Thoughts on information and integration in honey colonies Thomas D. Seeley

To cite this version:

Thomas D. Seeley. Thoughts on information and integration in colonies. Apidologie, Springer Verlag, 1998, 29 (1-2), pp.67-80. ￿hal-00891480￿

HAL Id: hal-00891480 https://hal.archives-ouvertes.fr/hal-00891480 Submitted on 1 Jan 1998

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés. 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 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 . 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. So let us consider further this specifically for spreading information same example, focusing on the bees that about a colony’s labor needs. But at the extensively follow a dancing bee and are same time, we must not overlook the fact aroused to leave the hive. A closer analy- that when a worker bee processes the sis of the response of these bees to the information in a signal, she often inte- dance signal may reveal large changes in grates the signal information with a large relation to contextual information about amount of contextual information. And as the colony’s forage needs. Specifically, the above discussion attempts to show, these bees may selectively follow dances this contextual information can strongly advertising either nectar sources, pollen influence a worker’s response to a signal. sources, or water sources, depending on Indeed, a worker may even choose to ignore a signal in light of certain contex- In table II, I have assembled a list of tual information. Hence it seems likely known or suspected cues used by honey that a full understanding of the way that bees. This list of cues is long, approxi- any particular signal contributes to the mately twice as long as the list of signals smooth functioning of a honey bee colony in table I. I hasten to add, however, that will also require a careful analysis of the sometimes it is not clear whether an item contextual information used by the work- should be categorized as a signal or as a ers in association with this signal. cue, so some of the things listed as cues may prove to be signals (and a few of the items listed as signals in table I may prove 3. A MULTITUDE OF CUES to be cues). For instance, I have placed the odor of dead bees in the cue category, In the preceding section I have argued since it is likely that this odor is an inci- that much of the information with which dental by-product of the decomposition worker bees make behavioral decisions of a dead bee, but it may turn out that bees comes from sources other than specially possess exocrine glands whose secretions evolved signals. I have not, however, dis- are released upon death to provide a con- cussed explicitly what these alternative spicuous marker of a dead bee. If so, then sources of information are. In principle, the odor of dead bees would be a signal, they are anything that a worker bee can not a cue. Conversely, I have placed the perceive. Following Lloyd (1983), I will brood pheromone in the signal category, use the term ’cue’ to denote any informa- since it is likely that the immature bees tive variable perceived by a bee that has have evolved a special signal of their pres- not been shaped by natural selection ence, but it may be that this pheromone specifically to convey information, but is simply an automatic by-product of the that instead conveys information inciden- biochemical processes underlying the tally. Cues and signals are, therefore, log- growth and maturation of bees. If so, then ically distinct categories of information- the brood pheromone would be a cue, not bearing variables. What, then, is the a signal. Much of the ambiguity about cues relationship between cues and contextual versus signals reflects the fact that many information? The two terms are not syn- signals have evolved from cues and have onymous. Although cues are the source retained their original information content of most contextual information, there are even though they have been molded by many cues used by worker bees not in natural selection to express this informa- association with a communication signal. tion more strongly and precisely than was Hence, we must view cues more broadly originally the case. Of course, there are than just as sources of contextual infor- also many cues and signals whose status is mation. Consider, for example, the level of not ambiguous. Nest temperature, ragged atmospheric carbon dioxide in a bee hive. cell cappings and unpacked pollen loads This is an important cue for worker bees, are all certainly cues, whereas the waggle one that indicates the need for hive venti- dance, tremble dance and Nasonov gland lation (Seeley, 1974), but there is no evi- pheromone are all undoubtedly signals. dence that it influences how bees respond to any communication signal. Therefore Although it is not always clear whether it is a cue, but apparently it is not a source a particular source of information inside a of contextual information. In short, most if bee hive is a signal or a cue, I believe that not all sources of contextual information the numerical predominance of cues over are cues, but not all cues are sources of signals, suggested by the relative lengths contextual information. of tables I and II, will prove correct. In

