The Princeton Guide to Evolution

VII.7

Evolution of Communication

Michael D. Greenfield

OUTLINE

1. Elements of animal communication

2. What communication is

3. How does communication originate and how does it evolve?

4. Evolutionary trajectories: Four examples

5. On the reliability of animal communication

The evolution of animal communication remains one of the more fascinating questions in evolutionary biology, but it is also presents us with some of the more complex problems. Owing to the diverse and often elaborate ways in which animals send, receive, and evaluate messages, animal communication attracts considerable attention from a wide range of professional scientists and the lay public. The conspicuousness of animal communication to the human observer and the expectation that an understanding of communication in nonhuman animals may shed light on our own behavior—the origin of human language in particular—are additional factors that draw our interest. Nonetheless, understanding how animal communication evolves has proven to be particularly challenging, for several reasons. First, communication invariably comprises markedly distinct traits: sending messages and receiving and interpreting them. These separate traits may be subject to very different selection pressures, as when males are the message senders and females are the receivers, or vice versa, but at the same time such disparate traits cannot evolve independently of one another. Second, as with most behavioral traits, animal communication seldom leaves direct fossil evidence, and attempts to reconstruct the evolution of communication normally resort to indirect, comparative (phylogenetic) methods and other means of logical inference.

GLOSSARY

Pheromone. A chemical substance, comprising one or several compounds, that an animal emits outside its body and that influences the behavior or physiology of another individual of the same species.

Phylogenetic Analysis. Use of an evolutionary model such as a tree or other diagram that depicts evolutionary relationships among species, to infer the evolutionary trajectory of a particular trait. Importantly, the model used must be based on traits independent of the trait of interest.

Sensory Bias. A perceptual trait that is widespread in a group of related species and that evolved in a context other than communication. At a later point in evolution, communication signals to which this ancestral perceptual trait is particularly sensitive may arise in one or more of these species, in which case the signals are described as having evolved by “exploiting” a sensory bias.

Signal Intensity. The energy transmitted by a mechanical (sound or vibration) or visual signal or the amount of matter disseminated by a chemical signal, as measured at a particular time and location in reference to the signal source (e.g., molecules per cubic centimeter for a pheromone dispersed in air or water, or watts per square meter for a sound signal in air or water).

Signal-to-Noise Ratio. The intensity of a signal, such as an animal’s communication message, divided by the intensity of that signal’s modality (e.g., sound, light, vibration), as measured in the local environment at a time when the signal is not being transmitted; the relative conspicuousness of a signal against the background.

1. ELEMENTS OF ANIMAL COMMUNICATION

Animal communication comprises four elements: (1) the individual who plays the role of a signaler, (2) the signal that this individual sends, (3) the individual(s) who plays the role of a receiver of the signal, and (4) the channel along which the signal travels from the signaler to the receiver. Signals include chemical, mechanical (sound and vibration), visual, and electrostatic messages transmitted along an environmental channel that may traverse air, water, or the substrate en route from the signaler to the receiver. Not everything that can be perceived can also be a signal. Whereas some animals are capable of perceiving infrared radiation or the earth’s magnetic field, no evidence indicates that they can transmit signals and communicate in these modalities. A feature shared by communication signals in all modalities is some modification of the physical or chemical environment that stands out against the background noise and may be perceived by a receiver.

Various constraints imposed by the physical and biotic environment may favor the evolution of one signaling modality over another in a particular species and for a given aspect of its communication. For example, nocturnal species cannot rely on visual signals unless they can generate bioluminescence, as in fireflies and various marine crustaceans. Similarly, signals intended as alarms that warn conspecifics of impending danger may need to be sent very quickly, a prerequisite that could preclude the use of chemical messages where neighbors are separated by relatively long distances: except in the case of odors transmitted by direct contact between individuals or that serve as territorial markers, chemical signals travel largely by convection in air or water, that is, wind or current. Diffusion of odors would be too slow a process for all but the shortest interneighbor distances, and convection would not allow signalers sufficient control over the direction in which messages are sent. Thus, in many species alarm signals tend to be visual, acoustic, or vibrational.

