This chapter will summarize and interpret that evidence on honey bee communication which has been gathered since publication of the comprehensive summary by Ribbands (1953, available as a reprint in 1964). The interpretation included herein differs markedly from that found in standard reference sources, but, hopefully, any disagreements with content will provide a firm basis for constructive experimentation. [See Lindauer, (1967) for a summary of the conventional interprétation.]
A prime difficulty in interpretation of experimental results arises from the different opinions held by various investigators on what constitutes acceptable evidence in support of a conclusion. Indiscriminate use of the terms “proof” and “prove” (or any equivalent term) has led to much confusion in the field of bee communication. In evaluating the evidence while writing this chapter, less credence has been given to subjective statements about what may occur during a communication act than to those statements which are based upon data (when an adequate description of experimental method is available). Similarly, less weight has been given to statements based on simple correlations than to statements based upon more direct evidence—that is, whether or not certain steps have been followed (measuring, dissecting, and imitating signals) during a study of communication after an initial deduction has been made that a signal(s) has influenced an animal’s behavior. In addition, this chapter is written with the assumption that the gathering of direct evidence for only part of a complex act of communication does not constitute a convincing demonstration of the entire communication act; evidence for a part does not indicate an understanding of the whole.
Students of bee communication have generally concentrated their attention on the identification of signals produced by bees and on the behavior observed after generation of these signals. Not enough attention, however, has been directed toward determining which signal or combination of signals is actually used by the bees. While correlations, such as those generally obtained, can suggest possibilities for explaining communication processes, they can never serve as substitutes for more direct evidence. It is assumed that ever better evidence can and must be obtained in a continuing study of each communication process and that any endeavor is only an interim report.
The writing of this chapter would have been virtually impossible without the aid of Apicultural Abstracts (a publication of the Bee Research Association). These abstracts permit one to consider the contributions of several authors on any one topic and are useful to look at before one turns to the contents of the original papers. The overwhelming number of papers on bees unavoidably leads to a good deal of selection; therefore, many works have not even been mentioned which do in fact contribute information pertinent to this chapter. This chapter also generally omits citation of those papers already summarized in Ribbands’ book on the assumption that copies of this book are generally available to all concerned.
ATTRACTION AND MATING
Replacement of a lost or failing queen (supersedure) and production of queens prior to the issue of a swarm appear to involve similar signals. The differences in communication during these two processes will be treated later. In either case, within ten days after the first queen emerges and disposes of her rivals in the colony, she is ready for a mating flight. Exactly when this occurs depends largely upon weather conditions, but prior to the actual mating flight she may make several short orientation flights and thereby learn the position of the colony. The distinctive hive odor and landmarks which figure so prominently in the orientation of bees to their particular colony are possibly learned on these orientation flights. Taber (1954), however, reports that 40 per cent of queens studied left directly on mating flights without having made prior orientation flights.
On the afternoon of the mating flight, general hive activity rises to a higher pitch than normal, and just before the queen leaves the hive, one can notice the workers throughout the colony milling about. Aggressive behavior and special feeding of the queen by workers reportedly precede queen departure (Allen, 1956; Hammann, 1957; Ruttner, 1957). Eventually, what resembles a small swarm leaves the hive via its entrance; the queen then leaves this group and flies off by herself. Within one-half to one hour she returns, generally with a “mating sign” still attached to the tip of her abdomen. Although she may have mated enough on this one flight to have a lifetime supply of sperm, queens usually take several mating flights before laying (Alber et al., 1955; Woyke and Ruttner, 1958).
What transpires during this queen flight is a difficult problem to study, since mating apparently occurs too high to be seen. Although there is by no means universal agreement on this point, certain sites appear to be drone congregation areas which queens eventually reach on their mating flights (see Zmarlicki and Morse, 1963; Butler and Fairey, 1964; and Ruttner and Ruttner, 1965). Some of these areas seem to be remarkably stable and persist in the same location for several successive years. A clearing, low hilltop, or some secluded spot free from high winds may function in such a capacity, but the precise requirements for such an area are not yet known.
Once one has seen such an area, however, he can have little doubt that drone congregation areas exist. The sound of thousands of drones just out of sight overhead (with the aid of binoculars they can be seen cruising about, once a general sound source has been located) resembles that of a swarm of bees in flight, but there is no precise center to this sound. Drones in these groups apparently come from several hives in the area. Small birds, dragonflies, and butterflies in the area are very likely to be pursued by a number of drones if they venture too high. Similarly, a stick or stone thrown high into the air results in a “comet” of drones pursuing the object until it hits the ground. Sometimes the drones even go into the grass as a result of their intense pursuit of an object and can be caught with a net. Ruttner and Ruttner (1965) investigated several of these areas over a period of time. They found that a queen tethered to a gas-filled balloon failed to attract drones when moved beyond a certain point with respect to the apparent center of the congregation area.
It would appear, then, that it is no longer a question of whether such congregations exist, but rather it is a problem of determining the factors responsible for such mating aggregations. Conceivably, a circular flight pattern by a virgin queen could result in a number of drones heading upwind and congregating about her—attraction of drones to queens tethered to balloons regardless of the location in which this was attempted supports this possibility (Butler and Fairey, 1964). Alternatively, especially in areas with varied topography, drones could be attracted to distinctive geographic sites—persistence of congregations for several weeks or under similar conditions in different years supports this notion. In the latter case, queens could be attracted either by the drone congregation itself or to the same site for the same reasons as the drones.
Formation of a mating aggregation has special advantages in increasing the probability of queens being found by drones and in providing some protection against predators. That is, predators would be unlikely to find a queen among the thousands of flying drones present in the same locality.
The factors responsible for mating aggregations relate to the mating activity which follows. Butler and Fairey (1964) suggest that drones get within sight of queens largely by utilizing odor stimuli from downwind and then employ visual cues during the final act of mating. The work of Gary (1963), together with my own observations of drone behavior in drone aggregations, however, reveals that a “comet” of drones forms about any moving object, even about other drones. If the proper chemicals are present on the moving object, more drones stay with the object longer and attempt to mount it. Gary also found that tactile stimuli may be involved, since mating did not proceed to completion unless the queens sting chamber was open.
Many investigators have attempted to find the chemical or chemicals responsible for attracting drones to queens during the mating flight. Fractions obtained from the head, thorax, or abdomen all attract drones when raised high into the air on filter paper, but the head fraction is by far the most attractive (Gary, 1962; Pain and Ruttner, 1963). At least one secretion of the mandibular glands, 9-ketodec-2-trans-enoic acid, is strongly attractive to drones, but it is not necessarily the sole factor responsible for initiating copulation. Successful mating by queens which have had their mandibular glands removed rules out this source of queen substance as necessary for successful mating (Morse et al., 1962). Renner and Baumann (1964) also report subepidermal cells located in the dorsal region of the queens abdomen, which appear glandular in nature. The odor from these glands is apparently very strong in young virgin queens and could function during attraction of drones to queens. Conceivably, one chemical could serve in attracting drones from a distance, visual cues could orient the drone into position behind a fast-moving queen, and another odor (or odors) could result in a final copulation attempt. Alternatively, some other combination of these events might be involved.