other words, I believe that a large major- need for heating or cooling activities, and ity of the information sources used by bees is a product of the thermoregulatory to decide what to do (and how to do it) behaviors (heat production, ventilation, will be cues rather than signals. My belief water collection, etc.) of hundreds or thou- is based on two insights. The first is that sands of colony members. pathways of information flow will evolve more readily when they involve cues rather than signals. The evolution of cuing 4. INDIRECT INDICATORS involves only the formation of an adap- OF LABOR SUPPLY-DEMAND tive response to a pre-existing stimulus RATIOS (the cue), whereas the evolution of sig- involves the modification naling adaptive The need for additional labor devoted to of both a stimulus (the signal) and a any given task depends ultimately on the response. All else being equal, then, we supply of and demand for the products of should expect more forms of cuing than this labor. Thus, a need for more of signaling in colonies. The second colony’s labor devoted to water collection depends insight is that the process of colony inte- on the relative rates of water collection gration is largely a matter of information and water consumption. Likewise, a flow from colony to individual, so that colony’s need for more labor devoted to each individual can adjust its behavior in comb on the of accordance with the activities of the other building depends supply empty comb relative to the demand for colony members. It seems likely that this comb (in which to store nectar and colony-to-individual information flow will empty rear brood). One suppose, there- occur via cues because might mainly any group- that bees inform themselves of their level indicator which individuals monitor fore, labor needs somehow sens- for their actions is to colony’s by co-ordinating likely both the of and the demand for be a of the combined activi- ing supply by-product the products of the different forms of labor ties of a group (a cue) rather than a group- in a nectar, water, level that has evolved colony: pollen, propolis, phenomenon specif- clean cells, drone cells, fed brood, warmed ically for information expression (a signal). brood, capped pupae, groomed bees, and One striking example of a group-gener- so forth. For one suppose ated cue is the time that a nectar example, might forager that a nurse bee senses the need for addi- just back from a foraging trip spends tional labor devoted to cell cleaning by searching for a receiver bee to take her measuring the rate at which her nestmates load of nectar. This cue provides the for- are cleaning dirty cells (the supply) and ager with information about her colony’s the rate at which young bees are emerging need for a higher rate of nectar intake (See- and so producing dirty cells (the demand), 1989). On the time that a for- ley, average, and noting whether the latter exceeds the ager must spend searching for a receiver is former. One might also imagine that a for- determined by the relative rates at which ager bee senses the need for additional arrive in the area foragers unloading ready the rate of to nectar and receivers arrive in pollen foraging by assessing give up pollen collection by fellow foragers and the unloading area ready to take in nectar the rate of pollen consumption by the (Seeley and Tovey, 1994). Search time is, nurse bees, and noting any discrepancy therefore, a of the activities consequence between these two rates. of all the nectar foragers and all the nectar receivers in a colony. Another example of The approach of directly determining a group-generated cue is the temperature the supply of and demand for various labor inside the hive. This is an indicator of the products would provide workers with a precise picture of their colony’s labor of bee labor. Thus bees may respond to needs, but almost certainly it is not the the prevalence of dirty or clean cells, hun- usual means by which workers assess their gry or well-fed larvae, and cool or hot colony’s labor needs. At least to date, there brood as indices of the need to clean cells, is no evidence that bees make direct mea- feed larvae and warm brood, respectively. surements of the supply of and demand for various goods and services in a bee Sometimes, however, the indirect indi- hive. Probably the reason for this is that cator of a particular labor need in a bee it is exceedingly