710050472 1 Does the volume of sucrose reward at a flower patch influence bumblebees’ spatial memory ? 710050472 Word count: 7,9 9 3 PSYM210 University of Exeter 710050472 2 Acknowledgements I would like to thank my supervisor for this project, Dr Natalie Hempel de Ibarra, who facilitated my work with these wonderful creatures. I would also like to give a huge special thanks to Dunia Gonzales, who was constantly available and offered a lot of support and guidance throughout my project, while also busy working on her own PhD project H er help was invaluable. I would also like to extend my sincerest gratitude now and always to my husband, who’s unconditional support helps me so much in everything that I do. I dedicate this piece of work to our family; us both and our first child who is due in December, as well as our two wonderful rabbits 710050472 3 Abstract Bumblebees conduct learning flights to retain information about the location of their nest or a feeding site. A suggested influence on these learning flights at flower patches is the amount of reward offered . The present study aimed to investigate this in terms of learning flight characteristics at different flower patches and volume of sucrose reward offered from flowers B umblebees were observed at two flower patch es in which flowers contained a high volume of sucrose, and a low volume of sucrose. Results and analysis showed that bumblebees at the flower patch offering a low volume of sucrose reward visited significantly more flowers and performed significantly longer learning flights in trajectory and duration. It was concluded that the reward gained from a flower patch may have a significant influence on learning flight s and the formation of spatial memories of a flower patch, potential ly due to the need to gather more reward from the entire patch and learn about additional flowers , due to singular flowers being less rewarding. Keywords: learning flight, flower patch, reward, spatial memory, sucrose volume 710050472 4 Introduction Bumble bees, Bombus terrestris , are essential pollinators that support agriculture and natural processes in our ecosystems. Their importance to wildlife, as well as the dramatic decline in population numbers worldwide, has drawn significant attention to this species, making them the second most studied group of bees worldwide (Ghisbain, 2021). The threat to this species stems largely from changes in land - use destroying and fragmenting natural and semi - natural habitats (Potts et al., 2016). This has modified and reduced the number of floral resources such as pollen and nectar, needed by bumblebees in addition to nest resources (Potts et al., 2016). Optimising and providing additional flower resources has therefore become a widely suggested mitigation strategy in order to stop the threat of extinction and threat to our ecosystems, meaning research into their behaviour pertaining to flower choice and flights at flowers has become essential in adding to information that can improve strategies to protect this species and our environment. In studying this species, m any pieces of research have investigated bumblebee s’ ‘learning flights ’ Bumblebees perform what is known as a ‘learning flight’, at important locations that they will return to, such as flowers and their nest site , to form spatial memories of these locations by gathering visual information to be able to navigate back to them (Robert, Frasnelli, Hempel de Ibarra, & Collett, 2018). Th ey are also capable of perform ing a learning flight just once and be ing able to return to the location afterwards (Robert et al., 2018). Past literature has explored the key characteristics of these flights for different species of bee For example, it has been found that bumblebees turn around to face the location they are leaving instead of flying in a straight line, to remember the visual characteristics around them as the y leave (Zeil, Kelber, & Voss, 1996). This was also found by Lehrer, (1993) and Lehrer and Collett (1994), i n which bees were found to encode certain visual aspects of a flower and its location on appro ach, and t o turn back to face the flower when conducting a learning flight on departure This is consistent with other research findings outlining that the learning flight ’s key functions are to allow the bee to encode the visual scene around a goal (i.e., flower) and to learn the distance of certain landmarks (Zeil et al., 1996). I t has also been suggested that the arcs involved in these learning flights serve the function of determining which objects are close to their goal, with learning flights allowing them to yield significant 710050472 5 information about their surroundings. Learning flights have been indicated to allow for efficient gathering of this information, with the process, consisting of phases of orientation, allowing bumblebees to learn the location of food and nest sites very quickly (Osborne et al., 2013). Findings pertaining to the pattern and direction of these learning flights, have also added to information on this process of form ation of spatial memories in bumblebees. In research by Hempel de Ibarra, Philippides, Riabinina, and Collett (2009) , it was found