter involved three rather different tasks and obstacles that needed skills of navigation and manipulation skills (sliding, lifting, and rotating), and how they related to problem solving, navigation and orientation, and cognition.

Those sequences correspond with sequence recognition and sequential decision making through several actions (goal-oriented, trajectory-oriented, and reinforcement-maximizing), all of which lead to the experiment’s goal: obtaining sugar syrup and returning home at the end (Sun

2001). Moreover, when presenting to bumblebee workers a series of tasks which may demonstrate that they organize learned behaviours hierarchically into behavioural response chains with goals and sub-goals (Byrne & Byrne 1993; Byrne & Russon 1998). After encountering unrewarded flower patches at the first and second compartments, and passing through the sliding door box, the bees had a choice: to return to their hive or proceed to the next compartment and feed at the floral array, lifting the occluding caps to forage before returning

154 home. Clearly, their priority was to proceed and forage. Pattern recognition and solving complex problems are part of the foraging strategies of bees. The results obtained from these experiments show that worker bumblebees have the capacity to solve different complex problems in a particular order that possibly involves their knowing where they are in a sequence of tasks

(possibly by counting and prioritizing the tasks) and being able to recognize a numerical array of objects as they perform some tasks. In contrast to the results obtained by Gegear and Laverty

(1998), these experiments demonstrated that bees’ abilities to master complex tasks were less constrained when the tasks were presented sequentially. The insects’ floral fidelity/constancy was constrained when required to perform tasks within a mixed patch of flowers, a drawback not evident when the bees were engaged in single foraging trips or moving between more than one flower patch. The importance of performing these intricate tasks in sequence may be one of the key factors in the bees’ ability to sample different flowers while foraging, and could help to clarify how they “major” and “minor” (Heinrich, 1979b) in tracking the evolving resources available. More experiments using the traditional training procedure of chain learning could prove bumblebees’ abilities to learn sequential tasks.

Moving on from the experiments on associative and problem solving to maze navigation in Chapter Five, with the absence or the minimum of visual and chemical cues, here, bumblebees learned to negotiate through complex mazes with multiple turns and several dead ends, like ones used with rats (Honzik 1936), by walking using a fixed motor program possibly linked to some form of path integration. These were more complex than the bifurcation mazes (e.g. as used by

Chittka 1998; Chittka and Thomson 1996; 1997) used to study cue discrimination (choice) and memory. Bumblebees could remember the routes through the mazes, although they were prevented from going through it, for up to16 days, which is the life time of scouts or foragers out

155 of their hive, although their performance in the maze deteriorated the longer they had been absent from it. This suggests that the bees invoked some sort of memory cascade; after making a correct initial choice, they were able to recall more accurately subsequent correct choices (see also Chittka 1998; Chittka and Thomson 1997).

Although bumblebees forage by flying to and from their nests, they perform many tasks by walking. Thus, although multi-turn maze learning and ambulatory navigation (such as I have used) may seem unnatural, it may be important in bumblebees’ lives. It is well known that bumblebees build their nests in piles of stones, underground in cavities, and in buildings where they must walk and navigate through topographically simple to complex spaces. The practical application of ambulatory navigation is exemplified in the technology that uses foraging bumblebees to deliver biological control agents against crop pests and pathogens from special walk-through dispensers on the hives (Kevan et al. 2008; Kevan et al. 2014). Finally, it seems that memory retrieval in workers of B. impatiens appears to work in a manner similar to that in humans and other vertebrates that have been tested (Anderson 2000; Schacter 2001). Although my experiments on maze navigation tested the bees’ abilities of navigation for first time, and their abilities to remember the route after having been absent from using it, additional experiments comparing colonies and nest-mates would be instructive. For instance, testing the abilities of nest-mates from the same colony for two or three different periods of being absent from passing through the maze could be useful for studying memory and memory deterioration.

I then moved away from associative, problem solving, and navigation. I wished to extend my understanding of the potential for social learning in bumblebees. The experiments in Chapter

Six were therefore designed to investigate information transmission following from the work of

Worden and Papaj (2005), Kawaguchi et al. (2006) and Leadbeater and Chittka (2007, 2008).

156

The results indicated that workers of B. impatiens are highly observant and learn through social communication. Although they were relatively slow to learn to forage from artificial flowers with dead conspecifics posed as foragers on them, they were much faster if they had the opportunity to observe, but not join, active foragers from either their own colony or from another. Surprisingly, when I allowed workers that had never had a chance to visit nor see an artificial flower, but had had contact with nest-mates that were successful foragers, the experimental (naive) workers were adept at handling the artificial flowers. All workers that were confronted with the artificial flowers but no opportunity to see posed dead bees, active foragers, or communicate within the colony, failed to forage successfully. It would be useful for other researchers to repeat the experiment, with appropriate modifications, to test if these results can be repeated or explained. It is often assumed that observational learning and imitation (or copying), lies at the heart of social transmission of information and learning (Brian 1957;

Leadbeater & Chittka 2005), but there are other ways novel behaviour can be transmitted socially (e.g. through tactile, vibratory, and olfactory senses), especially in bees. The mechanism of how workers that had never had a chance to visit nor even see an artificial flower, and only had contact with nest-mates that were successful foragers, became so quickly adept at handling the artificial flowers, are not known and could not be explained. They are an exciting area for future research. These results also indicated that social learning at flowers can be a component of foraging efficiency.

Evidence, including that which presented herein, continues to mount showing that there is no strict dichotomy between vertebrate and invertebrate cognition (Fiorito & Scotto 1992; Giurfa et al. 2001; Srinivasan & Zhang 2003). This work adds to the growing body of research in social

Hymenoptera which demonstrates that brain size does not necessarily limit an animal’s cognitive

157 abilities. More imaginative experiments are needed to determine the role of social learning, the amount and type of information that need to be transmitted, and how that body of information contributes to Darwinian fitness. Moreover, some behaviour has always been difficult to explain and cannot be clarified by a single cause. In addition to repeating those experiments that tested social learning by bumblebees, refinement of the methods could be used to elucidate some mechanism that the bees used for information transmission.

The evolution of learning, intelligence and brain function has presumably progressed from simple reflex and unconditioned behaviours through Pavlovian conditioning to operant learning. The results of these experiments indicate an unexpected capacity for learning and individualism in bumblebees in that the subjects learned to manipulate items that seem far from any natural objects or ‘contrivances’ (Darwin 1877) that they would encounter in the nature.

Adaptability, whether through behavioural plasticity, physiological adaptation or evolutionary change, is crucial to species’ survival. I suggest that the complex learning capacity demonstrated in bumblebees reflects hitherto unsuspected natural capacities, perhaps akin to those observed in vertebrates, to assemble sophisticated sequences of behaviour in response to novel challenges.

Finally, as additional evidence that the bumblebees were successfully learning a deliberated task, rather than merely modifying an unconditioned approach response, bumblebees developed a variable range of idiosyncratic behaviours to solve their problems. They also seemed to reduce the times it took them to accomplish the task. The bumblebee, as a model of conditioning learning, may be useful in understanding general and evolutionary principles of learning and cognition. Further experiments are required to quantify and analyse the kinds of different behaviours I noted. Other more refined researches using these studies’ techniques and concepts may elucidate such issues as age dependent effects (age of the workers),

158 individual differences in bees’ learning, and whether long term memory is affected differently by lack of use for varying periods. Overall, despite their small brains, I found that bumblebees can become skilled at rotational and lateral manipulations in moving objects from sources of food.