The need for fine control over message transmission may further constrain signaling modalities. A signaler normally has intended receivers (e.g., potential mates or conspecific neighbors), as well as unwanted eavesdroppers (e.g., male rivals or approaching predators), and signals may need to satisfy conflicting demands: the message may be expected to reach a certain intended receiver or as many of them as possible, but it may also be expected to remain hidden from unintended receivers. These opposing pressures may select for signals that are transmitted along a “private channel,” one where the signaler can retain control over the dissemination of the message. For example, we might expect a signaler in an acoustic species to forgo its sound messages, which disperse across long distances and more or less in all directions, and use substrate vibration instead when predators are present. Whereas vibrational messages may be limited by distance and require a particular substrate, for example, vegetation having certain mechanical properties, it is this discrete nature of vibration that may prevent unintended receivers, save those few who happen to be situated on the same substrate, from perceiving the messages. On an evolutionary scale, this shift from airborne sound to substrate vibration is observed in various clades of acoustic insects: katydid species found in regions subject to high levels of predation by insectivorous bats that localize the sounds of their prey tend to rely more heavily on vibration. Similarly, males in various insect species that broadcast intense calling songs attractive to females over relatively long distances tend to replace sound with vibration and tactile signals in the courtship that ensues once an attracted female is proximate.

In other cases signals may be primarily selected to reach as many receivers as possible, perhaps with relatively little regard to eavesdroppers. Territorial markers may fall into this category, as such signals are normally intended for conspecific rivals. Here, signals whose broadcast continues even in the absence of the signaler may be favored to provide a more or less permanent message for potential usurpers of the site. Signals that mark territories tend to be chemicals that have been applied to the substrate in key locations, often along the region’s boundary. The only other type of signal that could satisfy this demand of permanence would be the special visual message provided by architectural construction. While such construction may play a role in the courtship communication of some vertebrate and invertebrate species—for example, Australian bowerbirds or fiddler crabs whose elevated burrow entrances influence female visitation—it does not appear to be prominent in communication about territories.

A very general constraint on signaling modality, largely overlooked, is the body size of the signaler—and of the receiver. Size poses special physical constraints for both acoustic and visual signaling: animals below certain minimum dimensions are generally unable to radiate sound waves into the surrounding medium—air or water—and they are equally unable to bear pressure-sensitive organs (tympana) sensitive to far-field sound waves. Thus, the smallest insects, like ants, do not tend to rely on acoustic communication for courtship or other functions. Moreover, species at the lower end of the size range at which acoustic communication is feasible are typically restricted to using high sound frequencies, which can be more efficiently radiated and perceived by small organs. Vision, too, is subject to a size constraint, as eyes, either compound or single lens, capable of forming images must exceed a minimum diameter. Again, the smallest insects appear not to use visual communication, or they extract only very crude information such as patterns of movement from a signaler. However, chemical communication is ubiquitous among animals, and it is also found among single-celled organisms. There are apparently no fundamental size constraints on the production and perception of odor, and the chemical modality may have been the first one to evolve in communication by living organisms.

A general property of communication is the transfer of “information” from the signaler to the receiver. Information in this context may be considered as a reduction in uncertainty about the identity, quality, ability, or intentions of the signaler. A signaler may also transfer information about its environment, such as presence of predators or value of local resources. Because effective information transfer is often subject to strong selection pressure, certain signaling modalities may be favored more than others. For example, reliable information about a signaler’s individual identity may demand considerable detail so that he or she can be distinguished from other members of the local population. Constraints on both the production and perception of signals may preclude some modalities because the information content of signals would be too low: either the signaler may be unable to fine-tune its broadcast—for example, only a single compound might be available as a chemical message—or the receiver may be unable to discern subtle variants of the message, such as different levels of concentration of an odor or minor changes in its chemical composition. Consequently, acoustic or visual communication might be used, at least for individual recognition.

2. WHAT COMMUNICATION IS

One view, by no means universal, is that communication occurs when a signaler transfers information and the receiver then modifies its behavior and/or physiology such that both individuals benefit (have a net increase in expected fitness). This definition would include many examples of messages sent between different species, including mutualistic animal-plant interactions such as pollination. Cases in which information is inadvertently provided by prey to potential predators, however, would not be considered communication. At the intraspecific level, traditionally the level at which communication is considered to occur, situations arise in which the stipulation of mutual benefits accrued by both parties might be questioned. In sexual communication, a male whose “quality” is relatively inferior may nonetheless encounter a female and mate, possibly by remaining as a “satellite” in the vicinity of a male who was broadcasting a superior signal. Thus, the benefit that the female accrued by perceiving and evaluating that superior signal is unclear. However, if satellites are rare and a female would normally pair with a high-quality mate by virtue of perception and evaluation of the signals, the stipulation of mutual net benefit would be upheld.