that a significant influence on viewing direction an d flight pattern during a learning flight was the goal of exploiting cues to bee’s olfactory and visual systems. Similarly, a consistent influence found in bee’s viewing direction during a learning flight is the ir learning of the landmarks surrounding the location that they are encoding (Plante - Ajah, 2019) P atterns of flying involving loops and zig zags have also been investigated in previous research, such as in work by Philippides, Hempel de Ibarra, Riabinina, and Collett (2013), with it be ing found that loops away from and toward the nest are less common in return flights , and instead performed during initial le arning flights, and zig zag manoeuvres are more common during return flights These findings regarding bumblebee learning flights , however, only pertain to the nest. There have also been a wide range of findings from previous research regarding learning flight behaviour at flower s , specifically For example, both honeybees and bumblebees have been found to make assessments of flowers that influence their learning flights. It has been found that when a bee first comes across a flower, they will de cide whether the flower is worth utilizing (Frasnelli et al., 2021). A bee will learn a flower’s surroundings, and the flower itself on approach , h owever, whether this information is worth retaining is only decided upon after they have sampled from the flower and assessed its worthiness (Frasnelli et al., 2021). Bees also quickly learn to associate a flower’s reward with its shape, pattern, colour, and smell , when they do decide t o re tain a spatial memory of it (Cnaani, Thomson & Papaj, 2006). In this way, both honeybees and bumblebees have a significant capacity to learn to find and remember flowers of completely different characteristics (Laverty, 1994) I n a similar way, it has also been found that the reward gained from a flower can strongly influence the characteristics of a learning flight in bumblebees and honeybees. For example, 710050472 6 it was found in research by Wei and Dyer (2009) that an increase in either nectar concentration or volume was significant in inducing longer learning flights in honeybees. A similar pattern has also been found regarding flower choice, with bees being found to fly further or stay closer to a certain flower to continue to forage d epending on whether the flower they fed from was significantly rewarding. This was found in a study that investigated foraging patterns and decisions in bumblebees when presented with flowers that contained water or nectar, with those that obtained nectar staying within the vicinity, and those that did not obtain a ny nectar reward from the flower fl ying further away (Dukas & Real, 1993). Findings such as these pertaining to flowers have not only been seen in terms of single flowers, however, but also in t erms of entire patches of flowers . Bumblebees have been seen to perform learning flights of an entire flower patch, in which they will hover and rotate around the flower that they fed from, often also feeding from additional multiple flowers in the patch, which they also rotate and hover around before circling the entire patch in circles of increasing size (Robert et al., 2018). Previous research has also found that the memory and learning flights of entire flower patches can also differ, much like a t the nest and at singular flowers. For example, initial visitation of certain patches of flowers has been suggested to be influenced by factors such as the size of the flower patch, area , and density. It was found in multiple previous studies that there i s an association between bumblebee visitation at certain flower patches and the size of the patches. For example, floral density and flower patch size, referring to the number of open flowers present at the patch, was found to have a significant influence on bumble bee visitation in research by Walters and Stiles (1996). T ype of flower , as well as size and density, has also been indicated to have an influence on visitation, for example in bumble bees gravitating towards lavender when presented with a choice between this flower and borage (Garbuzov & Ratnieks, 2014). In a similar way, some bumble bees have shown a significant preference for radial patterns when tested against concentric patterns in flower patches, in flower naïve bees (Séguin & Plowright, 200 8), which also replicated prior findings of a bias towards radial patterns when compared with plain discs (Plowright, Simonds & Butler, 2006; Simonds & Plowright, 2004). It can be suggested that these differences in visitation of certain flower patches are then influences on bumblebees’ spatial memories of a patch, as they must select and visit a patch to decide to encode its location as a spatial memory through a learning flight 710050472 7 Landscape context has also been suggested to play a role in influencing bumblebee learning of flower patches in a similar way. It was found in research by Heard et al., (2007), that landscape context , rather than patch size , dictated the number of bumblebees foraging at sown patches, suggesting that the context and characteristics of the flower patch