159

REFERENCES

Aguilar, I., and Briceño, D. 2002. Sounds in Melipona costaricensis (Apidae: Meliponini): effect of sugar concentration and nectar source distance. Apidologie 33: 375-388

Alston, D. G., Tepedino, V. J., Bradley, B. A., Toler, T. R., Griswold, T. L., and Messinger, S. M. 2007. Effects of the insecticide phosmet on solitary bee foraging and nesting in orchards of Capitol Reef National Park, Utah. Environmental Entomology 36: 811-816

Amaya- Marquez, M. 2009. Memory and learning in bees' floral choices (Memoria y aprendizaje en la escogencia floral de las abejas). Acta Biol Colombiana 14: 125-135

Amaya-Marquez, M., Hill, P. S. M., Barthell, J. F., Pham, L. L., Doty, D. R., and Wells, H. 2008. Learning and memory during foraging of the blue orchard bee, Osmia lignaria Say (Hymenoptera: Megachilidae). Journal of the Kansas Entomological Society 81: 315-327

Anderson, J. R. 2000. Learning and Memory an Integrated Approach (Second Edition). Carnegie Mellon University, John Wiley and Sons, Inc

Baerends, G. P. 1941. Fortpflanzungsverhalten und Orientierung der Grabwespe Ammophila campestris Jur. Zietschrifte für Entomologie 84: 81-275

Balleine, B. W., Garner, C., Gonzalez, F., and Dickinson, A. 1995. Motivational Control of Heterogeneous Instrumental Chains. Journal of Experimental Psychology: Animal Behavior Processes 21: 203-217

Bandura, A. 1977. Social Learning Theory. General Learning Press, New York, NY, USA.

Bandura, A. 1986. Social Foundations of Thought and Action. Prentice-Hall, Englewood Cliffs, NJ, USA.

Bar-Shai, N., Keasar, T., and Shmida, A. 2011. How do solitary bees forage in patches with a fixed number of food items? Animal Behaviour 82: 1367-1372

Bar-Shai, N., Keasar, T., and Shmida, A. 2011. The use of numerical information by bees in foraging tasks. Behavioral Ecology 22: 317-325

Barth, F. G. 1985. Insects and flowers: The biology of a partnership. Princeton University Press, Princeton, NJ.

Barth, M., and Heisenberg, M. 1997. Vision affects mushroom bodies and central complex in Drosophila melanogaster.Learning & Memory. 4:219-229

160

Baude, M., Dajoz, I., and Danchin, E. 2008. Inadvertent social information in foraging bumblebees: effects of flower distribution and implications for pollination. Animal Behaviour 76:1863-1873

Baude, M., Danchin, E., Mugabo, M., and Dajoz, I. 2011. Conspecifics as informers and competitors: an experimental study in foraging bumble-bees. Proceedings of the Royal Society B- Biological Sciences 278: 2806-2813

Belew, R. K. and M. Mitchell. 1996. Introduction. In: Adaptive Individuals in Evolving Populations: Models and Algorithms. Proceedings of the Santa Fe Institute. Volume XXVI. Belew, R. K. and M. Mitchell (Eds.) pp 1-22. Addison-Wesley Publishing Company, Inc. Boston, Massachusetts.

Bell, I. R., Hardin, E. E., Baldwin, C. M., and Schwartz, G. E. 1995. "Increased limbic system symptomatology and sensitizability of young adults with chemical and noise sensitivities". Environ Res 70: 84-97, PMID 8674484, doi:10.1006/enrs.1995.1052

Bernstein, S., and Bernstein, R. A. 1969. Relationship between foraging efficiency and the size of the head and component. Brain Research 16: 85-104

Berstein, D., and Nash, P.W. 2008 Essentials of Psychology. Cengage Learning, US 544 pp.

Bicker, G., and Spatz, H. C. 1976. Maze-learning ability of Drosophila melanogaster. Nature 260: 371

Bierley, R. A., Rixen, G. J., and Tröster, A. I. 1986. Preserved spatial memory in old rats survives 10 months without training. Behavioural and Neural Biology 45: 223-229 Bitterman, M. E. 1984. Migration and learning in fishes. In I. D. McCleave., G. P. Arnold., I. I. Dodson., and W. H. Neill (Eds.), Mechanisms of migration in fishes (pp. 397-420). New York: Plenum

Bitterman, M. E., Menzel, R., Fietz, A., and Schäfer, S. 1983. Classical Conditioning of Proboscis Extension in Honeybees (Apis mellifera). J. Comp. Psych. 97: 107-119

Born, D. G., Snow, M. E., and Herbert, E. W.1969. Conditional discrimination learning in the pigeon. Journal of the Experimental Analysis of Behavior 12: 119-125

Bouton, M. E. 2007. Learning and Behavior: A Contemporary Synthesis, Sunderland, MA: Sinauer.

Braddick, O., and Qian, N. 2000. The organization of global motion and transparency. J.M Zanker, J Zeil (Eds.), Computational, neural and ecological constraints of visual motion, Springer, Berlin.

161

Braddick, O., Atkinson, J., and Wattam-Bell, J . 2003. Normal and anomalous development of visual motion processing: Motion coherence and ‘dorsal stream vulnerability. Neuropsychologia 13: 1769-1784

Bransford, J., Brown, A., and Cocking, R. 2000. How people learn: Mind, brain, experience and school. Washington, DC: National Academy Press.

Brembs, B. 2000. The Operant and the Classical in Conditioned Orientation of Drosophila melanogaster at the Flight Simulator. Learning and Memory 7: 104-115

Brembs, B. 2003. Operant conditioning in invertebrates. Current opinion in neurobiology 13: 710-717

Brian, A. D. 1957. Differences in the flowers visited by four species of bumble-bees and their causes. Journal of Animal Ecology 26: 71-98

Byrne, R., and Byrne, A. 1993. Complex leaf-gathering skills of mountain gorillas (Gorilla g. berengei ): variability and standardisation. American Journal of Primatology 31, 521-546

Byrne, R., and Russon, A. 1998. Learning by imitation: a hierarchical approach. Behaviour and Brain Science 21: 667-721. (doi:10.1017/S0140525X98001745)

Cameron, S. A. 1981. Chemical signal in bumble bee foraging. Behavioural Ecology and Sociobiology 9: 257-260

Cameron, S. A. 1985. Brood care by male bumble bees. Proceedings of the National Academy of Sciences of the USA 82: 6371 – 6373

Cameron, S. A., and Whitfield, J. B. 1996. Use of walking trails by bees. Nature 379: 125

Cammaerts , M. C., Rachidi, Z,. Beke, S., and Essaadi, Y. 2012. Use of olfactory and visual cues for orientation by the ant Myrmica ruginodis (Hymenoptera: Formicidae). Myrmecological News 16: 45-55

Carter, D. E., and Werner, T. J. 1978. Complex learning and information processing by pigeons: A critical analysis. Journal of the Experimental Analysis of Behavior 29: 565-601

Cartwright, B.A., and Collett, T.S. 1987. Landmark maps for honeybees. Biological Cybernetics 57: 85-93

Chameron, S., Schatz, B., Pastergue-Ruiz, I., Beugnon, G., and Collett, TS. 1998. The learning of a sequence of visual patterns by the ant Cataglyphis cursor. Proceedings of the Royal Society of London, B 265: 2309-2313

Chittka L. 1998. Sensorimotor learning in bumblebees: Long-term retention and reversal training. Journal of Experimental Biology. 201: 515-524