Communication may be further delimited as a process in which the information transferred by the signaler is perceived by the receiver’s nervous system, often to be stored in its memory. It is the retrieval of this stored information at a later time that results in the change in behavior and/or physiology noted earlier. According to this specification, communication would not include events in which one animal exerts physical force on a second animal, who then withdraws as a direct reaction to that force. Similarly, one animal might transfer a resource, such as a food item, to another, who then ingests the item and achieves improved growth and development as a result. Unless the second individual shows evidence of a behavioral response to the chemical, mechanical, or visual stimulus introduced by the food item, the interchange is simply nutritional.

The most difficult issue concerning the designation of an event as communication is the distinction between signals and cues. This distinction is most critical in the context of the present chapter because it is directly involved in the evolution of communication. Animals nearly always emit various inadvertent stimuli that may be perceived by another individual. In many cases such perception may benefit that individual, as well as the individual who emitted the stimulus, however inadvertently. For example, arthropods generally incorporate certain hydrocarbon substances in or on their cuticle that may function in protection against water loss or in other physiological processes, but these substances may also reveal the presence of the animal to other individuals. Thus, a male may localize and identify a conspecific female by detecting her cuticular substances, and even determine that she is mature and receptive. Here, both parties benefit from the message provided by the chemical stimulus, but does the message represent a communication signal? According to one perspective, it does not if the chemical stimulus shows no evidence of having undergone any specific evolution in the context of information transfer between individuals. Such evidence might include (1) energy expended specifically during the production and release of the substance that exceeds the expense necessary for the primary or original physiological function of the substance, (2) specialized structures that improve the release of the substance to the outside environment so that it can be more readily perceived by other individuals, and (3) a specific daily schedule for release of the substance that coincides with the periodicity of mating behavior and male receptivity rather than physiological needs, such as prevention of desiccation. Unless some of these indications are present, the stimulus is simply a cue, albeit one to which a male receiver responds. Should the system of chemical production and release evolve, however, toward a process in which any of the described features appear, one may consider the stimulus to be a specialized communication signal. In fact, it is not usual for cuticular hydrocarbons to evolve from ancestral, inadvertent cues to representing components of a sex pheromone.

3. HOW DOES COMMUNICATION ORIGINATE AND HOW DOES IT EVOLVE?

Animal communication requires an audience, either a specific receiver(s) that the signaler has perceived or expects to be in the surrounding area, or some unknown number of potential receivers who are likely to be present within broadcast range of the signaler. Thus, it is imagined that communication evolves in one of two ways, either by means of reciprocal modifications in the behavior of signalers and receivers that proceed in an alternating fashion or by means of modifications in the behavior of signalers in response to a preexisting preference or perceptual capability in receivers. Within the domain of sexual selection, wherein one typically focuses on the origin of mating signals, these two processes correspond with the well-known coevolutionary or Fisherian mechanism for the evolution of traits and preferences, or the evolution of traits by “exploitation” of sensory biases. In other words, the evolutionary origin and continued modification of signaler behavior in the context of communication does not seem likely without some corresponding evolutionary response in receiver behavior. A further clarification is that the descriptions of the two processes of behavior modification refer to events at their origin. Under the Fisherian (or coevolutionary) mechanism, evolutionary changes are expected in both parties, whereas under the sensory bias (or exploitation) process evolutionary change occurs only in signalers. However, the sensory bias process does not exclude the possibility that receiver traits may also undergo modification at a later point during evolutionary time. Rather, it is quite likely that a preexisting perceptual ability eventually experiences some fine-tuning in response to the newly evolved signal. Thus, the primary difference between the two processes is the relative timing of evolutionary change in signaler and receiver traits.

The need for an audience was recognized early in studies of animal communication, and one hypothesis proposed that communication could evolve provided the same or closely linked genetic elements controlled both signaler and receiver traits. This hypothesis, originally known as genetic coupling, was invoked on various occasions to explain how mating communication might change during the process of speciation: without such coupling, any change in mating signals would result in removal of the new variants from the population because no females would pay attention to them. The basic hypothesis continues to receive attention from evolutionary biologists studying species recognition in the speciation process as well as mate evaluation in sexual selection. Some limited support for the notion that common genetic factors pleiotropically influence both signaler and receiver traits has been found at the level of species recognition in some animal groups and signaling modalities, for example, visual communication in fish (Medaka) and Heliconias butterflies, acoustic communication in Hawaiian crickets (Laupala), and chemical communication in Drosophila. These findings invite the question, How might a motor trait and a perceptual trait share a common basis at mechanistic (physiological) and genetic levels?