may have a particular influence on visitation, and perhaps subsequently the formation of spatial memories. It was found specifically in this previous study that the number of bumbl e bees at a certain flower patch did not vary at all based on the size of the patch but varied with the context in which the patch was found. This suggests that factors pertaining to context are important in selecting flower patches to visit, which may indicate that these are also influences on the formation of spatial memories of a patch. This idea is potentially further supported by additional findings regarding learning at the flower patch, in which the more apparently complex a flower’s morphology was , the longer a bumble bee took to learn how to handle the flower – which could also translate to learning the overall location of the patch, though this has not yet been investigated explicitly (Laverty, 1994). I t has also been suggested by different findings that learning flights have bee n found to be potentially more demanding and take more time for bumblebees at the nest than at flowers, even if surroundings and characteristics are similar , however (Robert et al., 2018) In addition, in research by Goulson (2000), it was in fact found th at the proportion of inflorescences visited within a patch declined as patch size increased, and pollinators spent proportionally less time at larger patches , which means they may have spent less time learning about the patch . This highlights a potential c onflict in findings, regarding patch size having a significant effect on bee visitation to flowers at a flower patch , and therefore potentially the form ation of spatial memories of them, calling into question whether the physical attribute of patch size has an influence in visitation or learning flight behaviour. Alt ogether, this could potentially suggest that the reward, rather than the characteristics or phy sical attributes of a flower patch , may be the main influence in learning the location of a patch of flowers. This is a logical suggestion, based on the fact that in the wild, both honeybees and bumblebees that are foraging likely encounter a wide variet y of different flowers with different reward value, suggesting they may have to be selective in the patches at which they invest in forming a spatial memory. Many pieces of research investigating this idea, in fact, 710050472 8 suggest that this could be the case. For example, findings in research by Cartar (2004) indicate that bumblebees’ visitation and return rate to flower patches are strongly influenced by how much reward they gain, as an investigation into this effect showed that bumbleb ees strongly adjusted their visitation to flowers based on the quality of them. It was found that bumblebees visited more flowers in a patch when a flower they fed from was more rewarding, and this also made them return to the patch more frequently (Cartar , 2004) Similar findings were also shown in research by Cnaani, Thomson and Papaj (2006). The same adjustment of behaviour based on differing reward was investigated in this study, through the presentation of two artificial flowers, initially boasting the same reward, and then subsequently after bumblebees were familiar with them, researchers changed either the volume or concentration of the sucrose reward from one of the flowers and observed the change in their behaviour. It was found that bees then significantly shifted their visits to the more profitable flower type. It was also found that there was a stronger response to a change in concentration of sucrose rather than volume, supporting the idea also that bees prefer to forage on floral types asso ciated with nectar of a higher concentration (Bitterman, 1976; Whitham, 1977). These studies’ findings both suggest that a higher reward gained at a patch of flowers may strongly influence a bee’s visitatio n to them , suggesting that this also may have an e ffect on the formation of spatial memories of them Research by Frasnelli et al., (2021) can also be seen to support and extend this particular idea In this study it was investigated, in a similar way to previously mentioned studies, whether the concentration of sucrose in artificial flowers significantly influenced bumblebees’ behaviour at the flower patch ; specifically, their learning flight behaviour. This was also extended, through the additional inves tigation of differences between different sizes of bee, as to whether this would impact their learning based on reward from a flower Through studying learning flight patterns, comparing 20% and 50% concentration sucrose solution in an artificial flower si tuated in a greenhouse, and comparing small and large bumblebees in this behaviour, significant differences in learning flight behaviour were found. In terms of sucrose concentration alone, it was found that flower - facing behaviour significantly differed w ith increasing reward, increasing with a higher concentration of sucrose, and alongside this, flower - facing also increased with the duration of the learning flight. This potentially supports the idea that increased reward leads to bumblebees forming a spat ial memory of a patch, as both flower - facing