Evidence (Gary, 1963) indicates that drones normally approach a queen from behind and below and then assume a dorsal position during mating. As the drone’s genitalia move into the sting chamber, tracheal air sac pressure from within the drone leads to an eversion of his genitalia, a slight explosion, and a resultant separation of the paralyzed drone from the queen. The queen can repeat this procedure within seconds, mate with up to ten different drones on a mating flight (reviewed by Ruttner, 1964), and return to the hive with a lifetime supply of sperm.
Certain questions in view of the evidence above now become well defined in connection with the attraction and mating of bees. We must still determine the parameters that define a drone congregation area. When the necessary signals are understood, we should be able to create such aggregations artificially in certain locations. We must also determine which pheromone or combination of pheromones and visual signals is responsible for copulatory activity. Likewise, we need to determine the stimulus responsible for the opening of a queen’s sting chamber at the actual time of mating. Other, less well-defined questions need further study. Among these are a determination of which signals stimulate a queen to leave the hive on her mating flight and whether drone congregation areas are necessary for queen mating.
Queen Influence on Workers
A queen both directly and indirectly influences workers in the colony. The direct influences include all those signals produced by a queen which are immediately perceived by workers and which regulate the attitude of workers toward each other and toward the queen. The indirect influence is much more difficult to study and includes such factors as the number of eggs laid relative to available space, ratio of drone to worker eggs laid, and ratio of eggs laid to the available supply of adult workers. One would suspect that these indirect influences, which will be treated only incidentally in this account, are integral parts of cybernetic systems.
A colony must have a queen or be in the process of producing a queen if it is to survive long. As a laying queen moves over the combs, she is constantly surrounded by a “retinue” of workers who touch and lick her body. Obviously she maintains a certain attractiveness to workers in the colony, and workers show signs of queenlessness within a few hours after losing her. A few days after such a loss, queen cell building will begin, and some of the workers will exhibit enlarged ovaries.
If a queen is restrained in a wire-screen cage in her colony, workers continue many of their activities as if they still had a queen, and apparently they show no enlargement of ovaries (Verheijen-Voogd, 1959). On the basis of this and other reports in the literature, one might classify signals produced as a result of queen presence and which influence workers as (1) inhibition of queen cell building, (2) inhibition of ovary development, and (3) attraction of workers. These can function by worker reception of an odor emanating from some part or parts of a queen, by reception of a chemical compound or compounds as a result of workers licking some exudate from the queen (which would be a process which, only with great difficulty, could be separated from a reception of these same compounds via airborne odors), or by reception of some tactile stimulus or stimuli. One must also recognize the possibility that two or more of the above can operate synergistically—that one substance can function in two or more ways (depending on context) or that one stimulus might replace another upon failure of the primary source.
The signal(s) responsible for inhibiting queen cell construction was one of the first of these communication processes investigated. Early work indicated that 9-ketodec-2-trans-enoic acid (9-oxodecenoic acid, pheromone I, queen substance) from the mandibular glands of queens was responsible for this inhibition (see Butler et al., 1959; Barbier et al., 1960; and Pain, 1961b). Later work (Butler et al., 1961) revealed that neither this chemical nor “queen scent” inhibited queen rearing as effectively as did full access to a live laying queen.
Two additional facts indicate that more might be involved than a simple chemical inhibition. In one experiment Morse and Gary (1963) found that little construction of queen cells occurred during a period of fifty-two days in colonies which each had a single queen whose mandibular glands had been extirpated (examination of queens after sixty-five days revealed no regeneration of these glands).
A result which complements this first fact is that obtained by placing a queen excluder between two halves of a hive when both contain young brood. The queen excluder, a wire mesh (or a perforated zinc sheet) which permits passage of worker bees but not of the queen or drones, restricts the queen to one half of the hive. In such circumstances the half without a queen often constructs supersedure-type queen cells around several of the very young larvae. Some queen breeders in northern California raise dozens of queens in the upper half of a hive, while the bottom half contains a laying queen restrained by a queen excluder. After these new queens develop to maturity, one will kill the rest, leave the hive on a mating flight, return to the half from which it had come, and proceed to lay eggs (provided her half of the colony has its own entrance). Under such conditions a colony can function with two laying queens. A consideration of these facts, then, indicates that the queen must have physical access to all brood cells for effective inhibition of queen cell construction.
The chemical 9-ketodec-2-trans-enoic acid administered to a queenless colony also inhibits oogenesis, but it apparently does not function nearly as well as does worker access to a laying queen (see Butler and Fairey, 1963). Pain (1961b) reports that an additional factor of tactile stimulation is necessary for full inhibition. It would be interesting to know if worker ovaries develop in colonies which have queens whose mandibular glands have been removed (both in colonies which have brood and in those which do not have brood).
Pain (1961a) reported the presence of a second pheromone (II) in queen mandibular gland extract which appeared to be highly attractive to worker bees. Chauvin et al. (1961) state that this must be present in conjunction with pheromone I for maximum attracting powers. While this chemical has yet to be isolated, identified, and synthesized, the notion that a signal is necessary received support from the results of experiments by Gary (1961b) (see also Zmarlicki and Morse, 1964). Queens without mandibular glands attracted ever-decreasing numbers of workers until, after eleven months, at the experiment’s termination, such queens were virtually nonattractive. Butler and Simpson (1965) also describe the attractive qualities of chemicals produced by paired glands located in the sting chamber of queens (Koschewikow glands).
These different sets of results and interpretations from seemingly similar experiments could arise from a number of discrepancies. As implied earlier, a combination of factors may be more important than any single factor, or one factor may substitute for another. Also, in dealing with chemical signals, one is always plagued with the problems of purity of extracts and of synthetic products (possibly due to subsequent contamination by absorption or adsorption) as well as with the problem of correct dosage. Too high a concentration of an attractive substance, for example, can prove to be repellent rather than attractive, which may explain an antagonistic reaction of workers to mandibular gland extract (Gary, 1961a). Conceivably, colony condition and past history may influence some of these results. Jordan (1961), for instance, found a difference in attraction or repulsion to a queen substance depending upon the absence or presence of a queen in a colony.
In summary, certain signals produced by a queen have definite attractive and inhibitory effects in the normal life of a colony. If one kills a queen, workers immediately begin constructing queen cells, and an increased number of workers develop enlarged ovaries. Of the various substances isolated and tested so far, 9-ketodec-2-trans-enoic acid is the most active in directly attracting workers, inhibiting queen cell construction, and inhibiting ovary development among workers. The fact that, on the one hand, this chemical alone does not function nearly as well as the presence of a live, laying queen and that, on the other hand, the surgical removal of these glands (supposedly the primary, if not sole source of this pheromone) from the queen of a colony is not biologically equivalent to killing the queen suggests that more decisive experiments are in order.
Females in the colony regularly belong to either one of two castes, queen or worker. A single laying queen in a colony produces enough inhibitory substance(s) to maintain a clear distinction between these two castes. Removal of the queen, while initially resulting in a partial breakdown of the caste distinction (many workers undergo ovary development), also results in provision for queen replacement (in the building of queen cells). Enriched nourishment of very young larvae in these queen cells with an abundance of royal jelly leads to the virtually simultaneous development of several queens.
Thus, although the major distinction between the development of a worker and a queen appears to be initially determined by nourishment and later maintained by inhibitory substances, many gaps in our knowledge of this process remain to be filled (as discussed by Weaver, 1966). Somehow, bees concentrate on the development of a dozen or more queen cells although hundreds of suitable larvae are available. The implication is that bees must receive enough information about current hive conditions to provide extra food to certain larvae but not to others. Similarly, workers must utilize information about the age of the very young larvae chosen for extra nourishment, since overnourishment of older larvae would result in intercastes. We need to determine and to better understand the mode of action of these various substances and to know why a queen is immune to these chemical substances which she herself produces.