4. EVOLUTIONARY TRAJECTORIES: FOUR EXAMPLES

A classical approach to the evolution of animal communication described a general process of ritualization, wherein a postural movement or displacement activity, perhaps initially serving a physiological function such as thermoregulation, becomes incorporated in a message. Over the course of evolution, the various components of the activity gradually become standardized. One may recognize elements of this so-called ritualization process in the sensory bias mechanism noted earlier, as well as in the general transition of cues to signals. The following four examples present possible trajectories by which different forms of animal communication may have evolved.

Oviposition Marker Pheromones

The egg-laying behavior of female tephritid flies (true fruit flies) includes locating a fruit of the host species that is in an appropriate state of maturation and then depositing an egg if another female did not do so previously. The focal female determines whether she has been preceded by an earlier female via detecting an odor left on the fruit surface during oviposition. By sensing and responding to this odor, an oviposition marker pheromone, the focal female avoids having her offspring compete with a larva that is probably in a more advanced state of development and thus likely to win the resource in the event of a contest. By leaving an odor on the fruit surface, the first female to arrive prevents her offspring from having to compete for the resource, a desirable result even though her offspring would probably win. A parsimonious hypothesis for the evolution of oviposition marker pheromones is that they originated as a chemical cue left inadvertently on the fruit surface during deposition of an egg. Because a subsequent female would be under particularly strong selection pressure to shield her offspring from competition, any perceptual ability allowing her to detect the cues of a prior oviposition would have been favored. Once such detection had evolved to a certain level of sensitivity, selection would then have favored some modification of the cue rendering it more conspicuous and thereby ensuring that the offspring of the first female would avoid a costly encounter. Thus, a cue left during oviposition would have evolved into a specialized marker pheromone. Eventually, some refinement of the avoidance responses to the specific marker pheromone would also be anticipated. Similar signals may have evolved in parasitoid wasps that deposit one egg or larva on a host insect. In general, one might expect the evolution of this mode of communication in cases where a resource patch can accommodate one and only one offspring; while sufficient resource patches exist it will pay a second female to leave the patch and search for a previously undiscovered one.

Dance Language of Eusocial Bees

The communication by which Western honey bees (Apis mellifera) recruit numerous foragers to a distant food source has been largely deciphered owing to a series of painstaking observations and ingenious experiments performed over the past century by Karl von Frisch and his colleagues and students: having found a valuable patch of flowers, a worker honey bee returns to the colony and repeatedly performs a figure eight movement on a vertical surface in the darkness of the hive interior. Other workers sense the orientation and duration of the figure eight movement via substrate vibration, tactile stimuli, and near-field sound emanating from the dancer, and they use this information to infer the azimuth, that is, the horizontal angle between the sun and the floral patch as seen at the hive entrance, and approximate distance to the patch. Odor cues—scent from the discovered flowers—on the dancer may fine-tune the navigation of the recruited worker once she arrives in the vicinity of the resource patch.

Studies by Claudia Dreller and Wolfgang Kirchner on other members of the genus Apis found in Asia shed some light on how the complex recruitment communication in A. mellifera may have evolved. Various Asian honey bee species, such as A. dorsata, A. florea, and A. laboriosa, do not nest within enclosed cavities, and successful foragers in these species perform recruitment dances in the open while moving on upper, horizontal surfaces of the colony. Notably, dance signals in these species do not include sound. A. florea is believed to be the extant species of Apis closest to the root of the genus, suggesting that this dance style may retain some aspects of the ancestral form. One Asian species, A. cerana, is a cavity nester like A. mellifera and, like A. mellifera, performs its dances on vertical surfaces in the dark interior of the colony, and the dances do include sound. We can infer that foragers in the ancestral Apis species made certain displacement movements on the exposed hive surface on returning to the colony and that these movements reflected the distance they had just flown and the direction in which they might depart on their next trip. Thus, colony members would have been under some selection to pay attention to these cues because their own foraging would then be more efficient. That is, the colony would have served as an “information center” at which workers could learn the local distribution of floral resources. But because the colony is more than just a center at which information might be passively acquired, each colony member would benefit from an increased foraging efficiency by other members. Consequently, it has been proposed that selection favored the modification of inadvertent cues into specialized signals, that is, a dance that conveyed more accurate information about those resources. The switch to nesting inside dark cavities in the more derived species of Apis would have then demanded two further modifications of dance signals. First, these honey bee species could no longer rely on vision for evaluating information in the dance, but the incorporation of near-field sound signals appears to have retained high information content. Second, in the absence of an exposed horizontal surface on the colony, workers could not directly indicate the direction toward floral resources by means of dance movements. However, the development of an indirect mechanism in which the angle between the straight segment of the figure eight and a vertical line represents the horizontal angle between the sun and the floral patch seems to have solved this difficulty. It is this abstract representation of direction that has led these honey bee dances to be commonly referred to as a “language.”

