Chemical releasers, or pheromones as they are now commonly called (Karlson and Butenandt, 1959), were probably the first signals put to service in the evolution of animal communication. Haldane (1955) argued that communication among protozoan cells must have preceded the formation of metazoans, and this primitive communication was almost certainly chemical. Since cells in the metazoan body communicate with each other by hormones, a lineal evolutionary relationship between pheromones and hormones is indicated. The process of conversion from external to internal signals probably continued for some time into the evolution of the Metazoa. Whatever its relation to endocrine evolution, true pheromone communication has persisted in almost all animal taxa, and chemical systems turn up with regularity in most animal species where a deliberate search is made for them.
SOME GENERAL PROPERTIES
Modes of Action
Pheromones are secreted as liquids and transmitted either as liquids or gases. Some are normally detected at a distance by olfaction, others at the surface of the body of the emitting animal by smell or taste, and still others are encountered in liquid phase at olfactory “signposts” to be smelled by transient animals over relatively long periods of time. The substances function either by evoking immediate behavioral responses, called releaser effects, or by means of endocrinemediated primer effects that alter the physiology and subsequent behavioral repertoire of the receptor animal. Examples of releaser pheromones include the sex attractants and alarm substances. Examples of primer pheromones include the odors that alter rodent reproductive physiology (Parkes and Bruce, 1961). A single substance can have both releaser and primer effects. For example, 9-keto-2-decenoic acid serves as both the olfactory sex attractant of virgin honeybee queens and the signal by which the nest queens inhibit worker ovarian development and the rearing of new queens (Butler, 1964).
Chemical systems are capable of extreme energetic efficiency. Butenandt and Hecker (1961) have synthesized a geometric isomer of bombykol, the female sex attractant of the silkworm moth (Bombyx mori), which is still effective when diluted to 10-12 µg/cm3 of petroleum ether and presented in a film on the tips of glass rods 1 cm from the antennae of males. It can be estimated from simple diffusion models that no more than a few molecules strike the chemosensitive sensilla trichodea. In fact, Boeckh et al. (1965) have shown by means of the airstream-dilution technique that one molecule can fire a sensillum trichodeum. Experimental evidence of similarly low thresholds in other insect sex attractants is accumulating rapidly. For example, using only 0.004 µg of purified sex pheromone from females of the sawfly Diprion similis, Casida et al. (1963) were able to attract 500 to 1000 males over 100to 200-foot distances within five minutes in the field. It is therefore not surprising to find the quantities of pure pheromone produced by single individuals to be very small, as shown in Table 1. In passing it can be noted that although a single Bombyx mori female carries only about 1.5 µg or less of bombykol at any given moment, theoretically this is enough to activate at close range over one trillion males, probably more than exist in the entire world at any one time.
Organic odorants provide an immense array of potential signals. With an increase in molecular weight in any given homologous series, molecular diversity increases exponentially. In the methane series alone, to take the simplest possible example, only 3 geometric isomers can be constructed from 5 carbon members; but with 10 members 75 isomers become possible, and with 20 members 366,319 isomers become possible. If the substitution of double bonds and oxygen atoms is then permitted, the number of molecular species increases much more sharply. Since pheromones often contain ten or more carbon members, the potential number of odorants must be vastly greater than the number of actual pheromones in existence.
Pheromones are nearly unique in one feature: they can be left behind as a continuous signal after the animal has departed. We find chemical secretions commonly used as territorial and trail markers as well as individual nest odors. Slow fade-out, while convenient under special circumstances, has the potential disadvantage of severely limiting the rate of information transfer, at least in comparison with optical and acoustical systems. Yet it is possible to design special chemical systems with surprisingly rapid transfer rates. Wilson (1962) found that the total rate of spatial information transmitted in the fire ant odor trail is 0.1 to 5 bits/sec, a quantity fully comparable in magnitude to the transfer rate of the purely spatial information in the waggle dance of the honeybee. Bossert (1967) proved that patterned transmission of pheromones, consisting of amplitude and frequency modulation, is technically feasible over very short distances or in a steady, low wind. Under ideal conditions a pure substance could be used to transmit in excess of one hundred bits of information, or the equivalent of over twenty English words, per second. Patterned transmission has not yet been demonstrated in nature. But we should remember that it would be a most difficult property to measure or even detect, and biologists have not looked for it in the first place.