difficult for a worker bee colony does not have such an obvious rela- to accurately assess supply and demand tionship to the changes in the supply of as independent variables. Consider, for and demand for this labor. One example of example, the difficulty faced by a nurse this is the summary indicator of the need bee that is seeking to determine the need for additional labor devoted to water col- for more cell cleaning if she had to mea- lection (Lindauer, 1954; Kühnholz and sure the overall rate of cell cleaning and Seeley, 1997). Each bee engaged in water the overall rate of brood emergence within collection decides whether or not to recruit her colony. Likewise, imagine the diffi- additional bees to this task not by sensing culty faced by a forager bee that is seeking a rise in the level of water in the colony, to determine the need for more pollen for- but by sensing the difficulty of unloading aging if she had to measure her colony’s water to receiver bees. How exactly this total rates of pollen collection and pollen unloading difficulty is sensed remains consumption. unclear, for there are several variables of the unloading process that change simul- It now seems clear that instead of mon- taneously when a colony’s water need itoring the variables of supply and demand changes and it remains uncertain which directly, bees monitor variables that pro- one(s) the bees monitor. The critical vari- vide indirect summaries of the able may be something only distantly supply-demand ratios for the various related to the fundamental variables of goods and services produced in a colony. water collection and water consumption, This means of control is analogous to the such as how long a water collector must way a governor on a steam engine works. search to find a receiver, or how many It detects neither the supply of work being times a water collector experiences performed by the engine (i.e. the power unloading rejections from receiver bees output of the engine) nor the demand for seeking nectar rather than water. A sec- work by the engine (i.e. the load on the ond example of the way that a summary engine), but instead responds to a variable indicator of a labor supply-demand ratio that reliably summarizes the work sup- can have an obscure relationship to the ply-demand situation, namely the engine underlying variables of supply and speed. Any time there is a change in the demand is the indicator of need for pollen supply of or the demand for work, the foraging [Camazine (1993); see also the engine speed automatically changes. The discussion in Seeley (1995)]. Although governor responds to changes in engine the level of the pollen stores within a hive speed, increasing or decreasing the flow provides an accurate index of the relative of steam to maintain the engine speed and rates of pollen consumption and pollen so keeps the work supply matched to the collection, it seems that the foragers rely work demand. In the case of the bees, on another, less obvious indicator of the these indirect, summary indicators are need for more pollen foraging. Prelimi- often simply the levels of the different nary evidence suggests that this indicator substrates or products of the various forms is the level of protein hunger that a for- ager feels, for it appears that if a colony’s third bee might decide to plane this wall to rate of pollen consumption exceeds its rate a finer thickness, responding to the con- of pollen collection for long, the nurse struction of the previous bee. A second bees start to give the foragers little pro- example of information flow through the teinaceous food and soon these bees begin shared environment comes from the pro- to feel hungry for protein. cess of nest temperature control. A colony maintains the central broodnest of The lesson here is that region general although its nest at 32-36 °C in the face of ambient the need for labor in each task within a that from -20 to bee hive is determined temperature may range fundamentally by 40 °C. The coordination of the bees between the of and discrepancy supply involved in and cooling the nest demand for this labor, it that bees heating appears occurs with little or no direct communi- assess their labor needs without colony’s cation; each bee to the broad of rela- responds tempera- knowledge supply-demand ture of her immediate environment it that by tionships. Instead, appears they rely appropriately heating it (by making intense on indirect, summary indicators of dis- isometric contractions of her flight mus- crepancy between supply and demand. cles) or cooling it (by fanning her wings to Sometimes these indicators are summary draw cooler air into the area) (Heinrich, obvious - such as a buildup in the sub- 1985). Hence, the temperature of the air strate or the product of a particular form of and comb inside a hive provides an indi- labor - but other times are obscure. they rect means of information flow between the bees engaged heating or cooling their colony’s nest. 5. THE SHARED ENVIRONMENT

