Thoughts on Information and Integration in Honey Bee Colonies
Total Page:16
File Type:pdf, Size:1020Kb
Review Thoughts on information and integration in honey bee colonies Thomas D. Seeley Section of Neurobiology and Behavior, Cornell University, Ithaca, NY 14853, USA (Received 6 June 1997; accepted 27 October 1997) Abstract - Solving the puzzle of colony integration in honey bees requires understanding how a worker bee acquires the information that she needs to decide correctly, moment-by-moment, what task to perform and how to perform it. To help us understand how the bees inside a beehive acquire this information, I share some thoughts about information flow within honey bee colonies. These thoughts are based on recent findings about how a colony works as a unified whole in gathering its food. © Inra/DIB/AGIB/Elsevier, Paris Apis mellifera / communication / honey bee / information / social behavior 1. INTRODUCTION foragers among flower patches - these are some of the puzzles whose solutions have gradually emerged from scientific stud- Of all the embodied in a mysteries ies. But many other aspects of the unity bee the is honey colony, perhaps greatest of colonies remain enigmatic and so draw how thousands of bees can work together us onward. with such coherence that a colony func- tions as a single, smoothly running, indi- In this article, I hope to advance our vidually purposeful entity. The mystery understanding of the functional integra- of colony integration has intrigued humans tion of honey bee colonies by sharing for hundreds of years and, little by little, some thoughts that have emerged from much has been revealed from the treasure analyzing how a colony gathers its food chest of the bee hive. The ability of a [reviewed in Seeley (1995)]. These colony to control its nest temperature, to thoughts all concern the information used choose a home site, and to distribute its by worker bees as each one decides, * Correspondence and reprints Tel.: (1) 607 539 7897; fax: (1) 607 254-4308; e-mail: [email protected] moment-by-moment, how she should bees are primarily chemical and mechan- behave to contribute to the common good. ical stimuli, since they must be easily per- (In this article, the word ’information’ ceived by bees in the darkness that pre- denotes simply knowledge obtained by a vails inside the hive. Table I lists the bee from its environment; it does not known signals of honey bees. We can see denote a quantitative measure of the reduc- that in each case a signal is a means tion of uncertainty about conditions, as in whereby one bee can convey to her nest- formal information theory.) The close con- mates information that helps them decide nection between information and integra- what to do next (e.g. attack an intruder, tion is made clear by noting that the gen- forage for pollen, or feed a nestmate) and eral problem of colony integration can be how to do it (e.g. sting the intruder here, framed in terms of two more specific prob- obtain pollen from flowers just outside the lems: 1) maintaining a proper distribution hive, or give me just a little food). of individuals among the various tasks Because signals are specialized to be performed within a colony, and 2) achiev- informative, they are to be ing coordination among the individuals likely unusually pertinent among the other sources of infor- working on each task. These can be mation with which they occur. However, thought of as the problems of between- it should be that a worker bee task and within-task coordination. From recognized a has access to and can the perspective of the individual bee, the receiving signal process much information besides what first problem is one of knowing what to she obtains from the signal. Each time she do, while the second is one of knowing a she also how to do it. Because the solutions to these perceives particular signal may register information regarding her partic- problems depend critically upon workers ular ’location’ in the nest dance floor, possessing adequate information to decide (e.g. brood nest, combs), her correctly what to do and how to do it, it honey particular ’time’ time of season of the is clear that much of the of (e.g. day, challenge her ’behavioral context’ lies in year), particular understanding colony integration the analyzing the acquisition and processing of (e.g. defending nest, tending brood, resting), and her particular ’social iden- information by the workers in a colony. tity’ (e.g. age, experience, physiological state). I would like to suggest that the information a bee in con- 2. SIGNAL INFORMATION processed by with a which I VERSUS CONTEXTUAL junction receiving signal — will call ’contextual information’ — is often INFORMATION extremely rich, for in principle it com- prises anything that a bee can recall or Within a bee there is honey colony, at the time. And I wish to stress extensive of the inter- perceive overlap reproductive that to focus on as sources of ests of hence it is not only signals individuals, surpris- information for bees would be for that bees have evolved folly, ing many special to do so would blind one to all but a small means for sharing information, i.e. for part of the total body of information used communicating. Following Lloyd (1983), by bees as they make behavioral decisions. I will use the term ’signal’ to denote any structure or behavior that has been molded Although we do not know the full scope by natural selection for the purpose of con- of information acquisition and integration veying information. (Note: the term ’sig- by worker bees in even one behavioral nal’, as used here, is synonymous with the context, it is clear that sophisticated infor- classical ethological term ’sign stimulus’.) mation processing by a worker bee receiv- The signals that have evolved in honey ing a signal is not just a possibility, but a reality. Consider, for example, the case of will see various responses to the dance bees standing on the combs just inside the signals produced by the successful for- hive entrance during a period of good agers. Some bees will press in closely weather when the colony’s foragers are behind a dancing bee, then follow her busily bringing home load upon load of throughout several circuits of her dance, nectar, pollen and water. These bees are and finally turn away and scurry out the surrounded by foragers excitedly adver- hive entrance. Others will start to follow a tising with waggle dances the various dancing bee closely, seemingly with great sources of their foraging success. If one interest, but after a few seconds they will watches several of the unemployed bees, leave the dancer and crawl away. Still oth- individually, for several minutes each, one ers will completely ignore the dancing bees, and instead will seek to unload nec- their colony’s need for energy, protein or tar or water from the incoming foragers. water. Such sophistication by the dance These different responses to the waggle followers could be important as one of the dance signal reflect differences in the con- mechanisms underlying the ability of a textual information possessed by the var- colony to keep its foragers distributed ious bees on the dance floor (Seeley and among the tasks of nectar collection, Towne, 1992). The bees that follow a pollen collection and water collection in dancer closely and extensively are most accordance with its forage needs [dis- probably bees that possess prior foraging cussed further in Seeley (1995)]. Like- experience, but have abandoned their old wise, an analysis of the response of worker food source and so seek a new one. The bees to the queen bee’s mandibular bees that follow a dancer only briefly are pheromone signal may reveal dramatic probably also experienced foragers, but changes in behavior in relation to changes ones that have not yet abandoned their in contextual information about the prior food source. Their behavior suggests colony’s state and time of year. For exam- that they are seeking news of the renewal ple, most of the time workers respond to of their previous day’s food source, hence this signal by refraining from rearing addi- when each of these bees realizes that the tional queens, but under the influence of a dancing bee is advertising a different certain set of inputs of contextual infor- source, she turns away as if quickly real- mation (certain time of the year? certain izing, ’she is not advertising my flower abundance of brood? certain level of patch’. Finally, the bees that ignore the crowding in the hive?) the workers may foragers’ dances and instead relieve them alter their response and begin rearing of their loads are bees that have not yet queens in preparation for swarming. The begun foraging and instead are functioning workers may receive at all times the chem- as receiver bees (Seeley, 1989). These ical signal indicating that their queen is three types of bees perceive different sets alive and well, but in one context they of contextual information when encoun- may decide it is best to not rear queens tering the waggle dance signal, conse- while in a different context they may quently they show three kinds of response decide that rearing queens is the best to this one type of signal. course of action. It is likely that what I have written so The general message here is that as we far scarcely begins to express the com- seek to understand how individual work- plexity of the information acquisition and ers decide how to behave to foster colony processing that is performed by a worker functioning, we are right to place special bee, whose sensory system is acute, hive emphasis on the analysis of signals, environment is variable, and behavior is because many of them have evolved flexible.