162

Chittka, L., and Geiger, K. 1995. Can honeybees count landmarks? Animal Behaviour 49:159- 164

Chittka, L., and Thomson, J. D. 1996. The ecology of bumble bees in T- mazes. In: G?ttingen Neurobiology Report. Elsner N, Schnitzler H. (eds). Thieme, Stuttgart, p 130

Chittka, L., and Thomson, J. D. 1997. Sensori-motor learning and its relevance for task specialization in bumble bees. Behavioral Ecology and Sociobiology 41: 385-398

Chittka, L., and Thomson, J. D. 2001. Cognitive Ecology of Pollination. Animal Behavior and Floral Evolution. Cambridge University Press. 344p

Chittka, L., Geiger, K., and Kunze, J. 1995. The influence of landmarks on distance estimation of honeybees. Animal Behaviour 50: 23-31

Chmurzynski, J. A. 1964. Studies on the stages of spatial orientation in female Bembex rostrata (Linne 1758) returning to their nests (Hymenoptera, Sphegidae). Acta Biol Exp (Warsaw) 24:103-132

Collett, T. S.1992. Landmark learning and guidance in insects. Philosophical Transactions of the Royal Society of London, Series B. 337: 295-303

Collett, M., and Collett, T. S. 2000. How do insects use path integration for their navigation? Biological Cybernetics. 83: 245-259

Collett, T. S., and Baron, J. 1995. Learnt sensori-motor mappings in honeybees: interpolation and its possible relevance to navigation. Journal of Comparative Physiology, Series A. 177: 287- 298

Collett, T. S., and Zeil, J. 1998. Place and landmarks: An arthropod perspective. In: Spatial Representation in Animals (ed. S. Healy), pp. 18-53. Oxford University Press, Oxford.

Collett, T. S., Dillmann, E., Giger, A., and Wehner, R. 1992. Visual landmarks and route following in desert ants. Journal of Comparative Physiology, Series A. 170: 435-442

Collett, T. S., Fry, S. N., and Wehner, R. 1993. Sequence learning by honeybees. Journal of Comparative Physiology, Series A. 172: 693-706

Collett, T. S., and Graham, P. 2004 Animal Navigation: Path Integration, Visual Landmarks and Cognitive Maps" Current Biology. 14: 475-477 .doi:10.1016/j.cub.2004.06.013

Corbet, S. A., Kerslake, C. J. C., Brown, D., and Morland, N. E. 1984. Can bees select nectar- rich flowers in a patch? Journal of Apicultural Research 23: 234 – 247

163

Couvillon, P. A., and Biiterman, M. E. 1988. Compound-component and conditional discrimination of colors and odors by honeybees: Further tests of a continuity model. Animal Learning and Behavior 16: 67-74

Couvillon, P. A., and Bitterman, M. E. 1989. Reciprocal overshadowing in the discrimination of color–odor compounds by honeybees: Further tests of a continuity model. Animal Learning and Behavior 17: 213-222

Couvillon, P. A., and Bitterman, M. E. 1991. How honeybees make choices. In J. L. Goodman & R. C. Fischer (Eds.), The behaviour and physiology of bees (pp. 116-130). Wallingford, U.K.: CAB International.

Dacke, M., and Srinivasan, M. V. 2008. Evidence for counting in insects. Animal Cognition 11: 683-689

Dale, R. H. I. 1988. Spatial memory in pigeons on a four-arm radial maze. Canadian Journal of Psychology. 42: 78-83

D’Amato, M. R. 1991. Comparative cognition: Processing of serial order and serial pattern. In L. Dachowski and C.F. Flaherty (Eds.), Current topics in animal cognition: Brain, emotion, and cognition (pp. 165–185). Hillsdale, NJ: Erlbaum.

D'Amato, M. R., and Colombo, M. 1988. Representation of serial order in monkeys (Cebus apella). Journal of Experimental Psychology: Animal Behavior Processes 14: 131-139

Darwin, C. 1877. The Various Contrivances by which Orchids are Fertilized by Insects New York: D. Appleton. 2d ed., revised. 300pp.

Darwin, C.1873. "Origin of Certain Instincts". Nature 7: 417-418. doi:10.1038/007417a

Davis, R. L. 1993Mushroom bodies and Drosophila learning. Neuron. 11:1-14

Debelle, J. S., and Heisenberg, M. 1994. Associative odor learning in Drosophila abolished by chemical ablation of mushroom bodies. Science. 263:692-695

Dehaene, S. 1999. The number sense: how the mind creates mathematics. Oxford University Press, Oxford. de Marco, R. J., and Farina, W. M. 2003. Trophallaxis in forager honeybees (Apis mellifera): resource uncertainty enhances begging contacts. Journal of Comparative Physiology A 189: 125- 134

164

Dohl, J. 1968. Uber die Fahigkeit einer Schimpansin, Umwege mit selbststandigen Zwischenzielen zuuberblicken (The ability of a female chimpanzee to overlook intermediate goals). Z. Tierpsychol 25: 89-103

Domjan, M. 2003. The Principles of Learning and Behavior. Sixth Edition, Belmont, CA: Thomson/Wadsworth

Dornhaus A, Cameron S, 2003. A scientific note on food alert in Bombus transversalis. Apidologie 34:87–88

Dornhaus, A., and Chittka, L. 2001. Food alert in bumblebees, Bombus terrestris: possible mechanisms and evolutionary implications. Behav Ecol Sociobiol 50: 570-576

Dornhaus, A., and Chittka, L. 2004. Review article Information flow and regulation of foraging activity in bumble bees (Bombus spp.). Apidologie 35: 183-192

Dornhaus, A.. and Chittka, L. 2005. Bumble bees (Bombus terrestris) store both food and information in honeypots. Behavioral Ecology 16: 661-666

Dornhaus, A., Brockmann, A., and Chittka L, 2003. Bumblebees alert to food with pheromone from tergal glands. J Comp Phys A 189: 47–51

Drickamer, L. C., Vessey, H. S., and Jakob, E. 2002. Animal Behavior: Mechanisms, Ecology, Evolution, 5th Edition. McGraw-Hill Education.

Dukas, R. 1998. Evolutionary ecology of learning. In: Cognitive Ecology: The Evolutionary Ecology of Information Processing and Decision Making. Dukas, R.

Dukas, R. 2009. Learning: Mechanisms, Ecology, and Evolution. In: Cognitive Ecology II, (Ed. by R. Dukas & J. Ratcliff), pp. 7–26. Chicago: University of Chicago Press.

Durst, C., Eichmuller, S., and Menzel, R. 1994. Development and experience lead to increased volume of subcompartments of the honeybee mushroom body. Behavioral and neural biology. 62:259-263

Dyer, F. C. 1991. Bees acquire route-based memories but not cognitive maps in a familiar Landscape. Animal Behaviour. 41: 239-246

Dyer, A. G., Whitney, H. M., Arnold, S. E. J., Glover, B. J., and Chittka, L. 2007. Mutations perturbing petal cell shape and anthocyanin synthesis influence bumblebee perception of Antirrhinum majus flower colour. Arthropod-Plant Interactions. 1: 45-55

165

Dyer, A. G., Spaethe, J., and Prack, S. 2008. Comparative psychophysics of bumblebee and honeybee colour discrimination and object detection. Journal of Comparative Physiology, Series A. 194: 617-627

Dyer, F. C. 2002. The biology of the dance language. Annual Review of Entomology 47: 917- 949

Epstein, R. 1987. The Spontaneous Interconnection of Four Repertoires of Behavior in a Pigeon (Columba livia). Journal of Comparative Psychology 101: 197-201

Epstein, R., Kirshnit, C. E., Lanza, R. P., and Rubin, L. C. 1984. ‘Insight’ in the pigeon: antecedents and determinants of an intelligent performance. Nature 308: 61-62. (doi:10.1038/308061a0)

Erber, J., Masuhr, T., and Menzel, R. 1980. Localization of short-term memory in the brain of the bee, Apis mellifera. Physiological Entomology. 5:343-358

Fabre, J. H. 1915. The hunting wasps. Hodder and Stoughton, Toronto London New York.