AS A PATHWAY OF The use of the shared environment as a INFORMATION FLOW pathway of information flow between bees has several important advantages over Information can pass between colony direct means of information transfer. One members in two general ways: directly is that it allows easy asynchronous trans- from one bee to another, or indirectly fer of information between the sender and through some component of the bees’ the receiver of the information. Many shared environment (the combs and their forms of direct communication between contents, the shared food and other sub- bees require precise synchronization of stances, the hive atmosphere, and the nest- signal production and reception (since mates themselves). An example of the lat- most signals are ephemeral), and may even ter process is the passage of information in require that sender and receiver have a the process of comb building. The con- certain spatial relationship at the moment struction of a particular cell in a beeswax of signal transmission. For example, a bee comb involves numerous bees, yet these trying to obtain information from a nest- bees never need to come together and mate performing a waggle dance must exchange information directly. Their place her antennae within a few millime- building actions are completely and effi- ters of the dancer at the moments she pro- ciently co-ordinated by information duces her waggle runs (Kirchner, 1993). embodied in the structure of the partially Moreover, there is now evidence sug- completed cell. Thus one bee might begin gesting that successful acquisition of the a cell wall by depositing a small ridge of dance information depends on the dance beeswax; a second bee might continue follower aligning herself directly behind sculpting the wall, guided by the shape of the dancer at times of waggle run produc- the wax ridge left by the first bee; and a tion (Judd, 1996). This may explain, at least in part, why waggle dancing bees 6. MOBILE SENDERS AND often must produce 40 or more waggle RECEIVERS OF INFORMATION runs to get just one nestmate successfully recruited to a food source (Seeley and In analyzing the sources of informa- Towne, 1992). Such precise temporal and tion used by bees in deciding how best to spatial co-ordination between communi- behave for the common good, I think it is cating individuals is not required when critically important to keep in mind that the information passes between them bees are highly mobile creatures inside through the shared environment. their hives. This mobility means that bees The exploitation of the shared envi- are likely to have sophisticated adapta- ronment as an information channel also tions involving movements that help them has the important feature of providing easy send and receive information. Let us first passage of information from a group to consider the matter of sending informa- an individual. This will occur whenever tion. Already it is clear that to understand an individual responds to the environ- fully how a particular signal contributes mental effects of the group. Good exam- to the smooth functioning of a colony we ples of this include a comb building bee must know the spatial pattern of the sig- responding to the results of prior comb nal’s production. This is because where a construction by her nestmates, a ther- signal is produced within a hive strongly moregulating bee responding to the nest influences which colony members will temperature produced by her nestmates, receive the signal. One sees this if one and a foraging bee responding (though compares the spatial distributions and the perhaps indirectly) to the level of pollen audiences of waggle dances and tremble stored in the hive by her fellow foragers. dances (figure 1). Waggle dances are per- And as mentioned previously, group-to- formed just inside the hive entrance, where individual information flow is central to they are encountered mainly by unem- the process of colony integration. ployed foragers that have travelled to this location specifically to obtain information that the young and middle age bees in a about foraging opportunities (von Frisch, colony, the ones that work primarily inside 1967; Seeley, 1995). Tremble dances, the hive — cleaning cells, tending brood, however, are performed throughout the storing food, and so forth — do travel about hive and so are encountered by all types of a good deal, so it seems likely that they bees, including many inactive bees. Con- obtain and integrate information from var- sequently a broad range of bees is alerted ious locations within the hive, but we still to the colony’s need for additional receiver know very little about the phenomenon of bees, which is