The subject of male (drone) determination presents fewer problems, owing to the haploid nature of these individuals. Once a queen has laid an unfertilized egg, it will probably become a drone upon maturation. Anyone interested in communication within the colony might well concentrate on the following questions: (1) Why are drones produced and normally tolerated in colonies during only a part of the productive season? (2) Why does the queen generally lay drone eggs in oversized cells and not in worker cells? (3) How does the queen (or how do the workers) exercise control over the sphincter valve leading from the spermathaeca? (4) What signals result in the production of many drones prior to queen replacement? Concentrating on the signals responsible for these activities will be the most fruitful approach for further studies.
Preparation for Swarming
The preceding two sections provide some basis for understanding the factors leading to supersedure (replacement of lost or failing queens). During this process, a sudden or gradual loss of a signal(s) produced by the queen (conceivably aided by a signal(s) indirectly produced as a consequence of her absence) automatically results in her rapid replacement. Throughout this period of supersedure, which may occur at any time of the season, colony life proceeds without apparent disturbance. Although the first queen which emerges from her cell usually kills all rivals (even those still in their cells), this may not happen, and one can find both old and new queens laying in the same colony.
Within a few days after her emergence from a cell, the colony’s single queen takes her mating flight, returns to her own colony, and soon begins to lay eggs. In most of the ant and termite species hundreds or thousands of the offspring of a queen do the same and subsequently each begins a new colony. The fact that queen honey bees kill all rivals, mate, and return to the same colony makes it imperative that some other means of colony reproduction must occur if the species is to survive. For honey bees this occurs by means of the phenomenon of swarming, in which one-half to three-fourths of the bees in a colony leave with the old queen (Martin, 1963) and begin a new colony.
Preparations for swarming begin most frequently during colony expansion in the spring. Increased use of winter stores, availability of new external sources of pollen and nectar, and lengthening days with ever-warmer temperatures all result in a period of rapid increase in brood production. This increased egg-laying and overpopulation [restriction of space for adults (Simpson and Riedel, 1963)] of the colony generally precedes preparations for swarm production by a hive.
One can see the advantage to the species if swarming takes place in the early part of the season. This permits the new colony to find a suitable site, permits a rapid increase for both halves while nectar and pollen are still in plentiful supply, and gives the new colony several months time before winter arrives. This also allows enough time for building new combs, closing extra openings in the new nest cavity, and storing nectar and pollen.
Swarming preparations proceed with more apparent disturbance than occurs during supersedure. Some time before the swarm leaves, workers feed the queen less (Allen, 1960), force the queen to slow or stop egg-laying (Taranov, 1947; Allen, 1956), and begin constructing queen cells. This is necessary if the queen is to be light enough in weight to fly with the swarm, and it is partly responsible for a very different behavior on the part of this old queen. Instead of a ponderous movement over the comb, she now runs about as rapidly as the most active workers. Workers also increase the frequency of “shaking” the queen, a vertical movement of the abdomen by a worker in contact with the queen (Allen, 1959). The function, if any, of this action remains obscure.
With the maturation of queens in their cells, the old queen becomes even more active. Occasionally the mature queen leaves with a swarm before completion of this maturation of new queens, but more often she leaves only after queens in their cells are fully mature (dependent, apparently, on weather conditions—swarms usually leave the colony on the morning of a clear still day). As in supersedure, the free queen attempts to tear open queen cells and kill the queens inside, if these queens are near maturity. But whereas in supersedure workers permit and even aid the queen in this activity, in swarm preparation workers prevent the queen from engaging in it. Nearby workers cover these cells closely and prevent the queen from damaging them. Some even appear to pull her away if she gets too close.
In time, if adverse weather (which prevents a swarm from leaving) results in full maturation of new queens in their cells and before the old queen leaves, these new queens attempt to emerge. But, under such conditions, as these queens begin to cut the lids off their cells from the inside, workers can be seen to press these lids back into place and glue them shut again. Any nutritional requirements are met by feeding occupants through a small hole left in one edge.
An old queen, when restrained from killing maturing queens which are so restricted in their cells, eventually exhibits further signs of distress. She now frequently stops on her trips from royal cell to royal cell, lowers herself to the comb, and produces a distinctive signal (queen piping) which can be heard several meters from the hive.
Queens in their cells respond to the piping of free queens by piping themselves, and an active exchange of this sort can continue for up to a week. The sounds of these two types of queens are distinctly different, a fact which has led to the assigning of different names to them. A queen free on the comb produces a series of hollow sounding “toots” (Tüten), while those still in their cells produce a rapid series of harsher sounding “quacks” (Quaken). Although we have direct evidence that queens perceive and react to these sounds by emitting sounds themselves (Hansson, 1945; Wenner, 1962b), we have little evidence concerning the function of these sounds. One possibility is that they inform the hive of the number and disposition of various queens in the colony. The fact that a queen, once she has emerged, produces the sound of a free queen and not the sound she had previously produced while still in her cell lends support to this notion.
Occasionally hives repeatedly produce swarms (afterswarms), even to a point where neither the later afterswarms nor the parent hive has much opportunity for surviving the season. This could be a consequence of a misuse of information by workers in the colony. Once the old queen has left with a swarm, the first new queen to emerge (if inadvertently prohibited from killing other new queens in their cells) could function briefly as the “old” queen before leaving with a second swarm. In this manner repeated swarming could result in decimation of the original colony. The data of Gary and Morse (1962) indicated that conditions leading to afterswarming could occur repeatedly without resulting in afterswarming. These factors led to an extended period in which a colony apparently had no laying queen.
Clearly, preparation for swarming involves a complex of communication processes before a group of bees comprising the swarm forces the queen from the hive. Preparations for swarming and supersedure apparently share many of the same common elements, and preparations for one could conceivably lead to the accomplishment of the other. This could be responsible for the inadvertent substitution of supersedure by swarming if this process were begun early in the season.
Although overcrowding plays a major role in instigating swarming preparations, such an action occurs more frequently in the early part of the season than later in the year. But, in fact, even an uncrowded observation hive may swarm if such a hive begins supersedure préparations during that part of the season in which swarming is most likely to occur. Also, colonies which have passed the swarming stage in California valleys begin swarm preparations anew when moved to high altitudes. When we begin to understand which signals lead to swarming [and why a given bee leaves with the swarm while its neighbor bee remains in the hive (Simpson and Riedel, 1964)], then we should be able to produce swarming by using the signals themselves, even in uncrowded hives. It also will be possible to halt swarming by removal or neutralization of those signals responsible for one or more of the processes.
Division of Labor
One of the most puzzling, and perhaps most difficult, questions in the study of honey bee communication concerns the method(s) by which a colony regulates the activities of its members. Individual bees in a colony succeed in achieving what is necessary for normal functioning and progression of a colony, if conditions permit.
An early deduction, that engagement by bees in particular types of activities depends on their ages [a notion developed by Gerstung and elaborated upon by Rösch (Ribbands, 1953)], appears to have little basis other than the presence of a general trend resulting from sequential glandular development (Lindauer, 1953). That there is a gradual turnover of individuals engaged in a given project (e.g., Lindauer, 1953; Allen, 1960) and that any one individual can be seen to change from one occupation to another without apparent cues indicate there is efficient control of the division of labor [this point is summarized more completely elsewhere (Wenner, 1961)].