Wilson and Bossert (1963) argued that most classes of pheromones could be expected to display a carbon number within 5 to 20 and molecular weight within 80 to 300. Very briefly, the reasons are as follows. Below the lower limit, only a relatively small number of kinds of molecules can be readily manufactured. Above it, molecular diversity increases very rapidly. In at least some insects, and for some homologous series of compounds, olfactory efficiency also increases steeply. As the upper limit is approached, molecular diversity becomes astronomical, so that further increase in molecular size confers no further advantage in this regard. The same thing is probably true for intrinsic increases in stimulative efficiency, in so far as they exist. On the debit side, large molecules are energetically more expensive to make and to transport, and they tend to be far less volatile. Wilson and Bossert further predicted that the molecular size of sex pheromones, which generally require a high degree of specificity as well as stimulative efficiency, would prove higher than that of most other classes of pheromones, including, for example, the alarm substances. These generalizations have so far held fairly well in the face of recent chemical identifications, a point that will be taken up again in the discussion of specific functional categories.
METHODS OF ANALYSIS
Because of the special difficulties encountered in the identification of the pheromones, work on chemical systems tends to progress slowly. Faced with the problem of the extreme dilution of pheromones in the bodies of the signaling organisms, biologists must gather large quantities of fresh material, and chemists must learn to work with milligram or even microgram quantities of the pure material. Purification often entails unusual technical problems. Butenandt, Hecker, Beroza, Jacobson, and their associates were forced to extract hundreds of thousands of female moths to obtain on the order of 10 mg of bombykol and glyplure (Jacobson, 1964). Walsh et al. (1965) obtained only 0.25 mg of pure trail substance from 200,000 fire ant workers. Once purified, the substances often prove difficult to identify. Methyl-2,3-dihydro-lH-pyrrolizidin-l-one, one of the major components of the male secretion of the butterfly (Lycorea ceres), is a novel substance in animals; its identity was revealed only by exceptional skill on the part of Meinwald and his associates (1966). The structural formula of the female sex attractant of the American cockroach (Periplaneta americana) originally proposed by Jacobson, Beroza, and Yamamoto was so complex and novel that it caused immediate controversy among organic chemists and required efforts by several research groups to synthesize it. Finally, Day and Whiting (1964) were able to show that the formula was incorrect. They postulated a similarly complex compound with the same empirical formula, but at the time of their first publication they had not succeeded in its synthesis. As special techniques in microanalysis are improved and the necessary collaboration between chemists and biologists increases, the rate of identifications can be expected to increase rapidly. A next step is the study of biosynthesis and transport within the body of the insect. Already Happ and Meinwald (1965) have used radioactive-labeled compounds to show that citronellal, an alarm substance of the ant Acanthomyops claviger, can be synthesized from acetates and mevalonates and does not need to be acquired in finished form from the food.
Once identifications are made, or at least the glandular sources of the pheromones are pinpointed, analyses of the physical properties of the transmission can be undertaken. Bossert and Wilson (1963) have developed a series of diffusion models of pheromones released under a variety of conditions, making it possible to relate the Q/K ratio (ratio of numbers of molecules released to response threshold in molecules per cubic centimeter) to the fade-out time, maximum radius of the active space (the space within which the pheromone is at threshold concentration or denser), and the time required to reach the maximum radius. An example is given in Fig. 1. These models have been used in conjunction with laboratory measurements to estimate Q and K separately in cases where these parameters would be extremely difficult or impossible to measure by direct means. The Q/K ratio was further found to vary greatly among different kinds of chemical systems in ways that appear intuitively adaptive.
Where a direct releaser effect is obtained with crude extracts or gland squashes, a bioassay is easily designed. The bioassay need not measure all of the releaser and primer effects of the molecule; instead, it can consist of but one reliable component. Thus Butler and his associates used the presence or absence of queen-cell building activity of small groups of queenless worker bees to detect 9-ketodecenoic acid. Schneider (1962) invented the “electro-antennogram,” or display of the directly recorded pattern of discharge of the antennal nerve, to test male chemoreceptor responses to moth sex attractants. The EAG is an interesting and useful adjunct technique; but, at least in its initial crude application, where most or all of the neurons were monitored together, it has proved somewhat less sensitive than behavioral tests.
Behavioral tests can be conducted either on animals that move freely in the wild or within laboratory cages or else on animals confined in specially designed olfactometers and multiple-choice arenas (see, for example, Gaston and Shorey, 1964). These can be regarded as a crude first step. They allow very little to be inferred in most cases about the site of the chemoreceptors or the mode of orientation. An exception is the recent work of Lindauer and Martin (1963), who, by a series of ingenious experiments, showed that honeybee workers employ their antennae to scan for odor gradients. So long as both antennae were allowed to sample the air in a fixed position or at least one antenna continued to sweep the air, the insect was able to move directly up the odor gradient toward the odor source. In a word, bees can orient by osmotropotaxis.
The molecular basis of olfaction has been the object of speculation since the time of Faraday and Tyndall and has been approached anew in recent years by means of biophysical models. J. E. Amoore, J. T. Da vies, and R. H. Wright have attempted to relate odor “groups” in human smell to common properties in the odorant molecules. The subject has most recently been reviewed by Wright (1966a, b). Since it is for the moment at least only peripheral to pheromone communication, it will not be considered further here.