Fahrbach, S. E. 1997. Regulation of age polyethism in bees and wasps by juvenile hormone. Advances in the Study of Behavior. 26:285-316

Fahrbach, S. E., Giray, T., and Robinson, G. E. 1995. Volume changes in the mushroom bodies of adult honey bee queens.Neurobiol Learning and Memory. 63:181-191

Fiorito, G., and Scotto, P. 1992. Observational learning in Octopus vulgaris. Science 256: 545- 547

Free, J. B. 1987. Pheromones of social bees. Chapman and Halls. London, UK

Free, J. B., and Butler, C. G. 1959. Bumblebees. Collins, London.UK

Friederici, A. D., and R. Menzel. 1999. Learning: Rule Extraction and Representation. Walter de Gruyter, Berlin, New York. 290 pp.

Frisch, K. von. 1967. The Dance Language and Orientation of Bees, 2nd printing, 1993 Belknap Press, Cambridge, MA, USA.

Fu, Wai- Tat., and Anderson, J. R. 2006. From Recurrent Choice to Skill Learning: A Reinforcement-Learning Model. Journal of Experimental Psychology 135: 184-206

Funayama. E. S., Couvillon, P. A., and Bitterman, M. E. 1995. Compound conditioning in honeybees: Blocking tests of the independence assumption. Animal Learning and Behavior 23: 429-437

166

Galef, B. G., Kennett, D. J. & Wigmore, S. W. 1984. Transfer of information concerning distant foods in rats: a robust phenomenon. Animal Learning and Behavior 12: 292-296

Gegear, R. J., and Laverty, T. M. 1995. Effect of flower complexity on relearning flower- handling skills in bumble bees. Canadian Journal of Zoology 73: 2052-2058

Gegear, R. J., and Laverty, T. M. 2005. Flower constancy in bumblebees: a test of the trait variability hypothesis. Animal Behavior 69: 939-949

Gegear, R.J., and Laverty, T. M. 1998. How many flower types can bumblebees work at the same time? Canadian Journal of Zoology 76:1358-1365

Gelman, R., and Gallistel, C. R. 1978. The Child’s Understanding of Number. Cambridge, MA: Harvard University Press.

Gill, R. J., Ramos-Rodriguez, O., and Raine, N. E. 2012. Combined pesticide exposure severely affects individual- and colony-level traits in bees. Nature 491: 105-108

Giraldeau, L. A. 1984. Group foraging: The skill pool effect and frequency-dependent learning. The The American Naturalist 124:72-79

Giraldeau, L. A., and Caraco, T. 2000. Social Foraging Theory. Princeton University Press, Englewood Cliffs, NJ, USA.

Giraldeau, L. A., Caraco, T., and Valone, T. J. 1994. Social foraging-individual learning and cultural transmission of innovations. Behavioral Ecology 5: 35-43

Giurfa., M. 2003. Cognitive neuroethology: dissecting non-elemental learning in a honeybee brain. Current Opinion in Neurobiology 13:726-735

Giurfa., M. 2012. Social learning in insects: a higher-order capacity? Frontiers in Behavioral Neuroscience 6

Giurfa, M., and Núñez, J. A. 1992. Honeybees mark with scent and reject recently visited flowers. Oecologia 89: 113-117

Giurfa , M., Ntifiez , J., and Backhaus, W. 1994. Odour and colour information in the foraging choice behaviour. Journal of Comparative Physiology A 175: 773-779

Giurfa, M., Zhang, S., Jenett, A., Menzel, R., and Srinivasan, M. 2001. The concepts of sameness and difference in an insect. Nature 410: 930-933

Gould, J. L. 1986. The locale map of honey bees: do insects have cognitive maps. Science 232:861-86

Goetsch, W. 1957. The Ants. The University of Michigan Press.

167

Gong, Z. F., Xia, S., Feng, C., and Guo, A. 1998. Operant visual learning and memory in Drosophila mutants’ dunce, amnesiac and radish. Journal of insect physiology 44: 1149-1158

Goulson, D. 2010. Bumblebees: Behaviour, Ecology, and Conservation. Oxford University Pres.

Goulson, D., Hawson S. A., and Stout, J. C. 1998. Foraging bumblebees avoid flowers already visited by conspecifics or by other bumblebee species. Animal Behaviour 55: 199-206

Goulson, D., Stout, J. C., Langley, J., and Hughes, W. O. H. 2000. Identity and function of scent marks deposited by foraging bumblebees. Journal of Chemical Ecology 26: 2897-2911

Grasso, F. W., and Basil, J. A. 2009. The evolution of flexible behavioral repertoires in cephalopod mollusks. Brain, Behavior and Evolution 74: 231-245

Gronenberg, W., Heeren, S., and Holldöbler, B. 1996. Age-dependent and task-related morphological changes in the brain and mushroom bodies of the ant Camponotus floridanus. The Journal of Experimental Biology. 199:2011-2019

Gross, H.J., Pahl, M., Si, A., Zhu, H., Tautz, J., and Zhang, S. 2009. Number based visual generalisation in the honeybee. PLoSONE 4(1):e4263. doi:10.1371/journal. pone. 0004263

Guo, H., Tong, J., Hannan, F., Luo, L., and Zhong, Y. 2000. A neurofibromatosis-1-regulated pathway is required for learning in Drosophila. Nature 403: 895-898

Hadad, B.,Maurer, D., Lewis, T. L. 2001. Long trajectory for the development of sensitivity to global and biological motion. Developmental Science, 14:6, pp 1330–1339

Han, P., Niu, C., Lei, C., Cui, J., and Desneux, N. 2010. Use of an innovative T-tube maze assay and the proboscis extension response assay to assess sublethal effects of GM products and pesticides on learning capacity of the honey bee. Apis mellifera L. Ecotoxicology 19: 1612-1619

Hansell, M. H. 2005. Animal architecture. Oxford University Press. p. 1.ISBN 978-0-19-850752- 9

Hart, A. G., and Ratnieks, F. L. W. 2002. Task-partitioned nectar transfer in stingless bees: work organisation in a phylogenetic context. Ecological Entomology 27: 163-168

Hay, D. A. 1975. Strain differences in maze-learning ability of Drosophila melanogaster. Nature 257: 44 -46

Healy, S. 1998. Spatial Representation in Animals. Oxford University Press, Oxford.

Heinrich, B. 1976a. Bumblebee foraging and the economics of sociality. American Scientist 64:384-395

168

Heinrich, B. 1976b. The foraging specialization of individual bumble bees. Ecological Monographs 46:105-129

Heinrich, B. 1979a. Bumble Bee Economics. Harvard University Press, Cambridge, London. 245 pp

Heinrich, B. 1979b. "Majoring" and "minoring" by foraging bumblebees, Bombus vagans: An experimental analysis. Ecology 60: 245-255

Heinrich, B. 1984. Learning in invertebrates. In: The Biology of Learning. Marler, P. and H. S. Terrace (Eds.) pp 135-147. Springer-Verlag, New York.