Courtship Pheromones in the Lepidoptera

Long-distance communication during pair formation in the Lepidoptera typically involves visual signals (butterflies) or advertisement pheromones emitted by females (moths), but once the male and female have established contact, male courtship pheromones often mediate the final outcome of the encounter. Chemical analyses in some species indicate that these male courtship pheromones may be derived from substances acquired during feeding by larvae in some cases, and by adults in others. In the arctiine moth Utetheisa ornatrix, a species studied intensively in the laboratory of Thomas Eisner at Cornell University over many years, the courtship pheromone is a volatile substance that the male produces by converting a chemical, a pyrollizidine alkaloid (PA) acquired from the host plant—several species of legumes in the genus Crotalaria—during larval feeding and sequestered in the body. Males also transfer a small quantity of this unconverted host-plant chemical (PA) to the female along with sperm at the time of mating, and the toxic properties of the PA appear to confer some protection against natural enemies to the female, as well as protecting her eggs. Thus, transfer of the PA represents a “nuptial gift,” and the courtship pheromone serves to signal the presence, and possibly the quantity, of this gift to a discriminating female. Because both male and female U. ornatrix perceive and respond positively to various chemical cues from their host plant, one may propose that the male courtship pheromone evolved from an ancestral cue, the PA in this case. Evolution of the pheromone may have taken place for any of several reasons. First, the PA in its original form is probably not sufficiently volatile to serve in communication, even at close range. Second, by converting the PA to another substance, the male may achieve an improved signal-to-noise ratio when courtship occurs on or near the host plant: How would a female discern a courting male from the background odor of the plant were he to use an unconverted host-plant chemical as a pheromone? Finally, the primary behavior that an ancestral cue such as the PA might elicit in U. ornatrix females may be oviposition, not mating. However, by using a different but chemically related substance as a courtship signal, a male ensures that a female will still perceive and respond but that her response will occur in a different (nonoviposition) context, such as courtship. This specific courtship response assumes that the female receiver, like the male signaler, has undergone some modification during the course of evolution.

Acoustic Communication in the Lepidoptera

Whereas the vast majority of pair formation in moths is accomplished via long-range advertisement pheromones emitted by females and short-range courtship pheromones presented by males, a small but critical percentage of species use sound in mating communication. How did these exceptions originate, and what general lessons about the evolution of communication can we learn by studying them? We begin by noting that the perception of sound is actually quite common in moths, being prevalent in three major superfamilies (Pyraloidea, Geometroidea, and Noctuoidea) and having evolved independently between 7 and 10 times. Hearing in moths occurs largely in the domain of ultrasonic (> 20 kHz) frequencies and appears to have evolved and to function in the context of avoiding predatory bats: phylogenetic and biogeographic analyses suggest that moth hearing arose coincident with the origin of echolocation signals in insectivorous bats, about 55 million years ago. This inference is bolstered by the observation that hearing has been lost secondarily in various moth species found in regions lacking insectivorous bats or that fly at times when bats do not hunt. That is, the primary function of moth hearing is the detection of ultrasonic echolocation signals emitted by bats, which then allows the moth to make appropriate flight movements and evade predation.

Against the evolutionary background of widespread hearing, some moth species also emit sound, in most cases in the ultrasonic frequency range. In some of these species the sounds function, like hearing, in defensive behavior: they may “jam” the bat’s echolocation system, or they may serve as an aposematic (warning) signal in some chemically protected moth species. But in a few phylogenetically unrelated species of pyraloid and noctuoid moths the sounds are produced only by males and serve either in close-range courtship (similar to the use of the male pheromone by U. ornatrix, described earlier) or as a long-range advertisement to females. In these latter species the sounds are essentially the equivalent of male calling songs found in more familiar acoustic species such as crickets or frogs.