It is intriguing to look briefly toward a field of inquiry that is just about to come into being. Since pheromones frequently differ among closely related species and, in some cases, even among related individuals of the same species, genetic analyses of the control of pheromone production are feasible. A genetic analysis in the chemistry of insect quinones recently undertaken by Engelhardt et al. (1965) is at least very suggestive. These authors discovered an autosomal recessive mutant that modifies the adult-instar quinones of the beetle Tribolium confusum in two ways. As the beetles age, the quinones (2-ethyland 2-methyl-l,4-benzoquinone) present in exocrine gland reservoirs are reduced to one-twentieth of the amount found normally in the wild type, and the contents may further be modified into a highmolecular-weight polymeric substance which occurs as a black solid mass. In Tribolium, quinones are used as pheromones in a primitive manner. Mutations of lesser magnitude would be enough to alter the information content of most pheromones, which are generally manufactured in exocrine glands and can be deactivated with less severe chemical change.
A classification of pheromones according to function of the evoked behavioral acts is intrinsically crude and imperfect, but it at least serves to indicate the diversity of forms of communication dependent on chemical signals.
The members of some animal species congregate prior to feeding, mating, hibernation, etc., without the assembly signals being in any way directly related to the subsequent activity. The workers of social insects are especially prone to clustering, a behavioral response that is often most strikingly manifested when individuals are removed from the nest. Verron (1963) found that the principal attractant in the termite Kalotermes flavicollis is 3-hexen-l-ol, which is apparently manufactured in the gut. Wilson (1962) showed that in the fire ant Solenopsis saevissima the principal volatile attractant is apparently simply carbon dioxide, a waste product normally found in high concentrations inside the teeming nests of social insects.
Many aposematic (warningly colored) insects have the interesting habit of forming conspicuous aggregations. Since they are distasteful, and predators quickly learn this fact, the aggregations serve to increase the repellent effect and give superior protection to the individual members. Eisner and Kafatos (1962) discovered that the assemblies of the yellow-orange beetle Lycus loripes are convoked entirely by means of an olfactory pheromone secreted by the adult males.
Sex pheromones produce simple attraction, sexual behavior, or both. In many cases, attraction occurs at low concentration and sexual behavior at high concentrations as the responder draws near the source. Sex pheromones are best known in insects, but they have also been recorded from crustaceans, fish, salamanders, and mammals [see the review by Wilson and Bossert (1963) and more recent articles by Meinwald et al. (1966) on butterflies, Ryan (1966) on crabs, Schultze-Westrum (1965) on phalangers, and Mykytowycz (1965) on rabbits]. As Jacobson (1964) has documented at length for the insects, either sex or (in a few species) both sexes can produce the pheromones. An interesting example of male pheromonereleasing behavior is given in Fig. 2. In many species of butterflies, moths, true bugs, roaches, beetles, and flies, the male presents an aphrodisiac substance to the female that renders her more susceptible to his sexual approach. It is possible that in some Hymenoptera and mammals the same substances are used as both territorial markers and sexual stimulants, but this notion cannot be tested until precise bioassays are developed and chemical identifications achieved. In some groups sex pheromones by themselves are adequate to release full sexual behavior toward any object carrying the scent. Examples include the female pheromones of moths and honeybees. In other groups auxiliary stimuli are required. In the species of salticid spiders studied by Crane (1949), both visual stimuli and female odors are required to initiate courtship behavior in the male.
Wilson and Bossert (1963) predicted on a priori grounds that the optimal size of sex pheromones falls in the upper parts of the range 5 to 20 in number of carbon members and 80 to 300 in molecular weight. An examination of Table 2, in which are listed the known and probable sex pheromones identified to the present time, shows that the empirical evidence still matches this rough theoretical expectation. There is one striking apparent exception—the terpenoid alcohols produced by the Lasius and Acanthomyops ants. The evidence implicating these substances as pheromones is strong but still purely circumstantial. If they are truly sex substances, they are remarkable in another respect. They occur in medleys, each species manufacturing a blend in which the proportions of the constituents are peculiar to the species. Thus the male ant pheromones deviate from other sex substances in their smaller molecular size, but they contain an additional feature that has the same effect of increasing diversity.
Perhaps it is not premature to ask at this point why some species employ medleys while others rely on single compounds. A possible explanation, which apparently has not been considered before, is that medleys are employed at close quarters, whereas single substances are employed in more distant communication. A medley can generate very great signal diversity; it can also take advantage of synergistic effects among differing substances. But at a distance it tends to lose its specificity because of the effects of Graham’s law. In the butterfly Lycorea a medley of substances is applied directly to the antennae of the female. It will be interesting to see if the species of Lasius and Acanthomyops and other medley-producing species also employ their pheromones at relatively short distances.