Heisenberg, M., Borst, A., Wagner, S., and Byers, D. 1985. Drosophila mushroom body mutants are deficient in olfactory learning. Journal of Neurogenetics. 2:1-30

Heisenberg, M., Heusipp, M., and Wanke, C. 1995. Structural plasticity in the Drosophila brain. Journal of Neuroscience. 15:1951-1960

Heyes, C., and Galef, B. J. 1996. Social Learning and the Roots of Culture. Academic Press, San Diego, CA, USA.

Heyes, C. M. 1994.Social learning in animals : categories and mechanisms Biological Review 69: 207 -231

Hoffman, B., Ritchie, D. 1997. The problems with problem based learning. Instructional Science 25: 97-115

Honzik, C. H. 1936. The sensory basis of maze learning in rats. Comparative Psychology Monographs 13: 113

Hrncir, M., Jarau, S., Zucchi, R., and Barth, F. G. 2000. Recruitment behavior in stingless bees, Melipona scutellaris and M. quadrifasciata. II. Possible mechanisms of communication. Apidologie 31: 93-113

Jackson, D. E., and Ratnieks, F. L. W. 2006. Communication in ants. Current Biology 16: 570- 574

James, W. 1899. The principles of psychology. New York: Holt, USA

Janzen, D. H. 1971. Euglossine bees as long distance pollinators of tropical plants. Science 171: 203-205

Jeanne, R. L. 1981. Chemical communication during swarm emigration in the social wasp Polybia sericea (Olivier). Animal Behaviour 29: 102-113

169

Johnson, R. 1991. Learning, memory, and foraging efficiency in two species of desert seed- harvester ants. Ecology 72:1408-1420

Johnston, T. D. 1982. Selective costs and benefits in the evolution of learning. Advances in the Study of Behavior 12:65-106

Johnston, T. D. 1985. Introduction: Conceptual issues in the ecological study of learning. In: Issues in the Ecological Study of Learning. Johnston, T. D. and A. Pietrewicz (Eds.) pp 1-24. Lawrence Erlbaum. Associates, Publishers, Hillsdale, New Jersey

Kamil, A. C. 1994. A synthetic approach to the study of animal intelligence. In: Behavioral Mechanisms in Evolutionary Ecology. Real, L. (Ed.) pp 11-45. The University of Chicago Press, Chicago & London.

Karban, R., Black, C.A., and Weinbaum, S. A. 2000. How 17-year cicadas keep track of time. Ecology Letters 3: 253-256

Karni, A. and Sagi, D. 1993. The time course of learning a visual skill. Nature 365: 250- 252. doi:10.1038/365250a0

Kawaguchi, L. G., Ohashi, K., and Toquenaga, Y. 2006. Do bumble bees save time when choosing novel flowers by following conspecifics? Functional Ecology 20: 239-244

Kawecki, T. J. 2010. Evolutionary ecology of learning: insights from fruit flies. Population Ecology 1: 15-25

Kerr, W. E. 1969. Some aspects of the evolution of social bees, Evol. Biol. 3, 119–175

Kevan, P. G. 1972. Floral colors in the high arctic with reference to insect-flower relations and pollination. Canadian Journal of Botany 50: 2289-2316

Kevan, P. G. 1979. The spectral efficiency of phototaxis for some high Arctic Diptera. Arctic and Alpine Research 11: 349-352

Kevan, P. G., and Lane, M. A. 1985. Flower petal microtexture is a tactile cue for bees. Proceedings of the National Academy of Sciences USA. 80: 4750-4752

Kevan, P. G., and Manzel, R. 2012. The plight of pollination and the interface of neurobiology, ecology and food security. The Environmentalist 32: 300-310

Kevan, P. G., Kapongo, J. P., Al-Mazra'awi, M., and Shipp, L. 2008. Honey bees, bumble bees, and biocontrol. In: Bee Pollination in Agricultural Ecosystems (eds James, RR, Pitts-Singer TL). Oxford University Press. Pp. 65- 79

Kevan, P. G., Shipp, L., and Thomas, V. G. 2014. Using pollinators for crop protection. International Innovation (in press).

170

Kisch, J., and Erber, J. 1999. Operant conditioning of antennal movements in the honey bee. Behavioural Brain Research 99: 93-102

Kohler, W. 1925. The mentality of apes. 2nd edn. New York, NY: Harcourt, Brace and Co. (transl. from German by E. Winter)

Lajoie, S. 2005. Extending the scaffolding metaphor. Instructional Science 33: 541-557

Laverty, T. M. 1980. The flower visiting behaviour of bumblebee: floral complexity and learning. Canadian Journal of Zoology 58:1324-1335

Laverty, T. M. 1994. Bumble bee learning and flower morphology. Animal Behaviour 47: 531- 545

Laverty, T. M., and Plowright, C.R. 1988. Flower handling by bumblebees: a comparison of specialists and generalists. Animal Behaviour 36: 733-740

Lavertyt, M. 1978. Flower-visiting behaviour of experienced and inexperienced bumblebees (Hymenoptera: Apidae). M.Sc. Thesis, University of Alberta, Edmonton

Leadbeater, E. A., and Chittka, L. 2005. A new mode of information transfer in foraging bumblebees. Current Biology 15: 447-448

Leadbeater, E. A., and Chittka, L. 2007b. The dynamics of social learning in an insect model, the bumblebee (Bombus terrestris). Behavioural Ecology and Sociobiology 61:1789-1796

Leadbeater, E. A., and Chittka, L. 2008. Social transmission of nectar-robbing behaviour in bumble-bees. Proceedings of the Royal Society B 275: 1669-1674

Leadbeater, E. A., and Chittka, L. 2009. Social information use in foraging insects. In: S. Jarau and M. Hrncir (eds) Food Exploitation By Social Insects: Ecological, Behavioral, and Theoretical Approaches. CRC Press, Boca Raton, FL, USA. Pp. 135-146

Leadbeater, E. A., and Chittka, L. 2011. Do inexperienced bumblebee foragers use scent marks as social information? Animal Cognition 14: 915-919

Lefebvre, L. 1995. The opening of milk bottles by birds: Evidence for accelerating learning rates, but against the wave-of-advance model of cultural transmission. Behavioural Processes 34: 43- 54

Lehrer, M., Horridge, G. A., Zhang, S. W., and Gadagkar, R. 1995. Shape vision in bees: Innate preference for flower-like patterns. Philosophical Transactions of the Royal Society of London B. 347: 123-137

Lashley, K. S. 1938. Conditional reactions in the rat. Journal of Psychology 6: 311-324

171

Leppik, E. E. 1953. The ability of insects to distinguish numbers. American Naturalist 87: 229- 236

Lihoreau, M., Raine, N. E., Reynolds, A. M., Stelzer, R. J., Lim, K. S., Smith, A. D., Osborne, J. L., and Chittka, L. 2012. Radar tracking and motion-sensitive cameras on flowers reveal the development of pollinator multi-destination routes over large spatial scales. PLoS Biology 10(9): e1001392

Lindauer, M. 1963. Allgemeine Sinnesphysiologie, Orientierungn im Raum. Zeitschrifte für Zoologie 16: 58-140

Lindauer, M. 1971. The Functional Significance of the Honeybee Waggle Dance. The American Naturalist. 105: 89-96

Mackintosh, N. J. 1974. The Psychology of Animal Learning. London: Academic Press. Marler, P. and H. S. Terrace. 1984. Introduction. In: The Biology of Learning. Marler, P. and H. S. Terrace (Eds.) pp 1-13. Springer-Verlag, Berlin.