Since a colony rapidly reaches equilibrium after a disturbance, Wenner (1961) suggested that the division of labor might be an automatic (cybernetic) system. Testing this notion, however, proved to be more difficult than originally anticipated, largely because of the difficulty of defining the activities of a given bee at any one time. That is, although some activities are clear-cut (i.e., fanning, foraging, guarding), the majority of colony activities are yet to be clearly defined. It would appear that a more thorough understanding of division of labor awaits an objective analysis (rather than interpretation) of activities occurring in a colony.
Each colony apparently has a distinctive odor. Within a week after one splits a large hive, each of the two halves has its own odor. Subsequently, bees from one colony rarely enter another, even if the two are adjacent. This fact indicates an environmental rather than a genetic origin of an odor. Materials carried into the hive differ slightly in their chemical make-up, and subsequent metabolic processes would reinforce this difference in composite colony odor (Kalmus and Ribbands, 1952).
The ability of a homecoming bee to find its own hive among dozens or even hundreds of others in an apiary depends in part upon this characteristic odor, but this orientation depends more upon recognition of physical features in the immediate area, once hive location is learned. If, however, a hive is moved a meter or slightly more, or if prominent topographical features are changed in the immediate environment, colony odor perhaps becomes the single most important signal in orientation of homecoming bees. Should physical appearance or hive location change, returning foragers typically drop downwind until they perceive their hive odor and then commence a zigzag flight upwind until again reaching their colony (Ribbands and Speirs, 1953).
Although some popular accounts place much emphasis on the need for an incoming bee to carry the appropriate odor in order that it might not be attacked upon entering a hive, some flexibility exists in this communication system. Guard bees which challenge entering bees initially respond to visual signals from incoming bees—that is, if a bee enters a strange hive while engaged in a normal flight pattern, it is unlikely to be attacked. This is also true for a queen returning from her mating flight. Furthermore, should a bee from another hive be challenged, its posture, and contents of its honey stomach, dictate ultimate action on the part of guard bees (Free, 1954; Ribbands, 1954b; Meyerhoff, 1955). Thus a composite of signals determines admission of an individual into a colony. Beekeepers make use of this lack of rigidity in uniting colonies when circumstances warrant making one hive out of two. Often this uniting can be accomplished merely by placing the two colonies together, provided environmental conditions are favorable (e.g., the proper time of day, a high temperature, a good nectar flow, and a lack of robbing activity in the area).
Hive odor often functions in conjunction with an external secretion from the Nasanov gland in attracting workers into the hive once they have come close to the entrance. This phenomenon might be best illustrated by an example. Bees which are shaken from a comb onto the ground in front of their hive initially appear confused. Soon, however, some of these bees find their way to the hive. When this occurs, these first bees stand in place or walk slowly toward the hive entrance and expose their Nasanov glands (located at the posterior end of the abdomen between two dorsal sclerites) while fanning a current of air away from the hive. Presumably hive odor and Nasanov gland secretion mingle in this air current and aid other bees to find their way back to the hive. Soon a large number of bees forms a broad path, while all move slowly toward the hive as they fan and expose their scent glands. Within a short time, all bees are again in the hive from which they came.
Colony odor and Nasanov gland secretion also function in exploitation of food sources—a fact which possibly has not received enough attention in the past. In 1926 von Frisch and Rösch succeeded in training bees from adjacent colonies to forage at adjacent dishes at a point remote from the hive. Experienced foragers from one hive rarely attempted to forage at the second dish even when it was moved within two meters of the first dish. Likewise, recruited bees rarely visited the dish frequented by foragers from the second colony, even though both sources contained unscented food. From these results they concluded that regular visitors to a site leave an odor at the site which is characteristic of the colony from which they came and which aids in orienting hive mates to that food source.
Ribbands, Kalmus, and Nixon (1952) later confirmed and extended these results. After conducting an experiment similar to that described above in which bees from each hive were visiting regularly the dish specific to that hive, they switched the adjacent dishes. While experienced bees initially appeared confused after the exchange of dishes, most of them eventually settled at the same physical location they had visited earlier, indicating a strong place memory. Within the first fifteen minutes after the exchange, recruited bees from one hive went to the alternate location where concentration of their own hive odor was apparently strongest. After fifteen minutes, however, recruited bees showed little preference, perhaps because odors from each hive were now on both dishes. These results indicate that recruited bees depend strongly on colony odor left at the food place. I have also noticed that newly recruited bees often expose their nasanov gland before landing.
Ribbands (1955) further reported that one bee landing and feeding only once from a vial made this vial more attractive to itself and to others from its colony. Experiments by Renner (1955) indicated that this influence arises from the general hive odor and not from some specific substance or substances from the Nasanov gland. Lecomte (1957) extracted a substance from glass beads which were placed in the sugar solution from which bees obtained their loads. He found that this substance attracted bees from the parent hive but repelled bees from another hive.
All these results suggest that colony odor, while aiding in the orientation of homecoming bees and providing an identification for bees entering the hive, also serves as a means by which a colony extends its influence far beyond the physical confines of the hive. Results of experiments on foraging patterns of apiaries when faced with competition from other apiaries (Levin and Glowska-Konopacka, 1963) further suggest a possible “staking out of claims” by colonies.
COMMUNICATION ABOUT ENVIRONMENTAL FACTORS
Alarm and Distress Signals
A skilled beekeeper can detect severe disturbance in a colony either by odor or by sound (aside from the more obvious attacking by bees). These apiarists learn to proceed carefully upon the advent of either signal and may even discontinue work with a colony once one of these signals becomes intense.
The odor a human can perceive which emanates from a disturbed hive is particularly characteristic. (A severely disturbed bee can be seen to raise the tip of its abdomen into the air and exude a droplet at the tip of its everted sting.) The presence of this sweet-smelling, pungent odor in the air results in attacks of likely objects by other bees; if the substance exuded from the sting is smeared on an object and brought near a group of guard bees, such an object is likely to be attacked and stung, whereas an unscented object may be ignored. Vision plays a role in this attack, because bees attack a moving body containing sting exudate more often than they attack a stationary object containing the same amount of substance (Ghent and Gary, 1962). The exudate can be collected in quantity by inducing bees to sting through cloth mesh. The danger of this practice, due to the release of large amounts of an alarm odor into the atmosphere, became clear during a large-scale operation of this sort—bees from the apiary attacked any moving object in the area (Morse and Benton, 1964).
It is not clear whether there are one, two, or more sources of alarm odor. Maschwitz (1964) describes the mandibular glands as a source of alarm substance in honey bees. Bary (1961b) describes the antagonism of workers toward objects coated with mandibular gland extract obtained from queens and suggests a dosage effect. This indicates that a small amount would attract workers while a large amount would repel them.
Sounds produced by disturbed bees are not as clearly defined as are odors. Although a skilled beekeeper can detect a major disturbance by listening for particular sounds in a hive, we have no conclusive evidence that such sounds are normally involved in communication by bees. The most common sound produced by disturbed bees is a series of short bursts (¼ to ½ sec) of about 250 Hz frequency. If large numbers of bees produce these at about the same time, this results in the collective sound one gets by jarring a hive. Soon after such a general burst, other workers in the hive produce short (½-sec) notes which are less variable in tone. These latter types of sounds (one variation of worker piping) continue until the hive is once again calm. These correlations suggest an alerting function for the former and a quieting function for the latter, but we do not yet have direct evidence for either. Displays and descriptions of some of these sounds are given by Wenner (1964).