behaviour and duration of learning flights have been associated with 710050472 9 quality of a spatial memory of the location of the nest or a flower (Wei & Dyer, 2009). In addition, it was also found that the size of a bumblebee influenced learning flight behaviour differences based on the sucrose reward from the flower. Smaller bees w ere found to invest equally in both largely rewarding and less rewarding flowers, whereas larger bees focused predominantly on the highly rewarding flowers, not learning much information about those offering smaller rewards i.e., those containing less sucr ose concentration. The longest learning flights, and most flower - facing behaviour overall, also came from large bumblebees who fed from the highly concentrated flowers (50% sucrose concentration). This was not to say that smaller bees were not discerning a t all in reward, as there was still suggested to be a threshold of reward below which small bumblebees w ould choose not to feed fro m a flower These findings are also indicated to mirror the behaviour differences of bumblebees in the wild, with larger bees flying further afield to find highly rewarding flowers, and smaller bumblebees staying closer to the nest with more activity at the hive , mean ing less travelling for the sake of flowers that are more rewarding , as well as them potentially having a lower threshold for highly concentrat ed nectar. This study and its findings significantly indicate differences in learning flight behaviour based on t he reward a bumblebee gains from a flower or patch of flowers and implies that the formation of a spatial memory could differ based on the reward gained from a flower patch. This idea is further echoed by research by Robert (201 7 ) , in which spatial memory was investigated through learning flight behaviour, also related to different amount of reward through different concentrations of sucrose. B umblebee learning flights were recorded to compare their learning flights interacting wit h different concentrations and volumes of sucrose in an artificial flower. It was found that bumblebees’ drinking time was positively associated with the volume of sucrose in the flower, with them also dr inking for longer when the concentration and volume of sucrose was higher. It was also found that bumblebees’ learning flights were significantly longer when the reward from the flower was higher ; specifically , when the concentration of sucrose was highest. Bumblebees also increased their rate of fixation and flower - facing as the reward quality increased, with fixations on the flower increasing in frequency with the concentration of sucrose. Learning flight duration, flower - facing and fixation are all suggested in literature to translate to a stro nger spatial memory being formed of a location through a learning flight (Wei & Dyer, 2009), meaning that with a higher reward also came a stronger memory of the flower ’s location. Flower fixations 710050472 10 specifically are suggested to be highly significant in acquiring visual information regarding the physical features of a flower and its location (Robert et al., 2018; Wei & Dyer, 2009). Taken together, these pieces of research outline a possible robust influence of sucrose reward on the spatial memories form ed through the learning flights of bumblebees. There has also been slightly less investigation into differences specifically in the volume of reward, with most previous studies focusing on sucrose concentration as a measure of reward from a flower. This st udy therefore aims to build upon this idea and fill this gap in literature , investigating whether the volume of sucrose reward in a flower patch influences learning flight characteristics and spatial memories of the flower patch in bumblebees. Specifically , this study aims to answer: i) whether the number of flowers bumblebees visit and sample from is different when flowers at a patch offer a higher volume of reward, ii) whether bumblebees spend more time looking at a flower they fed from when the reward is higher , and whether learning flights are structured in the same way in terms of iii) the length of their trajectory and iv) their duration when the volume of reward fro m flowers in a flower patch is higher or more limited. Based on previous literature, it is hypothesised that learning flights at a five - flower patch will significantly differ based on the volume of sucrose reward offered at each patch Sp ecifically, it is hypothesised that bumblebees will i) visit and sample from more flowers, ii) look significantly more at the flower they originally fed from in the patch (more flower - facing behaviour) , and the learning flight conducted will be more comprehensive ; iii) the trajectory of bumblebees’ learning flights and iv) the total duration of their learning flights will be longer , for the bumblebees visiting a flower pat ch in which the volume of sucrose offered from flowers is higher than for bumblebees visiting a patch in which the sucrose volume from flowers is more limited ; due to the motivation to form a better spatial memory of the patch because they received higher reward from flowers. 