Martin, H. 1965. Leistungen des topochemischen sinnes bei der Honigbiene. Zeitschrift für vergleichende Physiologie 50: 254-292

Mayer, R. E. 1992. Thinking, problem solving, cognition. Second edition. New York: W. H. Freeman and Company

Meller, V. H., and Davis, R. L. 1996. Biochemistry of insect learning: Lessons from bees and flies. Insect Biochemistry and Molecular Biology. 26:327-335

Menzel, R. 1969. Das Gedächtnis der Honigbiene für Spektralfarben. II. Umlernen und Mehrfachlernen. Zietschrifte für vergliechende Physiologie, Series A. 63: 290-309

Menzel, R., Erber, J., and Masuhr, T. 1974. Learning and memory in the honeybee. In: Barton- Brown L, editor. Experimental analysis of insect behaviour. New York, NY: Springer-Verlag. pp. 195-217.

Menzel, R. 1981. Achromatic vision in the honeybee at low light intensities. Journal of Comparative Physiology Series A. 141: 389- 393

Menzel, R. 1990. Learning, memory and ‘cognition’ in honey bees. In: Neurobiology of Comparative Cognition (ed. RP Kesner and DS Olten), pp. 237-292. Hillsdale, NJ: Erlbaum Inc.

Menzel, R., Leboulle, G. and Eisenhardt, D. 2006. Small brains, bright minds. Cell 124: 237-239

Mery, F., and Kawecki, T. J. 2003. A fitness cost of learning ability in Drosophila melanogaster. Proceedings of the Royal Society B: Biological Sciences 270: 2465-2469

172

Miltenberger, R. 2012. Behavior modification, principles and procedures. (5th ed., pp. 87-99). Wadsworth Publishing Company.

Mirwan, H. B., and Kevan, P. G. 2013. Social learning in bumblebees (Bombus impatiens): worker bumblebees learn to manipulate and forage at artificial flowers by observation and communication within the colony. Psyche 2013:1–8. doi:10.1155/2013/768108

Mirwan, H. B., and Kevan, P. G. 2014. Problem solving by worker bumblebees Bombus impatiens (Hymenoptera: Apoidea). Animal cognition (doi 10.1007/s10071-014-0737-0)

Mirwan, H. B., and Kevan, P. G. Submitted. Maze learning and route memorization by worker bumblebees (Bombus impatiens Cresson (Hymenoptera: Apidae).

Mittelstaedt, M.-L., and Mittelstaedt, H. 1980. Homing by path integration in a mammal. Natur wissenschaften 67:566–567

Moran, N. A. 1992. The evolutionary maintenance of alternative phenotypes. The American Naturalist 139: 971-989

Mizunami, M., Weibrecht, J. M., and Strausfeld, N. J. 1993. A new role for the insect mushroom bodies: Place memory and motor control. In: Beer RD, Ritzman RE, McKenna T, editors. Biological neural networks in invertebrate neuroethology and robotics. New York, NY: Academic Press; pp. 199-225.

Mostofsky, D. I. 1965. Stimulus Generalization. 1St Edition. Stanford University Press.

Muller, H., and Chittka, L. 2012. Consistent interindividual differences in discrimination performance by bumble bees (Hymenoptera: Apidae: Bombus terrestris) in colour, shape and odour learning tasks. Entomol General 34: 1-8

Muller, M., and Wehner, B. 1988. Path integration in desert ants, Cataglyphis fortis. Neurobiology 85: 5287-5290

Munn, N. L. 1950. The role of sensory processes in maze behavior. In: Handbook of Psychological Research on the Rat, (Munn NL) pp. 181-225. Houghton Mifflin, Boston.

Nieh, J. C. 2004. Recruitment communication in stingless bees (Hymenoptera, Apidae, Meliponini). Apidologie 35: 159-182

Niggebrügge1, C., Leboulle1, G., Menzel, R., Komischke1. B., and Hempel de Ibarra, N. 2009. Fast learning but coarse discrimination of colours in restrained honeybees. The Journal of Experimental Biology 212: 1344-1350

North, A. I., Maller, 0., and Hughes, C. 1958. Conditional discrimination and stimulus patterning. Journal of Comparative and Physiological Psychology 51: 711-715

173

Ohashi, K., and Thomson, J. D. 2012. Trapline foraging by bumblebees: VI. Behavioral alterations under speed-accuracy trade-offs. Behavioral Ecology 24: 182-189

O’Keefe, J., and Nadel, L. 1978. The hippocampus as a cognitive map (Clarendon Press).

Olson, J., and Platt, J. 2000. The instructional cycle. Teaching children and adolescents with special needs (pp. 170-197) Upper Saddle River, NJ: Prentice-Hall, Inc

Olton, D. S. 1977. Spatial memory. Scientific American 236 (6): 82-98

Pankiw, T., Page, Jr. R. E. 1999. The effect of genotype, age, sex, and caste on response thresholds to sucrose and foraging and behavior of honey bees (Apis mellifera L.). J Comp Physiol 185: 207-213

Papaj, D. R. and A. C. Lewis. 1993. Insect Learning. Chapman and Hall, Inc., New York. 398 pp.

Pavlov, I. P. 1927. Conditioned reflexes (G. V. Anrep, translation). Oxford University Press, London.

Pearce, J. M. 2008. Animal learning and cognition. An Introduction. Psychology Press. PP432

Pessotti, I. 1972. Discrimination with light stimulus and a lever-pressing response in Melipona rufiventris. Journal of Apicultural Research 11:89- 93

Pick, C. G., and Yanai, J. 1983. Eight arm maze for mice. International Journal of Neuroscience 21: 63- 66

Pickard, R. S. 1991. Applications of silicon engineering to the electrophysiological study of mushroom bodies in the honeybee brain. In: The Behavior and Physiology of Bees. Goodman, L. R. and R. C. Fisher (Eds.) pp 253-269. CAB International,Wallingford, UK.

Platt, S. A., Holliday, M., and Drudge, O. W. 1980. Discrimination learning of an instrumental response in individual Drosophila melanogaster. Journal of Experimental Psychology: Animal Behavior Processes 6: 301-311

Plowright, C. M. S., Simonds, V. M., and Butler, M. A. 2006. How bumblebees first find flowers: Habituation of visual pattern preferences, spontaneous recovery, and dishabituation. Learning and Motivation 37: 66-78

Plowright, R. C., and Laverty, T. M. 1984. The ecology and sociobiology of bumble bees. Annual Review of Entomology 29: 175-99

Punchihewa, R.W.K., Koeniger, N., Kevan, P. G., and Gadawski, R. M. 1985. Observations on the dance communication and natural foraging ranges of the three honeybee species in Sri Lanka. Journal of Apicultural Research 24: 168-175

174

Purdy, J. 2012. Learning in Invertebrates. In: Seel N. (Ed.) Encyclopedia of the Sciences of Learning: SpringerReference (www.springerreference.com). Springer-Verlag Berlin Heidelberg, 2012. DOI: 10.1007/SpringerReference_320307 2012-05-03 07:59:08 UTC.