Exploitation of Food Sources
Perhaps more investigators have concerned themselves with a study of the honey bee “waggle dance” than with any other communication act within an animal species (human language excepted). Honey bees have always been known to have an efficient system of exploitation of food sources, but only since 1940 has there been an attempt at approaching the problem experimentally. By 1950 von Frisch had formulated a rather complete statement of the now familiar theory of “dance” recruitment among honey bees.
The explanation proposed for this recruitment to a newly discovered source of food is remarkably simple. After a successful forager has made several round trips to a rich source of food, it may execute a figure8pattern on the vertical comb in the hive. During this maneuver, other bees in contact with this first bee become excited and a majority of them leave the hive. Some of these recruited bees succeed in finding the site visited by the first successful forager(s).
Initially von Frisch and Rösch (1926) deduced that the figure-8 pattern communicated information about a pollen source and that the related “round dance” conveyed nectar source information. Later studies, however, indicated that round and figure8dances are part of a continuum (Hein, 1950) related to the location of either nectar or pollen sources (von Frisch, 1950).
The details of the theory of honey bee exploitation of food sources as early stated by von Frisch are presented clearly in many sources, but Lindauer (1961, p. 32) summarizes this early statement quite concisely:
If a worker bee has found a good source of food, she announces her discovery at home by means of either a “round dance” or a “tailwagging dance.” If the food is to be found less than 80 meters away, she performs a round dance, running around rapidly in a circle, first to the left, then to the right, and the surrounding bees become excited by this dance. They follow interestedly behind the dancer and thus receive the message: “Fly out from the hive; right in the neighborhood is food to be fetched.” The smell of the blossoms still remaining on the dancers gives the further information how the food source smells, and in this way the informed seekers can hunt out the fragrant blossoms.
If, however, the food source is a good distance away, then the tail-wagging dance gives additional information about the exact location of the newly discovered goal. The rhythm of this dance shows the distance: the farther away the goal, the fewer cycles of the dance in a given time. When the goal is close, there are more turns per minute. The direction of the goal is also conveyed by the tail-wagging part of the dance, the sun being used as a reference point. On the vertical honeycomb in the dark beehive the angle between the direction to the sun and that to the food source is transposed into an angle with respect to gravity, according to the following rule: a tail-wagging run pointed upward means that the source of food lies in the direction of the sun; the same part of the dance directed downward announces that the food is opposite to the sun. If the tail-wagging run points 60° left of straight up, the food source is 60° to the left of the sun, and so on . . .
These soliciting dances last only as long as the food is in abundance, and only first class, highly concentrated nectar solutions are advertised. The quality of each source of food can be seen from the duration and vivacity of the dance. Since the number of newly won followers is dependent on the same factors, these dances harmonize supply and demand in the search for food in an almost perfect manner.
Although some of the information contained in dances enables an investigator to determine the approximate distance and direction of a food source, there has always been some question as to which elements of the dance maneuver, of those we have chosen to measure (if any), are used by bees in finding the food source (Wenner, 1962a). One can obtain a reasonably precise estimate of distance and direction to a food source by gathering data from many dancing bees, but an individual bee attending a single dancing bee is exposed to a rather crude estimate of distance and direction. Both within-bees and between-bees variance [which are approximately equal (Wenner, 1962a)] arise, in part, from influences of temperature and wind on the foragers while they are in flight, before they dance in the hive on their return, and also from the inherent differences among bees. As a consequence, the information available to a potential recruit is not very accurate (Schweiger, 1958; Wenner, 1962a; Levchenko, 1961, 1963)—for a food source 200 meters from the hive, standard deviation in the dance is equivalent to about 30 meters.
The possibility that other elements which could contain these types of information might also be produced during the dance maneuver has provided the basis for further studies. The correlation between the distance a bee has traveled while on its way to a food source and the number of times it executes a figure-8 maneuver per unit time when it returns to the hive (dance rhythm) proved to be only the first of several possibilities. In a comprehensive study of various elements, for example, Esch (1956) and von Frisch and Jander (1957) reported that the time spent while waggling or the number of times the abdomen moved laterally during the straight-run part of the figure-8 pattern each proved to be better indicators of distance traveled than did the dance rhythm.
A short time after this, both Wenner (1959) and Esch (1961) found an amplitude-modulated sound signal produced during the waggle dance. The sound, which is generally produced only during the straight-run portion of the figure 8, has an average frequency of about 280 Hz. [279.4 ± 1.27 Hz, N = 439 (Wenner et al., 1967)], a figure in remarkable agreement with the vibration sensitivity of antennae reported by Heran (1959). (In his study, the antennae were most receptive to vibrations of about 274 Hz in the lateral direction and 285 Hz in the dorsoventral direction.)
Later work showed that two elements of this sound signal, the time of sound production during a straight-run portion of the figure 8 and the number of amplitude-modulated pulses produced within this straight run, correlate highly with the distance traveled from the hive before a bee reaches the food site (Wenner, 1962a). The high correlation between these two elements makes it impossible to separate one from the other clearly. Esch (1964) confirmed the relationship between sound production time and distance traveled from the hive.
Esch (1963) reported another correlation, one between rate of modulation and concentration of the sugar solution obtained at a food site. A more extensive study (Wenner et al., 1967), however, revealed that one can obtain many correlations in studies of this type but that interpretation of such correlations is exceedingly difficult. In this latter study, a large number of correlations was obtained between pairs of variables—none of which could be construed as acts of communication. The correlation and slight regression obtained between sugar concentration and modulation rate could be dismissed as a consequence of a temperature influence on the signal rather than as a means by which bees could communicate information about sugar concentration (bees flying through cool air have their metabolism increased more upon entering a warm brood nest than bees which have flown through warm air).
One point has not been appreciated fully in connection with the theory of recruitment by means of the waggle dance. To date we have no direct evidence that the elements we have chosen to measure among the various elements in the dance maneuver are utilized by recruited bees in their search for a food source. The entire theory rests upon a remarkable set of correlations between what occurs during the dance in the colony and those environmental parameters extant during visitation of a food source by bees (together with correlations between occurrence of elements in the dance maneuver and success of bees at exploiting food sources in the field). The evidence can be summarized as follows:
1. Successful foragers, after returning to the hive, produce a dance which contains information about location, type, and (possibly) quality of food.
2. The number of foraging bees visiting a newly replenished source of food increases exponentially until a plateau is reached.
3. Bees recruited tend to go only to the food location indicated in the dance (step and fan experiments of von Frisch, 1954).
The conclusion reached from this line of evidence is that recruits can find the source independently by interpreting and using abstract information contained in the dance.