adaptive since during a patrolling. How wide-ranging is a strong nectar flow a colony may need to patrolling bee’s reconnaissance? What is devote up to half of it’s members to the the temporal pattern of patrolling? What task of processing nectar (Seeley et al., signals and cues are registered during 1996). patrolling? Are patrolling bees selective about the information that Another illustration of the idea that the they register, as a function of the tasks currently being movement of worker bees can be patterns These are but a few of the for of infor- performed? adaptations improved sending questions that must be answered before mation is what one sees with messenger we will understand what information a bees. These bees the pick up queen’s patrolling bee is actually acquiring. And pheromone and then travel about the given that information collection is a crit- broodnest actively dispersing this chemi- ical component of a worker bee’s deci- cal signal of the queen’s presence sion-making process as she chooses what (Velthuis, 1972; Seeley, 1979; Ferguson to do and how to do it, we must conclude and Free, 1980). This pattern of messenger that our lack of knowledge about patrolling bee movement can be viewed as a means remains a obstacle to our of telecommunications major solving achieving rapid the mystery of colony integration. from one sender (the queen) to many receivers (the workers). Likewise, it seems clear that the of buzz-runners scrambling ACKNOWLEDGMENTS over the combs at the moment of swarm departure from the hive is an adaptation The research that inspired the thoughts pre- to send quickly and broadly the message sented here was National Sci- ’let’s I that supported by go!’ (Martin, 1963). suspect ence Foundation grants BNS-8606778, BNS- it is generally true that bees produce their 8916006, IBN92-21150 and IBN96-30159, for signals with particular spatial distributions which I am most grateful. inside the hive, and that this enhances the effectiveness of the signals, but this remains a largely unexplored aspect of Résumé - Réflexions sur l’information honey bee communication. et l’intégration dans les colonies The mobility of worker bees probably d’abeilles, Apis mellifera L. Résoudre le also underlies adaptations for the acqui- mystère de l’intégration de la colonie chez sition of information. In principle, the les abeilles mellifères implique de com- capacity for independent travel about the prendre comment une ouvrière acquiert hive enables the members of a colony to l’information nécessaire pour décider cor- personally gather information over a wide rectement, instant après instant, quelle area inside the hive. Lindauer (1952) aptly tâche accomplir et comment. À ce sujet named this process ’patrolling’. It remains je fais part de quelques réflexions. Pre- unclear, however, exactly how important mièrement, lorsqu’on étudie le flux patrolling is to bees. Lindauer documented d’informations dans les colonies, il faut mettre l’accent sur l’analyse des signaux, développé des adaptations sophistiquées c’est-à-dire des variables porteuses d’infor- qui impliquent le mouvement et les aident mation qui se sont développées spécifi- à envoyer et recevoir l’information. quement pour transmettre l’information. « Patrouiller », courir çà et là dans la ruche Nous ne devons pourtant oublier que pour collecter l’information, est peut-être lorsqu’une abeille traite l’information la plus importantes de ces adaptations, exprimée par un signal, elle intègre sou- mais nous savons à ce jour peu de choses vent cette information dans tout un concernant ce moyen visiblement crucial contexte informationnel. C’est pourquoi de collecte de l’information. © Inra/DIB/ nous devons aussi analyser l’information AGIB/Elsevier, Paris contextuelle utilisée par les abeilles en association avec chaque signal. Deuxiè- Apis mellifera / communication / infor- mement, une grande partie de l’informa- mation / comportement social / inté- tion qui permet à l’abeille de prendre des gration colonie décisions provient d’indications, c’est-à- dire de variables porteuse d’information qui ne se sont pas développées spécifi- Zusammenfassung - Überlegungen quement pour transmettre l’information, über Information und Integration bei mais qui l’ont fait incidemment. La Honigbienenvölkern. Ein Teil der Lösung majeure partie des sources d’information des Rätsels der Integration der Honigbie- de la ruche sera probablement des indica- nenvölker liegt in dem Verständnis, wie tions plutôt que des signaux. Si l’on com- die einzelne Arbeiterin die adaequate pare les signaux et les indications dans la Information aufnimmt, anhand derer sie colonie d’abeilles (tableaux I et II), on richtig und von Augenblick zu Augen- s’aperçoit que ces dernières surpassent blick entscheidet, welche Aufgabe sie ver- largement les premiers. Troisièmement, richtet und