Qu, Z. Song, Y and Ding, Y. 2010. ERP evidence for distinct mechanisms of fast and slow visual perceptual learning. Neuropsychologia 48- 1869-1874..doi:10.1016/j.neuropsychologia. 2010.01.008

Quinn, W. G., Harris, W. A., and Benzer, S. 1974. Conditioned behavior in Drosophila melanogaster. Proceedings of the National Academy of Sciences of the USA 71:708-712

Raine, N. E., and Chittka, L. 2007. Pollen foraging: learning a complex motor skill by bumblebees (Bombus terrestris). Naturwissenschaften 94: 459-464

Raine, N. E., and Chittka, L. 2008. The correlation of learning speed and natural foraging success in bumblebees. Proceedings of the Royal Society of London Series B-Biological Sciences 275: 803-808

Raine, N. E., and Chittka, L.2012. No trade-off between learning speed and associative flexibility in bumble bees: A reversal learning test with multiple colonies. PLoS ONE 7(9): e45096. doi:10.1371/journal.pone.0045096

Raine, N. E., Ings, T. C., Ramos-Rodríguez, O., and Chittka, L. 2006. Intercolony variation in learning performance of a wild British bumble bee population (Hymenoptera: Apidae: Bombus terrestris audax). Entomol General 28: 241-256

Ray, S., and Ferneyhough, B. 1997. The effects of age on olfactory learning and memory in the honey bee Apis mellifera. NeuroReport Journal 8: 789-793

Raygor R. 2005.The Science of Psychology, New York; McGraw-Hill 2005. New York.

Real, L. A. 1994. Behavioral Mechanism in Evolutionary Ecology. The University of Chicago Press, Chicago and London. 480 pp.

Regolin, L., Vallortigara, G. and Zanforlin M. 1995. Detour behaviour in the domestic chick: searching for a disappearing prey or a disappearing social partner. Animal Behaviour 50: 203- 211

Renner, M. A., and Nieh, J. C. 2008. Bumble bee olfactory information flow and contact-based foraging activation. Insectes Sociaux 55: 417-424

Ritter, F. E., and Schooler, L. J. 2002. The learning curve. In International encyclopedia ofthe social and behavioral sciences. 8602-8605. Amsterdam: Pergamon. http://www.iesbs.com

175

Rohrseitz, K., and Tautz, J. 1999. Honey bee dance communication: waggle run direction coded in antennal contacts. Journal of Comparative Physiology A 184: 463-470

Roiblat, H. L. 1987. Introduction to Comparative Cognition. W. H. Freeman, New York. 377 pp

Rosengren, R. 1971. Route fidelity, visual memory and recruitment behaviour in foraging wood ants of the genus Formica (Hymenoptera, Formicidae). Acta Zoologica Fennica 133: 1-106

Roubik, D. W. 1992. Ecology and natural history of tropical bees, Cambridge University Press, New York.

Rubadeau, D. O., and Conrad, K. O. 1963. An apparatus to demonstrate and measure operant behaviour of Arthropoda. Journal of the experimental analysis of behaviour 6: 429-430

Saavedra, M. A. 1975. Pavlovian compoundconditioning in the rabbit. Learning and Motivation 6: 314-326

Saleh. N,, Scott, A. G., Bryning, G. P., and Chittka, L. 2007 Distinguishing signals and cues: bumblebees use general footprints to generate adaptive behaviour at flowers and nest. Arthropod-Plant Interactions 1: 119-127

Sanz, C., Call, J., and Boesch, C. 2013. Tool use in animals: Cognition and Ecology. Cambridge University Press.

Sawyer, K. R. 2006. The Cambridge handbook of the learning sciences. New York: Cambridge University Press

Schacter, D. L., Gilbert D. T., and Wegner, D. M. 2011. "B. F. Skinner: The role of reinforcement and Punishment", subsection in: Psychology; Second Edition. New York: Worth, Incorporated, 278-288

Schacter, D. L., Gilbert, D. T., Wegner, D. M. 2011. Psychology, Second Edition. Worth Publishers. p. 264.

Scheiner, S. M. 1993. Genetics and evolution of phenotypic plasticity. Annual Review of Ecology and Systematics 24:35-68

Schmidt, V. M., Zucchi, R., and Barth, F. G. 2006. Recruitment in a scent trail laying stingless bee (Scaptotrigona aff. depilis): Changes with reduction but not with increase of the energy gain. Apidologie 37: 487-500

Schmitt, U., and Bertsch, A. 1990. Do foraging bumblebees scent-mark food sources and does it matter? Oecologia 82: 137-144

Schmitt, U., Lübke, G., and Franke, W. 1991. Tarsal secretions mark food sources in bumblebees (Hymenoptera: Apidae). Chemoecology 2: 35-40

176

Schneirla, T. C. 1929. Learning and orientation in ants. Comparative Psychology Monographs 6: 1-143

Schneirla, T. C. 1941. Social Organization in insects, as related to individual function. Psychological Review 48: 465-486

Schöne, H., and Tengö, J. 1981. Competition of males, courtship behaviour and chemical communication in the digger wasp Bembix rostrata (Hymenoptera, Sphecidae). Behaviour 77: 44-66

Schooler, J. W., Ohlsson, S., and Brooks, K. 1993. Thoughts beyond words: When language overshadows insight. Journal of Experimental Psychology: General 122: 166-183

Schrier, A. M., and Thompson, C. R. 1980. Conditional discrimination learning: A critique and amplification. Journal of the Experimental Analysis of Behavior 33: 291-298

Séguin, F. R., and Plowright, C. M. S. 2008. Assessment of pattern preferences by flower-naïve bumblebees. Apidologie 39: 215-224

Sherry, D. F. 2008. Social learning: Nectar robbing spreads socially in bumble bees. Current Biology 18: 608-610

Shettleworth, S. J. 1998. Cognition, evolution and behaviour. Oxford, UK: Oxford University Press.

Shettleworth, S. J. 2010. Cognition, evolution and behaviour. Oxford, UK: Oxford University Press.

Shettleworth, S. J. 2012. Do Animals Have Insight, and What Is Insight Anyway? Canadian Journal of Experimental Psychology. Canadian Psychological Association 66: 217-226

Shettleworth, S. J., J. R. Krebs, D. W. Stephens and J. Gibbon. 1988. Tracking a fluctuating environment: A study of sampling. Animal Behaviour 36:87-105

Skinner, B. F. 1938. "The Behavior of Organisms:An Experimental Analysis", New York: Appleton-Century-Crofts

Skinner, B. F. 1950. "Are theories of learning necessary?", Psychological Review,57, 193-216

Skinner, B. F. 1953. Science and human behavior. pp. 92-3. Oxford, England: Macmillan.

Skinner, B. F. 2005. Science and Human Behavior. the b.f. skinner foundation.

Sladen, F. W. L. 1912. The humble-bee its life history and how to domesticate it. With description of all British species of Bombus and Psithyrus. Macmillan and Co., Ltd.