By 1955 at the earliest and 1960 at the latest, however, enough evidence had accumulated in the literature to permit the formulation of an alternative explanation of a rapid exploitation of food sources. Two findings, in particular, contribute to a growing body of evidence for this alternative theory. The first of these is the fact that regular visitors to a food site leave a characteristic odor at the site (von Frisch and Rösch, 1926) which is apparently attractive to bees from the same hive (Ribbands, 1955) but not to bees from another hive (Lecomte, 1957). The second is the ease with which bees can be conditioned to come to a site upon presentation of the conditioned stimulus [simple discrimination conditioning (see Wenner and Johnson, 1966)]. Although various Russian workers have discussed the importance of discrimination and other types of conditioning in exploitation of food sources (see Chesnokova, 1959; Lobashev et al., 1961; and Lopatina and Chesnokova, 1962), neither the attraction which a characteristic odor holds for bees from a common hive nor discrimination conditioning has received general recognition in discussions of food exploitation. (See the theory statement at the beginning of this section.) Ribbands (1953), Heran (1958), and von Frisch (1965), however, include a discussion of the value of hive scent left at the food site.
On the basis of the information available before 1960 (and impiemented by evidence published later), it is now possible to state an alternative theory which can account for the rapid exploitation of available food sources without invoking a need for bees to communicate (utilize) abstract information. Successful recruitment of bees to a newly yielding source of food depends upon whether a bee is experienced (conditioned bee) or inexperienced (recruit bee) at visiting a given species of plant. Experienced bees rely most heavily upon certain parameters, arranged below according to their relative importance:
1. Experienced bees routinely monitor known sources of food once these sources become empty. If food again becomes available:
a. Returning laden bees enter the hive, carrying the characteristic odor of the food source on their bodies.
b. Bees experienced at visiting such sources leave the hive in response to the odor stimulus (simple discrimination conditioning) provided by the first successful forager(s) and travel to the site at which they had earlier been successful (choice discrimination conditioning). This can explain a rapid, exponential increase in the number of foragers at a newly refilled dish [see Wenner and Johnson (1966) and Johnson and Wenner (1966) for discussion of the use of these phrases]. This site need not be the same as the site visited by the first forager (s) (Johnson, 1967a).
c. Upon arrival at the suspected site, these experienced bees can rely upon their previous learning of the area’s visual configuration, characteristic odors of the location, the odor of other feeding bees, the scent of the hive left at the site (von Frisch and Rösch, 1926; Ribbands, 1955), and odor of the food (Ribbands, 1954a). The great value of hive odor left at the site becomes apparent if one washes and refills the dish—experienced visitors may require several minutes before feeding from the dish again. This occurs especially if one uses unscented sugar solution both before and after refilling (sucrose solution has no scent).
2. Recruited bees (it should be emphasized that most of the foraging force is experienced at any given time) certainly appear to be stimulated to leave the hive as a result of the dance motions (Steche, 1957), and dance sounds appear to be a necessary part of this alerting (Esch, 1963). After leaving the hive, recruited bees apparently drop downwind from the hive and pick up the odors to which they have been recruited. These odors can come from one or a combination of several sources:
a. Bee scent at or near the site appears to be the strongest single attractant (von Frisch and Rösch, 1926; Ribbands et al., 1952). Bees landing and feeding result in the increased attractiveness of that site to hive mates (Ribbands, 1955). (Such an odor also changes the design of many experiments the moment the first bee lands.)
b. Exposure of the scent gland by a feeding bee reportedly attracts other bees to a source of food, but not nearly as well as the hive odor left at the site (Lecomte, 1957; see also Renner, 1955). (Newly recruited bees often expose their scent gland before landing.
c. Odor of the food, possibly carried in the honey stomach (von Frisch, 1950), but more likely on the body (Shaposhnikova, 1959; Nixon and Ribbands, 1952), apparently aids a recruited bee in its search.
d. Conceivably, the unique scent of any location can also function in orientation (see Casey, 1950).
The notion that inexperienced bees find a prospective site by using odor cues in the environment and not distance and direction information obtained from the dance maneuvers of successful bees is supported by experimental results. Johnson (1967b) and Wenner (1967) have repeated the fan and step experiments of von Frisch (1954). If bees from an experimental hive routinely forage at one site and not at other nearby (control) sites, then inexperienced bees are normally recruited to or near the same site visited by hive mates. If one provides bee visitation (from a second, control, hive) at control sites as well as at the experimental site, however, inexperienced recruits from the experimental hive no longer preferentially arrive at or near the experimental site. The results of these repeat experiments suggest that the classic experiments of von Frisch lack the essential controls alluded to in the previous discussion. [See von Frisch, (1967) and Wenner and Johnson, (1967) for more details on this point.]
Another argument against the familiar theory is the low number of recruit bees arriving at a site regularly visited by experienced foragers [fewer than 5 per cent of the bees observed attending a dancer and leaving the hive arrived at the site (Johnson and Wenner, 1966)]. Studies of dance communication, in fact, would have been virtually impossible if efficiency of recruitment were any higher—because of the potentially large number of unmarked bees arriving at the site.
After a swarm leaves the hive, a cluster or two may develop on a nearby branch or projecting object. Soon all bees form one tight cluster. Within a week after such clustering, weather permitting, the cluster again becomes a swirling mass of individuals and moves off toward the site which is to be the permanent hive for the new colony. Sometimes swarms go directly to the new site (Root, 1950, p. 606). In that case, orientation to the new location must occur before the swarm leaves the hive.
The mechanisms responsible for successful swarm relocation appear to be closely related to those utilized in exploitation of external food sources (Lindauer, 1961). Any theory which purports to explain orientation in the surrounding environment for one process should be able to offer an explanation for the mechanism(s) involved in the other. The following outline can replace earlier theories of swarm relocation without invoking any need for swarm bees to utilize abstract information contained in the dance maneuver (see Wenner et al., 1967 for an expanded presentation).
1. The new colony develops a distinctive odor while clustered on a nearby object.
2. Scout bees which have found a site spend time at this location, and colony odor accumulates at the site.
3. Bees alerted to leave the swarm cluster as a result of dance motions search the surrounding countryside until they find their own colony odor. Repeated trips aid experienced bees in learning the route well.
4. The most suitable site receives the largest number of visitors for the greatest number of round trips, accumulates the greatest amount of colony odor at that site, and eventually becomes the site for the new hive (positive feedback).
5. As the colony disbands its cluster and moves through the air, the bees which have been conditioned to the new location “lead” the swarm by means of Nasanov gland secretion or some suitable combination of odors. The mass of bees thereby does not need to use abstract information concerning distance or direction traveled.
6. Queen odor must accompany the swarm as it moves through the air, or the colony will return to or near its original site (Morse, 1963).
7. As the destination is reached, the mass of bees can stop when the conditioned bees alight at the new site. The swarm soon clusters at the entrance of the new location, aided by Nasanov gland secretion and presence of colony odor.
Although this alternative explanation lacks support of critical evidence (as do earlier theories), it successfully explains how a swarm, having traveled the required distance and direction, can come to a stop. (An individual bee traveling in a swarm constantly circles and consequently travels in every direction and for a much greater distance than that indicated in the dance by “scout” bees.) However, we must still determine the signal(s) responsible for the swarm leaving as a unit. Likewise, we must assess the relative roles of queen and scout bees and the interaction between these two as possible sources of odor (or whatever other signals may be responsible for the swarm’s exit).
A comprehensive evaluation of all lines of evidence regarding honey bee communication indicates that we are beginning and not ending our studies. If one demands direct evidence (as defined above) rather than indirect evidence in a construction of explanatory schemes, few positive statements can be made about bee communication. Although honey bees and honey bee colonies are remarkable, results from recent studies provide evidence that they may not be as superior to other animals at communicating abstract information as previously supposed.