710050472 11 Method Location Experiments were conducted at a greenhouse at the University of Exeter, Streatham Campus, UK, with a floor area of 8x12m. Materials A commercially reared Bumble bee colony, of species Bombus terrestris audax (Linnaeus, 1758) , were used , sourced from Koppert Products, Haverhill, UK. Each bumblebee was marked with a coloured tag with a number on, to identify each bumblebee. Bumblebees were kept in a nest box, made of wood and Perspex, with a plastic tube attached to the entrance that came out the top of the nest table, where bumblebees could exit. The nest table was 1.5m x 1.8m, covered in white gravel that was raked frequently to prevent any arbitrary cues being picked up by bees from the gravel instead of the flowers The ‘nest hole’, in the midd le of this nest table was the location that bumblebees would exit out of the tube. In addition, the nest table also had three ‘landmarks’, consisting of three black cylinders, measuring 17 cm x 5cm , in a 120 - degree arc around the hole in the middle of the nest table, the centres of them measuring 24.5cm from th is nest hole. Another table, 5m away, was used as the ‘flower patch table’ (Figure 1) This table also measured 1.5m x 1.8m and was also covered in white gr avel that was frequently raked to keep the physical surroundings consistent and prevent any arbitrary cues being picked up by bees from the gravel instead of the flowers The flower patch table also consisted of three landmarks, identical black cylinders to the nest table, measuring 17cm x 5cm, in a 120 - degree arc around the flower patch in an identical position to that of the nest table. The ‘flower patch’ on the flower patch table consisted of 5 art ificial flowers , equally spaced 5cm away from each other, and classified by their compass - point direction ( Southwest , Southeast , Northwest , Northeast , and Central flower) . E ach artificial flower in the patch consist ed of a purple - coloured matte plastic rin g, measuring 5cm in the outer diameter, with a plastic Eppendorf tube in the centre dispensing 50% (w/w) sucrose solution for the bumblebee to drink from. In the higher volume condition, each fl ower w as filled with 0.5ml of 50% (w/w) sucrose solution, and in the lower volume condition, each flower was filled with 20 μL , filled 710050472 12 using a n adjustable micropipette. This smaller volume was intended to represent the limited amount bumblebees would drink from flowers in a singular foraging bout in the wild, (Cavalcante et al., 2018; Power et al., 2017; Pyke et al., 2020) , with the higher volume therefore representing a flower patch with comparatively unlimited reward. The plastic ring of each flower and the micropipette were frequently cleaned, with the flowers being wiped with 70% ethanol. A video camera (Panasonic HC - V720, HD 1080p) was placed approximately 1.35m above the flower patch table, secured with tape around a metal beam, to capture learnin g flights on video. Experimental procedure Bumblebees were housed in the nest box under the nest table previously described and left the next box through the plastic tube attached to the box , which allowed them to exit out of the nest hole in the middle of the nest table when they wished to leave t he nest . Bumblebees Eppendorf tube Artificial flower Landmark White gravel Figure 1 - Artificial Five Flower Patch table experiment set - up 710050472 13 were then allowed to perform a learning flight of the nest table location, before being carefully captured in a butterfly net one by one. The bees were then tagged with a coloured number tag , and then left to settle down after being tagged for a period of ten minutes. Ten bumblebees were allowed to leave the nest to take part in the experiment to make the number of bees in the greenhouse manageable and minimise interference , with the nest box opening le ading to the plastic tube being closed once ten in total had exited out of the nest table. Bumblebees were then placed on an artificial flower, individually, with only one bee being at the flower patch at one time and allowed to drink from the flower they were placed o n The flower that the bee was placed on was randomised so that no flower was used more than others as the initial flower that bee s drank from , with them either being placed on the Southeast , South w est, C entral, Northeast, or North w est flower The bee was then allowed to drink from any additional flowers, and then allowed to do a learning flight before leaving the patch The bumblebees’ behaviour at the flower patch table was recorded continuously at 50 fr ames per second, using the video camera secured above the table, capturing an area of approximately 70cm x 90cm, in an image consisting of 1920 x 1080 pixels. Once they had performed a learning flight, bees were allowed to fly back to the nest hole or elsewhere in the greenhouse, whilst the process was repeated for the next bee. Researchers present made sure no bees that had already drank from the flower patch interfered with the indiv idual at the flower patch by capturing them or encouraging them away if they got too close to the flower patch. All artificial flowers were cleaned in between each bee visiting , with 70% ethanol solution being wiped over each