177

Sokolowski, M. B. C., Disma, G., and Abramson, C. I. 2010. A paradigm for operant conditioning in blow flies (Phormia terrae novae Robineau-Desvoidy 1830). Journal of the experimental analysis of behavior 93: 81-89

Spiegel, R., and McLaren, I. P. L. 2006. Associative sequence learning in humans. Journal of Experimental Psychology: Animal Behavior Processes 32: 156-163

Srinivasan, M., and Zhang, S. W. 2003. Small brains, smart minds: vision, perception, and cognition in honeybees. IETE Journal of Research 49: 127-134

Staddon, J. E. R., and Cerutti, D. T. 2003. Operant conditioning. Annual Review of Psychology 54: 115.144

Stephens, D. W. 1991. Change, regularity, and value in the evolution of animal learning. Behavioral Ecology 2:77-89

Sun, R. 2001. Introduction to Sequence Learning. Sequence Learning. Lecture Notes in Computer Science, Computer Science 1828/2001

Sun, R., and Giles, C. L. 2001. "Sequence Learning: From Recognition and Prediction to Sequential Decision Making". IEEE Intelligent Systems and Their Applications 16 (4)

Takeda, K. 1961. Classical conditioned response in the honey bee. Journal of Insect Physiology 6:168-180

Tarsitano, M. 2006. Route selection by a jumping spider (Portia labiata) during the locomotory phase of a detour. Animal Behaviour 72: 1437-1442

Tarsitano, M. S., and Jackson, R. R. 1994. Jumping spiders make predatory detours requiring movement away from prey. Behaviour 131: 65-73

Taylor, A. H., Elliffe, D., Hunt, G. R., and. Gray, R. D. 2010. Complex cognition and behavioural innovation in New Caledonian crows. Proceeding of The Royal Society B Biological Sciences 277: 2637-2643

Terrace, H. S. 1987. Chunking by a pigeon in a serial learning task. Nature 325: 149-151

Terrace, H. S. 1991. Chunking during serial learning by a pigeon: I. Basic evidence. Journal of Experimental Psychology: Animal Behavior Processes 17: 81-93

Thomas, D. R., Curran, P. J., and Russell, R. J. 1988. Factors affecting conditional discrimination learning by pigeons: II. Physical and temporal characteristics of stimuli Animal Learning & Behavior 6: 468-476

178

Thompson, H. M. 2003. Behavioural effects of pesticides in bees - Their potential for use in risk assessment. Ecotoxicol 12: 317-330

Thomson, J. D., Slatkin, M., and Thomson, B. A. 1997. Trapline foraging by bumblebees. Definition and detection from sequence data. Behavioral Ecology 8: 199-210

Thorndike, E. L. 1901. Animal intelligence: An experimental study of the associative processes in animals. Psychological Review Monograph Supplement 2: 1-109

Thorpe, W. H. 1950. A note on detour experiment with Ammophila pubescens Curt. (Hymenoptera; Sphecidae). Behaviour 13: 257-263

Thorpe, W. H. 1956. Learning and Instinct in Animals. Methuen, London. 568 pp.

Tinbergen, L. 1960. The natural control of insects in pinewoods. I. Factors influencing the intensity of predation by songbirds. Archives Neerlandaises de Zoologie 13:266-343

Tomina, Y., and Takahata, M. 2010. A behavioral analysis of force-controlled operant tasks in American lobster. Physiology and behavior 101: 108-116

Tully, T. 1984. Drosophila learning: Behavior and biochemistry. Behavior Genetics 14: 527-557

Vygotsky, L. S. 1987. Thinking and speech. In L. S. Vygotsky, Collected works (1: pp. 39-285) (R. Rieber and A. Carton, Eds; N. Minick, Trans). New York: Plenum. (Original works published in 1934, 1960)

Wcislo, W. T. 1989. Behavioral environments and evolutionary change. Annual Review of Ecology and Systematics 20:137-169

Wehner R., and Menzel, R. 1990. Do insects have cognitive maps? Annual Review of Neuroscience 13: 403-414

Wehner, R. 1992. Arthropods. In: Animal Homing (ed. F. Papi), pp. 45-144. London: Chapman and Hall.

Wehner, R., Michel, B., and Antonsen, P. 1996. Visual navigation in insects: coupling of egocentric and geocentric information. Journal of Experimental Biology 199: 129-140

Weisman, R. G., Dodd, P. W. D., Wasserman, E. A., and Larew, M. B. 1980. Representation and retention of two-event sequences in pigeons. Journal of Experimental Psychology: Animal Behavior Processes 6: 312-325

Weiss, K. 1953. Versuche mit Bienen und Wespen in farbingen Labyrinthen. Zeitschifte für Tierpsychologie 10: 29-44

179

Wells, H. and P. H. Wells. 1986. Optimal diet, minimal uncertainty and individual constancy in the foraging of honey bee Apis mellifera. Journal of Animal Ecology 55: 881-892

West-Eberhard, M. 2003. Developmental Plasticity and Evolution. Oxford University Press, Inc., New York. 794 pp.

Whiten, A., and Ham, R. 1992. On the nature and evolution of imitation in the animal kingdom - reappraisal of a century of research. Advances in the Study of Behavior 21: 239-283

Wiese, H. 2003. Numbers, Language, and the Human Mind. Cambridge University Press.

Wilms, J., and Eltz, T. 2008. Foraging scent marks of bumblebees: footprint cues rather than pheromone signals. Naturwissenschaften 95: 149-153

Wilson, E. O., and Hölldobler, B. 1990. The Ants. Springer-Verlag, Berlin Heidelberg.

Withers, G. S., Fahrbach, S. E., and Robinson, G. E. 1993 Selective neuroanatomical plasticity and division of labour in the honeybee. Nature. 364:238-240

Witjes, S. and Eltz, T. 2007. Influence of scent deposits on flower choice: experiments in an artificial flower array with bumblebees. Apidologie 38: 12-18

Wong, M., Peters, R. M., Goldreich, D. 2013. "A Physical Constraint on Perceptual Learning: Tactile Spatial Acuity Improves with Training to a Limit Set by Finger Size". Journal of Neuroscience 33 : 9345-9352. doi:10.1523/JNEUROSCI.0514-13.2013

Wood, D. C. 1988. Habituation in Stentor produced by mechanoreceptor channel modification. Journal of Neuroscience 8: 2254-2258

Woodward, G. L., and Laverty, T. M. 1992. Recall of flower handling skills by bumble bees: a test of Darwin's interference hypouthesis. Animal Behaviour 44: 1041-1051

Worden, B. D., and Papaj, D. R. 2005. Flower choice copying in bumblebees. Biology Letters 1: 504-507

Wustmann, G., Wustmann, G., Rein, K., Wolf, R., and Heisenberg, M. 1996. A new paradigm for operant conditioning of Drosophila melanogaster . Journal of Comparative Physiology A. Neuroethology Sensory Neural and Behavioral Physiology 179: 429-436

Wyatt, T. D. 2003. Pheromones and Animal Behaviour: Communication by Smell and Taste. Cambridge University Press, Cambridge, UK.

Zhang, S. W., Bartsch, K., and Srinivasan, M. V. 1996. Maze learning by honeybees. Neurobiology of Learning and Memory 66: 267-282

180

Zhang, S. W., Lehrer, M., and Srinivasan, M. V. 1998. Stimulus- conditioned sequence learning in honeybees. In: Proceedings of the 26th Goettingen Neurobiology Conference 1998 (eds Elsner N and Wehner R), 2: 519. Stuttgart: Thieme.

Zhang, S. W., Lehrer, M., and Srinivasan, M. V. 1999. Honeybee memory: Navigation by associative grouping and recall of visual stimuli. Neurobiology of Learning and Memory 72: 180-201

Zhang, S., Mizutani, A., and Srinivasan, M. V. 2000. Maze navigation by honeybees: Learning path regularity. Learning and Memory 7: 363-374

Zohar, O., and Terkel, J. 1996. Social and environmentam factors modulate of pine-cone stripping techniques by black rats, Rattus rattus. Animal Behaviour 51: 611-618

Zucca, P., Antonelli, F., and Vallortigara, G. 2005. Detour behaviour in three species of birds: quails (Coturnix sp.), herring gulls (Larus cachinnans) and canaries (Serinus canaria). Animal Cognition 8: 122-128

181