Only an extremely small percentage of papers on bee communication deal with species other than our common honey bees (Apis mellifera Linnaeus). Results from experiments dealing with conditioning of bees suggest that closely related bee species may be more similar to honey bees in their communication systems than earlier believed. This tentative conclusion precludes any discussion of these other species at this time.
So much has now been published by so many investigators, sometimes with and often without good supporting evidence, that an analytical summary becomes difficult. A variety of opinions has been expressed on each subject at one time or another—often, even conflicting opinions have been stated by the same author at different times (without reference to the apparent conflict). Consequently, by carefully selecting sources, one can build a rational argument for almost any position. In writing this chapter, a deliberate attempt has been made to avoid this pitfall. Further experimentation will test the success of this attempt.
Many earlier studies show bias in having been designed to gather evidence in support of previously established concepts [see Popper (1957) for a discussion of the subtle weaknesses inherent in such an approach]. Perhaps it is now past the time that we ask, “How can we prove our hypothesis?” and turn to the more rigorous question, “What experiment will contradict our hypothesis?” When the second question has been asked, studies of bee communication should move ahead more deliberately.
Alber, M., R. Jordan, F. Ruttner, and H. Ruttner, 1955. Von der Paarung der Honigbiene. Z. Bienenforsch. 3:1-28.
Allen, M. D., 1956. The behaviour of honeybees preparing to swarm. Brit. J. Animal Behaviour 4:14-22.
——, 1959. The occurrence and possible significance of the “shaking” of honeybee queens by the workers. Animal Behaviour. 7:66-69.
——, 1960. The honeybee queen and her attendants. Animal Behaviour, 8:201-208.
Barbier, M., E. Lederer, and T. Nomura, 1960. Synthese de l’acide ceto-9decene-2-trans oique (substance Royale’) et de l’acide ceto-8Înonene-2trans oique. C. R. Acad. Sei. Paris. 251:1133-1135.
Butler, C. G., and E. M. Fairey, 1963. The role of the queen in preventing oogenesis in worker honeybees. J. apic. Res. 2:14-18.
——, and E. M. Fairey, 1964. Pheromones of the honeybee: biological studies of the mandibular gland secretion of the queen. J. apic. Res. 3:65-76.
——, and J. Simpson, 1965. Pheromones of the honeybee (Apis mellifera Linnaeus). An olfactory pheromone from the Koschewnikow gland of the queen. Ved. Pr. vyzk. Ust. vcelars. Dole. 33-36.
——, R. K. Callow, and N. C. Johnston, 1959. Extraction and purification of ‘queen substance’ from queen bees. Nature. 184:1871.
——, R. K. Callow, and N. C. Johnston, 1961. The isolation and synthesis of queen substance, 9-oxodec-trans-2֊enoic acid, a honeybee pheromone. Proc. roy. Soc. B 155:417-432.
Casey, A. L., 1950. Conditioned behaviour in bees. Aust. beekpr. 52:7.
Chauvin, R., R. Darchen, and J. Pain, 1961. Sur l’existence d’une hormone de construction chez les abeilles. C. R. Acad. Sei. Paris. 253:1135-1136.
Chesnokova, E. G., 1959. Conditioned reflexes established in bees through a chain of visual stimuli. Trud. Inst. Fiziol. Pavlova. 8:214-220.
Esch, H., 1956. Die Elemente der Entfernungsmitteilung im Tanz der Bienen (Apis mellifica). Experientia. 12:439-441.
——, 1961. Ueber die Schallerzeugung beim Werbetanz der Honigbiene. Z. vergl Physiol 45:1-11.
——, 1963. Ueber die Auswirkung der Futterplatzqualität auf die Schaller zeugung im Werbetanz der Honigbiene. Zool Anz. Suppl 26:302-309.
——, 1964. Beiträge zum Problem der Entfernungsweisung in den Schwanzeltanzen der Honigbiene. Z. vergl Physiol 48:534-546.
Free, J. B., 1954. The behaviour of robber honeybees. Behaviour. 7:233-240.
Frisch, K. von, 1950. Bees: Their Vision, Chemical Senses and Language. Great Seal Books (Cornell Univ. Press). 118 pp.
——, 1954. The Dancing Bees. Methuen & Co. Ltd: London.
——, 1965. Tanzsprache und Orientierung der Bienen. Berlin: Springer Verlag.
——, 1967. Honeybees: Do they use direction and distance information provided by their dancers? Science. 158:1072-1076.
——, and R. Jander, 1957. Ueber den Schwanzeltanz der Bienen. Z. vergl Physiol 40:239-263.
——, and G. R. Rösch, 1926. Neue Versuche über die Bedeutung von
Duftorgan und Pollenduft für die Verständigung im Bienenvolk. Z. vergl Physiol 4:1-21.
Gary, N. E., 1961a. Antagonistic reactions of worker honeybees to the mandibuÎ lar gland contents of the queen. Bee World. 42:14-17.
——, 1961b. Queen honey bee attractiveness as related to mandibular gland secretion. Science. 133:1479-1480.
——, 1962. Chemical mating attractants in the queen honey bee. Science 136:773-774.
——, 1963. Observations of mating behaviour in the honeybee. J. apic. Res. 2:3-13.
——, and R. A. Morse, 1962. The events following queen cell construction in honeybee colonies. J. apic. Res. 1:3-5.
Ghent, R. L., and N. E. Gary, 1962. A chemical alarm releaser in honey bee stings. Psyche. 69:1-6.
Hammann, E. von, 1957. Wer hat die Initiative bei den Ausflügen der Jungkönigin, die Königin odor die Arbeitsbienen? Insectes Sociaux. 4:91-106.
Hansson, Â., 1945. Lauterzeugung und Lautauffassungsvermögen der Bienen. Opusc. ent. 6:1-122.
Hein, G., 1950. Ueber richtungweisende Bienentänze bei Futterplätzen in Stocknähe. Experientia. 6:142-144.
Heran, H., 1958. Die Orientierung der Bienen im Flug. Ergebn. Biol. 20:199-239.
——, 1959. Wahrnehmung und Regelung der Flugeigengeschwindigkeit bei Apis mellifica L. Z. vergl Physiol 42:103-163.
Johnson, D. L., 1967a. Communication among honey bees with field experience. Animal Behaviour. 15:487-492.
——, 1967b. Honey bees: Do they use the direction information contained in their dance maneuver? Science. 155:844-847.
Johnson, D. L., and A. M. Wenner, 1966. A relationship between conditioning and communication in honey bees. Anim. Behav. 14:261-265.
Jordan, R., 1961. Versuche hinsichlich des Reagierens der Bienen auf Königinnenduft-Extrakt. Bienenvater. 82:3-5, 45-48, 78-81, 110-112, 145-146, 177-178.
Kalmus, H., and C. R. Ribbands, 1952. The origin of the odours by which honeybees distinguish their companions. Proc. roy. Soc. B. 140:50-59.
Lecomte, J., 1957. Sur le marquage olfactif des sources de nourriture par les abeilles butineuses. C. R. Acad. Sei. Paris. 245:2385-2387.
Levchenko, I. A., 1961. (Studying bee dances with the help of a cine camera.) Pchelovodstvo. 38:37-39.
——, 1963. (The indication by bees of the whereabouts of a honey source.) Pchelovodstvo. 40:30-33.
Levin, M. D., and S. Glowska-Konopacka, 1963. Responses of foraging honeybees in alfalfa to increasing competition from other colonies. J. apic. Res. 2:33-42.