Given the relative prevalence of hearing and rarity of courtship and calling songs, it is most likely that acoustic communication in moths evolved via “exploiting” the ancestral auditory perception. Beyond this basic inference, however, various questions remain unsolved. As in U. ornatrix, a female moth would not be expected to respond to a male sound as if it were a courtship message but rather in an antipredator manner, that is, by undertaking a negative, evasive movement. Thus, the origin of acoustic communication in this group is not completely clear. One possible explanation is that moths situated on the substrate, as courting females normally are, may evade bats by remaining stationary, a response that eliminates inadvertent sound that can reveal the moth’s presence to a bat searching for prey on vegetation or on the ground. Consequently, a male who emits an ultrasonic song may render a nearby female immobile and therefore more readily courted. But in some moth species that broadcast long-range male calling songs, the male call is delivered with a rhythm that is fundamentally different from that found in most bat species. Here, females distinguish between bat echolocation signals and male calls, and they actively run toward the latter. Evidently, the female response, as well as the male signal, has experienced some fine-tuning subsequent to the initial appearance of male song.

5. ON THE RELIABILITY OF ANIMAL COMMUNICATION

While examples of cheating and deception among nonhuman animals always draw considerable interest from human onlookers, it is becoming increasingly evident that animal communication is by and large an “honest” affair. Honesty in this sense refers to the communication signal’s being a reliable indication of the signaler’s identity, quality, physical ability, intention, information, and so forth. The expectation that signals are reliable is based on economic grounds: should receivers pay attention to signals that do not reflect the quality of the signaler, for example, they will ultimately suffer reduced fitness. Under such circumstances, receivers are likely to diminish and eventually cease their evaluation of these signals, which, in turn, should influence signalers to discontinue such broadcasts. Production of ignored signals would not be worth the inevitable costs in time, energy, and risk.

It is argued that the need for signal reliability has played a major role in shaping the evolution of animal communication wherever the potential for conflict exists between signalers and receivers. Importantly, some degree of conflict is expected in most interactions, including those between members of the same species. For example, a male and a female may have a common goal of producing offspring, but they also have very different specific objectives related to achieving this goal. A male might be less discriminating and be expected to mate with most females encountered, whereas a female may forgo mating with a particular male if he is of low quality, and future mating possibilities exist. Thus, any signal that pretends to represent male quality would have to be reliable in that it could be produced only by a male who actually bears that quality. The male courtship pheromone of the moth U. ornatrix shows how signal reliability might function. The chemical precursor of this pheromone is a substance in the host plant that forms part of a nuptial gift transferred at mating, and the only way the male can acquire the precursor is by feeding on the host plant and sequestering it. Thus, on detecting the pheromone, a female is at least assured that the courting male has successfully foraged on and accumulated the substance—behaviors that might be inherited by her offspring. Additionally, the male can potentially transfer a nuptial gift, the size of which may be commensurate with the concentration of the pheromone, although it is yet possible that the male could withhold his offering at the last moment. That is, he might conserve his gift materials for a future mating opportunity, particularly in the case where he has evaluated the female to be of low reproductive potential.

Another common way in which signal reliability can function is via intensity. If the broadcast of an intense signal demands considerable energy, and available energy is an indication of male quality, only high-quality individuals are expected to signal strongly. Should a low-quality individual imitate a high-quality one by broadcasting an incongruously intense signal, he would deplete so much of his energy store that he might not be able to avail himself of the rewards normally accruing to a strong signaler. Thus, signal strength is expected to be proportional to quality.

Signal reliability may be much less critical in communication in highly social species, particularly those in which most individuals tend to be close genetic relatives. Interindividual conflict is expected to be relatively weak here, and there are few a priori reasons for individuals to broadcast signals that misrepresent themselves or their information. For example, the dance language in honey bees, while not perfectly accurate, is a reasonably good indication of the location of floral resources, and there are no apparent mechanisms by which this accuracy is safeguarded against returning foragers who might deliberately mislead recruited individuals.

The preceding arguments for honesty notwithstanding, dishonest communication is sometimes observed in animals, and it merits our consideration. Some cases of unreliable signals can be explained as communication that on average affords a net benefit to the parties concerned. That is, signals may misrepresent quality or intention until a certain point, beyond which their deception would be selected against. But perhaps a more common source of signal unreliability is the dynamic nature of evolution itself. Populations are continually subject to environmental change, as well as to the arrival of migrants from neighboring populations, in which traits may be somewhat different. Our observations of animal communication are only snapshots of signaler and receiver traits, which may sometimes be less than fully reliable: for example, females may evaluate male signals according to criteria that promised quality under previous environmental conditions but not necessarily under current ones. Thus, the level of reliability in animal communication may reflect the extent to which populations have attained an equilibrium state in which traits are fully adapted to an environment that is, for the moment, stable.