wie sie dies tut. Hier möchte le besoin d’une colonie d’abeilles en un ich einige Gedanken mitteilen, die uns das matériel ou un service donnés dépend en Verständnis der Informationsaufnahme dernier ressort du rapport entreson offre et durch die Bienen innerhalb des Volkes sa demande. Pourtant, plutôt que de sur- erleichtern sollen. Zunächst halte ich es veiller les variables d’offre et de demande, für das Studium des Informationsflusses in les abeilles surveillent habituellement den Bienenvölkern für angemessen, eine d’autres variables qui résument les rap- besondere Betonung auf die Analyse von ports d’offre/demande, car souvent ces Signalen im Sinne von spezifisch zu die- indicateurs condensés sont à la fois d’une sem Zweck evolvierten informationstra- perception facile et des sources d’infor- genden Variablen zu legen. Allerdings mation très fiables. Quatrièmement, une müssen wir uns in diesem Zusammenhang large part du flux d’information entre les klar machen, daß eine Biene, sobald sie membres d’une colonie parvient indirec- die in einem Signal enthaltene Informa- tement par l’intermédiaire d’éléments de tion verarbeitet, diese oftmals mit einer l’environnement commun plutôt que direc- grossen Menge kontextbezogener Infor- tement d’une abeille à l’autre. L’utilisa- mation verrechnet. Wir müssen daher tion de l’environnement commun comme ebenso die Information aus dem Kontext canal d’information permet une commu- berücksichtigen, die von den Arbeiterinnen nication asynchrone facile et un passage in Verbindung mit jedwedem Signal aisé de l’information du groupe à l’indi- benutzt wird. Zweitens entstammt sehr vidu. Et cinquièmement, étant donné la viel der Information, die die Arbeiterin- mobilité des abeilles à l’intérieur de la nen zu einem bestimmten Verhalten ver- ruche, celles-ci ont vraisemblablement anlaßt, aus Hinweisen (’cues’), d.h. aus informationstragenden Variablen, die sich von Information. © Inra/DIB/AGIB/Else- nicht spezifisch zur Übertragung von vier, Paris Information entwickelt haben, sondern die dies nur beiläufig tun. Höchstwahr- Apis mellifera / Kommunikation / scheinlich wird sich herausstellen, daß der Honigbienen / Information / Sozialver- überwiegende Teil der Informationsquel- halten len einer Innendienst-Bienenarbeiterin eher aus Hinweisen als aus Signalen besteht. Ein Vergleich der bekannten Sig- REFERENCES nale und der Hinweise in Honigbienen- völkern (Tabelle I und II) zeigt auf, daß Breed MD (1985) How honeybees recognize their die letzteren bei weitem häufiger sind als nestmates: a re-evaluation from new evidence. Bee World 66, 113-118 die ersteren. Zum Dritten hängt der Bedarf Camazine S (1993) The of an einem bestimmten Material oder Dienst regulation pollen foraging by honey bees: how foragers assess the colony’s in einem Honigbienenvolk letztendlich need for pollen. Behav Ecol Sociobiol 32, von dem Verhältnis zwischen dessen 265-272 Bereitstellung und der Nachfrage ab. Darchen R ( 1968) Le travail de la cire et la con- anstelle und struction dans la ruche. In: Traité de Biologie de Allerdings, Bereitstellungs - l’Abeille, Vol. 2 (R Chauvin, ed), Masson, Paris, Nachfragevariablen direkt zu messen, 241-331 erfassen Bienen zumeist weitere Varia- Ferguson AW, Free JB ( 1980) Queen pheromone blen, die das Verhältnis zwischen Bereit- transfer within honeybee colonies. Physiol Ento- stellung und Nachfrage zusammenfassen. mol 5, 359-366 Der Grund ist, daß solche zusammenfas- Free JB (1967) The production of drone comb by senden Indikatoren oft einerseits leicht honeybee colonies. J Apic Res 6, 29-36 wahrzunehmen sind, andererseits aber Frisch K von (1967). The Dance Language and Ori- entation of Bees. Harvard Univ Press, Cambridge auch hochzuverlässige Informationsquel- Heinrich B (1985) The social of tem- len darstellen. Viertens wird physiology ein großer perature regulation in honeybees. In: Experi- Teil des Informationsflusses zwischen den mental and Sociobiology (B Koloniemitgliedern eher indirekt über Hölldobler, M Lindauer, eds), Sinauer, Sunder- land, 393-406 Komponenten aus der gemeinsamen Heran H (1952) Untersuchungen über den Tempera- vermittelt als direkt von einer tursinn der Honigbiene (Apis mellifica) unter Biene zur anderen. Die Nutzung der besonderer Berücksichtigung der Wahrnehmung gemeinsamen Umgebung als Informati- strahlender Wärme. Z vergl Physiol 34, 179-206 onskanal ist eine leichte Möglichkeit zur Hess C von (1916) Messende Untersuchung des Lichtsinnes der Biene. Arch 163, Kommunikation und zur ges Physiol asynchroner 289-320 von Information von der Übermittlung Judd TM (1995) The waggle dance of the honey bee: zum Individuum. Und fünftens Gruppe which bees following a dancer successfully besteht in Anbetracht der Mobilität der acquire the information? J Insect Behav 8, Bienen eine hohe Wahrscheinlichkeit für 343-354 die Entstehung