Lindauer, M., 1953. Division of labour in the honeybee colony. Bee World. 34:63-73, 85-90.
——, 1961. Communication among Social Bees. Cambridge: Harvard Univ. Press. 143 pages.
——, 1967. Recent advances in bee communication and orientation. Ann. Rev Entomol. 12:439-470.
Lobashev, M. E., N. G. Lopatina, J. A. Nikitina, and E. G. Chesnokova, 1961. Studying of honeybee orientation in space by provisory reflex method (Papers from U.S.S.R.) Moscow: State Publishing House for Agricultural Literature. 272 pp.
Lopatina, N. G., and E. G. Chesnokova, 1962. Motive stereotype of conditioned reflexes in the foraging activities of the honeybee. Trud. Inst. Fiziol. Pavlova. 10:245-254.
Martin, P., 1963. Die Steuerung der Volksteilung beim Schwärmen der Bienen. Zugleich ein Beitrag zum Problem der Wanderschwärme. Insectes sociaux. 10:13-42.
Maschwitz, U., 1964. Gefahrenalarmstoffe und Gefahrenalarmierung bei sozialen Hymenopteran. Z. ver gl. Physiol. 47:596-655.
Meyerhoff, G., 1955. Ueber die Demuthaltung bei der Honigbiene Apis mellifica L. Mit. dtsch. ent. Ges. 14:5-7.
Morse, R. A., 1963. Swarm orientation in honeybees. Science. 141:357-358.
——, and A. W. Benton, 1964. Notes on venom collection from honeybees. Bee World. 45:141-143.
——, and N. E. Gary, 1963. Further studies of the responses of honey bee (Apis mellifera L.) colonies to queens with extirpated mandibular glands. Ann. ent. Soc. Amer. 56:372-374.
——, N. E. Gary, and T. S. K. Johannson, 1962. Mating of virgin queen honeybees (Apis mellifera L.) following mandibular gland extirpation. Nature. 194:605.
Nixon, H. L., and C. R. Ribbands, 1952. Food transmission within the honeybee community. Proc. roy. Soc. B. 140:43-50.
Pain, J., 1961a. Sur la phéromone des reines d’abeilles et ses effets physiologiques. Ann. Abeilles 4:73-158.
——, 1961b. Absence du pouvoir d’inhibition de la pheromone I sur le développement ovarien des jeunes ouvrieres d’abeilles. C. R. Acad. Sei. Paris. 252:2316-2317.
——, and F. Ruttner, 1963. Les extraits de glandes mandibulaires des reines d’abeilles attirent les maies, lors du vol nuptial. C. R. Acad. Sei. Paris. 256:512-515.
Popper, K. R., 1957. “Philosophy of science: A personal report.” In C. A. Mace (Ed.), British Philosophy in the Mid-century. Macmillan, N.Y.
Renner, M., 1955. Neue Untersuchungen über die physiologische Wirkung des Duftorganes der Honigbiene. Naturwissenschaften. 42:589.
——, and M. Baumann, 1964. Ueber Komplexe von subepidermalen Drüsenzeilen (Duftdrüsen?) der Bienenkönigen. Naturwissenschaften. 51:68-69.
Ribbands, C. R., 1953. The Behaviour and Social Life of Honeybees. London: Bee Research Assoc. (Reprint: New York: Dover, 1964).
——, 1954a. Communication between honeybees. I: The response of crop-attached bees to the scent of their crop. Proc. R. ent. Soc., London (A) 29: 141-144.
——, 1954b. The defence of the honeybee community. Proc. roy. Soc. B. 142:514-524.
——, 1955. The scent perception of the honeybee. Proc. roy. Soc. B. 143:367-379.
——, and N. Speirs, 1953. The adaptability of the homecoming bee. Brit. ƒ. Anim. Behav. 2:59-66.
——, H. Kalmus, and H. L. Nixon, 1952. New evidence of communication in the honeybee colony. Nature. 170:438-440.
Root, A. I., 1950. ABZ and XYZ of Bee Culture. A. I. Root: Medina, Ohio.
Ruttner, F., 1957. Sexualfunktionen der Honigbienen im Dienste Ihrer Sozialen Gemeinschaft. Z. vergl. Physiol. 39:577-600.
——, 1964. Was wissen wir heute von der Begattung der Bienenkönigin. Südwestdtsch. Imker. 16:78-83.
——, and H. Ruttner, 1965. Untersuchungen über die Flugaktivität und das Paarungsverhalten der Drohnen 2. Beobachtungen an Drohnensammelplätzen. Z. Bienenforsch. 8:1-8.
Schweiger, E. M., 1958. Ueber individuelle Unterschiede in der Entfernungs-und Richtungsangabe bei den Tänzen der Bienen. Z. vergl. Physiol. 41:272-299.
Shaposhnikova, N. G., 1959. The comparative significance of two ways of foodscent transfer into the hive (on the body and in the honey sac) in the recruiting activity of the honeybees. Ent. Obozr. 38:523-528.
Simpson, J., and I. Riedel, 1963. The factor that causes swarming by honeybee colonies in small hives. J. apic. Res. 2:50-54.
——, 1964. The emergence of swarms from Apis mellifera colonies. Behaviour. 23:140-148.
Steche, W., 1957. Soziale Steuerung des Alarmwertes der Bienentänze Natur-Wissenschaften. 44:597-598.
Taber, S., III, 1954. The frequency of multiple mating of queen honeybees. J. econ. Ent. 47:995-998.
Taranov, G. F., 1947. (Occurrence and development of the swarming instinct in the colony.) Pchelovodstvo. 24:44-54.
Verheijen-Voogd, C., 1959. How worker bees perceive the presence of their queen. Z. vergl. Physiol. 41:527-582.
Weaver, N., 1966. Physiology of caste determination. Ann. Rev. Entomol. 11: 79-102.
Wenner, A. M., 1959. The relationship of sound production during the waggle dance of the honey bee to the distance of the food source. Bull. Entomol. Soc. Am. 5:142 (An abstract).
——, 1961. Division of labor in a honey bee colony—a Markov process? J. Theoret. Biol. 1:324-327.
——, 1962a. Sound production during the waggle dance of the honey bee. Anim. Behav. 10:79-95.
——, 1962b. Communication with queen honey bees by substrate sound. Science. 138:446-448.
——, 1964. Sound communication in honey bees. Sei. Amer. 210:116-124.
——, 1967. Honey bees: Do they use the distance information contained in their dance maneuver? Science. 155:847-849.
——, and D. L. Johnson, 1966. Simple conditioning of honey bees. Anim. Behav. 14:149-155.
——, and D. L. Johnson, 1967. Honey bees: Do they use direction and distance information provided by their dancers? (rebuttal). Science. 158:1076-1077.
——, P. H. Wells, and F. J. Rohlf, 1967. An analysis of the waggle dance and recruitment in honey bees. Physiol Zool. 40:317-344.
Woyke, J., and F. Ruttner, 1958. An anatomical study of the mating process in the honeybee. Bee World. 39:3-18.
Zmarlicki, C., and R. A. Morse, 1963. Drone congregation areas. J. apic. Res. 2:64-66.
——, 1964. The effect of mandibular gland extirpation on the longevity and attractiveness to workers of queen honey bees, Apis mellifera. Ann. ent. Soc. Amer. 57:73-74.