flower , and researchers topping up the sucrose in the centre of the flower , after rinsing the Eppendorf tube to prevent build - up of sugar. Once all bees had drank from the patch once, the patch was monitored for return flights of each individual bee, to record if they came back to the patch. They were then captured and placed back into the nest. The same cleaning and replenishing of sucrose process was repeated between each return flight. This same process was used for the experiment involving an identical 5 - flower p atch with 0.5ml of sucrose solution instead of 20 μL, conducted by different researchers in July 2021, 710050472 14 with the lower volume ( 20 μ L) experiment conducted in July 2022, and compared with this existing set of data with a higher volume of reward offered in an identical flower patch. Overall, due to some bees being eliminated from analysis due to struggling to perform learning flights, a sample of 9 bumblebees and 15 bumblebees were analysed in their learning flights at the low volume and high volu me of sucrose flower patches (respectively). Although recorded, return flights were not included in analysis due to interference of other bumblebees and researcher cleaning. Data Analysis The orientation of bumblebees’ heads and bodies and their positions and learning flight trajectories in the video recordings were extracted using MATLAB software with custom - written codes (Philippides et al., 2013) , and the number of flowers visited and sa mpled from by bumblebees was recorded from video footage The number of flowers visited by bumblebees at each type of patch was recorded from video footage of each bumblebee at each flower patch, as mutually exclusive categories of i ) flowers that bumblebees approached but did not land on, ii) flowers that bumblebees landed on but did not feed from, and iii) flowers that bumblebees fed from. These categories were combined into one outcome variable for analysis due to there only being three cases in which bumblebees fed from an additional flower to the one that they originally fed from , and therefore one outcome variable being more meaningful in analysis R software was then used to conduct statistical analyses on all data (4.1.2; https://www. r - project.org/). 710050472 15 Results Difference s in number of flowers visited T he spread of the data for the low volume flower patch appears to be wider than that of the high - volume patch, with a wider range of the number of flowers visited by each bumblebee at that flower patch. It is also indicated t hat the bumblebees in the low volume flower patch may have visited more flowers overall than those at the flower patch in which flowers had a higher volume of reward offered (Figure 2). Figure 2 : T he number of flowers visited by bumblebees at the low volume and high - volume flower patches . Shown are the median and lower/upper quartiles (box) with one outlier. The whiskers depict the highest and lowest values recorded, and the mean is depicted by an ‘ x’. # 710050472 16 A Mann - Whitney U test was conducted to investigate whether the number of flowers visited by bumblebees was significantly different at the low - volume flower patch compared to the high - volume flower patch. A Mann - Whitney test indicated that the number of flo wers visited by bumblebees at the five - flower patch was significantly greater at the low - volume flower patch (Mdn = 5), than at the high - volume flower patch (Mdn = 3), U = 15, p = 0.0015. A Mann - Whitney U test was conducted due to normality assumptions of parametric tests not being met Frequency of looking at flowers The frequency at which bumblebees looked at each angle in relation to the flower that they originally fed from was recorded and compared between the five - flower patch containing the low volume of sucrose solution and the five - flower patch containing the high volume of sucrose solution. (Figure 3 ). I t is indicated that bumblebees faced the flower that they originally fed from at a similar frequency at the five - flower patch in which the flowers offered a high volume of sucrose solution and the five - flower patch in which the flowers offered a low volume of sucrose solution (Figure 3) Figure 3 : Distribution of frequencies at which bumblebees’ retinas were at each angle of orientation relative to the flower they fe d from at both flower patches, in degrees (°) 710050472 17 Data was analysed using a two - way repeated measures ANOVA to investigate whether there was a significant difference in the number of frames in which bumblebees lo oked at each angle in relation to the flower they originally fed from during video footage of their learning flights. Looking angles (in degrees) were treated as categorical for analyses. It was found that there was no significant difference in the number of frames that bumblebees looked at each angle towards and away from the flower between the flower patch in which flowers offered a high volume of sucrose and the flower patch in which fl owers offered a low volume of sucrose solution, F (1 , 15 ) = 0.52 , p = 0.471 A Shapiro Wilk test found that a ll normality assumptions were met (W = 0. 