höchstentwickelter Anpas- Kirchner W (1993) Acoustical communication in honeybees. Apidologie 24, 297-307 sungen, die sowohl für das Aussenden als N (1970) Factors the of auch das von Information eine Koeniger determining laying Empfangen drone and worker eggs by the queen honeybee. Bewegungskomponente enthalten. Die Bee World 51, 166-169 wichtigste dieser Anpassungen ist mögli- Koeniger N, Veith HJ (1983) Glyceryl-1,2-diole- cherweise das ’Patrouilliern’, das Umher- tate-3-palmitate as a brood pheromone of the bee 39. laufen von Arbeiterinnen zum Informati- honey (Apis mellifera L.). Experientia 1051-1057 onsgewinn, allerdings wissen wir bislang Kühnholz S, Seeley TD ( 1997) The control of water nur sehr wenig über dieses offensichtlich collection in honeybee colonies. Behav Ecol wesentliche Mittel für das Ansammeln Sociobiol 41, 407-422 Lensky Y, Slabezki Y (1981) The inhibiting effect of cation regulating foraging on two time scales. the queen bee (Apis mellifera L.) foot-print Anim Behav 34, 377-385 on the construction of pheromone swarming Seeley TD (1974) Atmospheric carbon dioxide reg- queen cups. J Insect Physiol 27, 313-323 ulation in honey-bee (Apis mellifera) colonies. Lindauer M (1952) Ein Beitrag zur Frage der Arbeits- J Insect Physiol 20, 2301-2305 im Bienenstaat. Z teilung vergl Physiol 34, Seeley TD (1979) Queen substance dispersal by mes- 299-345 senger workers in honeybee colonies. Behav Ecol Lindauer M (1954) Temperaturregulierung und Sociobiol 5, 391-415 Wasserhaushalt im Bienenstaat. Z vergl Physiol Seeley TD (1989) Social foraging in honey bees: 36, 391-432 how nectar foragers assess their colony’s nutri- Lloyd JE (1983) Bioluminescence and communica- tional status. Behav Ecol Sociobiol 24, 181-199 tion in insects. Annu Rev Entomol 28, 131-160 Seeley TD (1992) The tremble dance of the honey Martin H, Lindauer M (1966) Sinnesphysiologische bee: message and meanings. Behav Ecol Socio- Leistungen beim Wabenbau der Honigbiene. Z biol 31, 375-383 vergl Physiol 53, 372-404 Seeley TD (1995) The Wisdom of the Hive. Harvard Martin P (1963) Die Steuerung der Volksteilung Univ Press, Cambridge beim Schwärmen der Bienen. Zugleich ein Seeley TD, Morse RA (1976) The nest of the honey Beitrag zum Problem der Wanderschwärme. bee (Apis mellifera L.). Insectes Soc 23, 495-512 Soc 13-42 Insectes 10, Seeley TD, Morse RA (1978) Nest site selection by Milum VG (1947) Grooming dance and associated the honey bee. Insectes Soc 25, 323-337 activities of the Illinois Acad honeybee colony. Seeley TD, Tovey CA (1994) Why search time to Sci Trans 40, 194-196 find a food-storer bee accurately indicates the Park OW (1925) The storing and ripening of honey relatives rates of nectar collection and nectar pro- by honeybees. J Econ Entomol 18, 405- 410 cessing in honey bee colonies. Anim Behav 47, Parker RL (1926) The collection and utilization of 311-316 pollen by the honeybee. Mem Cornell Agric Exp Seeley TD, Towne WF (1992) Tactics of dance Sta 98, 1-55 choice in honey bees: do foragers compare Pratt SC Condition-depend timing of comb con- dances? Behav Ecol Sociobiol 30, 59-69 struction by honeybee colonies: experimental Seeley TD, Kühnholz S, Weidenmüller A (1996) tests of an optimization model (submitted) The honey bee’s tremble dance stimulates addi- Pratt SC, Kühnholz S, Seeley TD, Weidenmüller A tional bees to function as nectar receivers. Behav (1996) Worker piping associated with foraging in Ecol Sociobiol 39, 419-427 undisturbed queenright colonies of honey bees. Simpson J (1973) Influence of hive-space restric- Apidologie 27, 13-20 tion on the tendency of honeybee colonies to rear Ratnieks FLW (1995) Evidence for a queen-pro- queens. J Apic Res 12,183-186 duced egg-marking pheromone and its use in Simpson J, Cherry SM (1969) Queen confinement, worker policing in the honey bee. J Apic Res 34, queen piping and swarming in Apis mellifera 31-37 colonies. Anim Behav 17, 271-278 Renner M, Vierling G (1977) Die Rolle des Taschen- Velthuis HHW (1972) Observations on the trans- drusenpheromons beim Hochzeitsflug der mission of queen substances in the honey bee Bienenkönigin. Behav Ecol Sociobiol 2, 329-338 colony by the attendants of the queen. Behaviour Rinderer TE (1982) Volatiles from empty comb 41, 105-129 increase hoarding by the honey bee. Anim Behav Visscher PK (1983) The honey bee way of death: 29, 1275-1276 necrophoric behaviour in Apis mellifera colonies. Sandeman DC, Tautz J, Lindauer M (1996) Trans- Anim Behav 31, 1070-1076 mission of vibration across honeycombs and its Winston ML (1987) The Biology of the Honey Bee. detection by bee leg receptors. J Exp Biol 199, Harvard Univ Press, Cambridge 2585-2594 Winston ML, Slessor KN (1992) The essence of roy- Schneider SS, Stamps JA, Gary NE (1986) The alty: honey bee queen pheromone. Am Sci 80, vibration dance of the honeybee. I. Communi- 374-385