846 , p = 2.2 ). Learning flight duration at each flower patch In order to assess whether bumblebees spent more time performing learning flights at either flower patch, t he total duration of learning flights for each bumblebee at both the five - flower patch containing flowers with a high volume of sucrose solution and the five - flower patch conta ining flowers with a low volume of sucrose solution was recorded from video footage and compared ( Fi gure 4 ). 710050472 18 I t is indicated that the bumblebees at the flower patch offering a low volume of sucrose in each of the five flowers showed longer learning flights than the bumblebees that performed learning flights at the flower patch offering a high volume of sucrose in each of the five flowers , and that there is a wider variation of learning flight durations for bumblebees at the low volume flower patch compared with the high - volume flower patch (Figure 4) Data was analysed using an independent samples t - test in order t o investigate whether there was a significant difference in learning flight duration bet ween the bumblebees at the flower patch containing flowers with the high volume of sucrose solution (0.5ml) and the low volume of sucrose solution (20μL). It was found that t here was a significant difference between the flower patch with the high volume of sucrose (M = 13.352 , SD = 8.347 ), and the Figure 4 : T he learning flight durations of bumblebees at the low - volume and high - volume flower patches . Shown are the median and lower/upper quartiles (box) with one outlier. The whiskers depict the highest and lowest values recorded, and the mean is depicted by an ‘x’. 710050472 19 low volume of sucrose (M = 25.644 , SD = 13.831 ), in ter ms of the duration of bumblebees’ learning flights at the flower patch , t ( 22 ) = - 2.547 , p = 0.018 , with learning flights being significantly longer at the low volume flower patch . A Shapiro Wilk test found that a ll normality assumptions were met (W = 0.965 , p = 0.546 ). Learning f light trajectories at each flower patch The trajectory / flight path length of learning flights was calculated for each individual bumblebee at both five - flower patches For each bumblebee, the first time they crossed each 2cm distance increment (up to 20cm) away from the flower that they fed from was recorded, and then averaged for all bumblebees at each patch to create an average trajectory length for each distance increment . The average learning flight trajectory was then plot ted and compare d between the flower patch in which flowers offered a low volume of sucrose solution, and the flower patch in which flowers offered a high volume of sucrose solution (Figure 5 ). Figure 5 : T he average learning flight trajecto ries for bumblebees at each flower patch when they first cross 2cm distance increments from the flower they originally fed from , including standard error bars. 710050472 20 It appears that bumblebees at the flower patch in which flowers offered a low volume of sucrose had a longer average trajectory length in their learning flights, compared to the flower patch in which flowers offered a high volume of sucrose (Figure 5). Data was analysed using a two - way mixed model ANOVA in order to determine whether there was a significant difference in the trajectory length of bumblebees’ learning flights between the flower patch in which flowers offered a low volume of sucrose and the flower patch in which flowers offered a high volume of sucrose. It was found that there was a significant difference in the length of learning flight trajectories between bumblebees performing a learning flight at the flower patch containing a low volume of sucrose and the flower patch containing a high volume of sucrose, F (1 , 8 ) = 12.22, p = < .001 , with the learning fli ght trajectory length of bumblebees being significantly longer at the low volume flower patch (M = 64.546, SD = 51.701), compared with the learning flight trajectory of bumblebees at the high volume flower patch (M = 43.008, SD = 41.487) A Shapiro Wilk test found that a ll normality assumptions were met (W = 0.863, p = 7.827). Discussion The aim of the present research was to determine whether the learning flights of bumblebees , and subsequently the spatial memories of a flower patch was significantly influenced by the volume of sucrose solution reward gained from the patch , investigating the difference in learning flights between bumblebees at an artificial five - flower patch that offered a high volume of sucrose reward from each flower, and bumblebees at an artificial five - flower patch that offered a low volume of sucrose reward from each flower. This was investigate d in terms of : i) how many flowers were visited after initially b eing released and drinking from one of the five flowers in the patch, ii) how much bumblebees at each flower patch looked at the flower that they originally fed from (flower - facing), iii) the duration of learning flights, and iv) the lengths of learning fl ight trajectories It was hypothesised that bumblebees at the five - flower patch in which artificial flowers offered a high volume of sucrose reward would form more comprehensive spatial memories of the patch , through learning flights being longer in