Although the variation in chemical structure among sex pheromones is great, it is not true, as was once thought, that all such pheromones are species-specific. They do tend to be differentiated at a low taxonomic level, at most, at the level of the family or genus. In some cases they are specific at the species level as, for example, within the scarabaeid beetle genus Rhopaea (Soo Hoo and Roberts, 1965) and mecopteran genus Harpobittacus (Bornemissza, 1966). Within the moth families Pyralidae, Lymantriidae, Saturniidae, and Noctuidae, on the other hand, specificity is slight or lacking in the small samples of genera and species thus far tested (R. Barth, 1937; Schwinck, 1955; Schneider, 1962; Shorey et al., 1965). In roaches, pheromones tend to be genus-specific but not species-specific in action (R. H. Barth, 1962). As Wilson and Bossert (1963) showed in their analysis of the data of Schneider and of Rau and Rau on the North American saturniids, weak pheromone specificity is supplemented by species differences in seasonality and diel schedules. In combination, all these differences result in strong, prezygotic sexual isolation among all the saturniid species sampled.
It is thus well demonstrated that differences exist among major taxa of insects with respect to the level of sex pheromone differentiation. At the moment, no good explanation has been offered to explain this newly discovered phenomenon. Perhaps the existence of multiple prezygotic devices in the Saturniidae offers us a clue. It is possible that differentiation is evolved down to the species level in groups where opportunities are lacking for the formation of other prezygotic isolating mechanisms.
Bossert and Wilson (1963) estimated the Q/K ratio for gyplure, the sex attraction of female Porthetria dispar, to lie in the range 1.87 X 1010 cm3/sec to 3.03 X 1011 cm3/sec, where Q is the number of molecules released per second. These estimates were based on an assumed response threshold density K similar to that of Bombyx mori. Even if the Porthetria K were several orders of magnitude higher, the Q/K ratio would still be extremely large. As a consequence, females of these and other moth species are able to “call” downwind for distances of hundreds or thousands of meters by releasing only microgram quantities, or less, of the pheromone per second. The active space, that is the space within which the pheromone concentration is at threshold or greater, is a huge semiellipsoid which forms downwind from the emitting female (Fig. 3). Within most of the space, the molecular concentration provides too shallow a gradient for insects to follow by osmotaxis. In most cases studied to date the males respond by becoming positively anemotactic on entering the active space; in other words, they fly directly upwind on smelling the pheromone, a maneuver that eventually brings them to the female.
Territory and Home-range Marking
The use of chemical trails and scent spots to mark territories and home ranges is widespread in the mammals. The odorants are either passed in excrement or else secreted by special exocrine glands, according to taxon, in one or more of a variety of body sites: pedal, carpal, tarsal, metatarsal, preorbital, occipital, caudal, preputial, anal, etc. European rabbits, for example, mark their territories by “chinning” objects within them, which leaves behind secretions from the submandibular glands (Mykytowycz, 1965). The flying phalanger Petaurus breviceps uses secretions from at least three glands in territorial marking (Schultze-Westrum, 1965). Our knowledge of the chemistry of these substances is still virtually nil. The Crocodilia also release scent associated with territories, at least during the mating season. In the South American alligatorids Caiman sclerops and C. latirostris, musk glands located near the anus yield 4 to 35 grams of a mixture composed of about 80 per cent fatty acids (including isovaleric and palmitic), some cholesterol, some cetyl alcohol, and about 4 per cent “yacorol,” which is principally d-citronellol (Lederer, 1950). The males of bumblebees and some solitary bees mark their flight paths with scent. Male bumblebees (Bombus) daub vegetation with scent spots placed at intervals in a circuit, around which they travel hour after hour and day after day (Haas, 1952).
The intraspecific use of repellents, locomotory stimulants, or deterrents should result in increased species dispersal. Adults of the flour beetle Tribolium confusum aggregate at low population density, apparently distribute themselves randomly at intermediate density, and distribute uniformly at high density (Naylor, 1959). The last effect is evidently due to the secretion of quinones, which act as repellents above a certain concentration. Loconti and Roth (1953) have shown that these substances are produced by thoracic and abdominal glands and, in the case of T. castaneum, consist primarily of 2-ethyl-l,4-benzoquinone and 2-methyl-l,4-benzoquinone. A third (trace) substance, 2-methoxy-l,4-benzoquinone, is not effective as a repellent. A similar effect may be responsible for the dispersal of young Zinaria millipedes, which secrete a substance, apparently hydrocyanic gas, during dispersal from their initial aggregations. However, the evidence is anecdotal. Dispersal can also be achieved by a reproductive deterrent, as shown by Salt (1936). Females of the parasitic wasp Trichogramma evanescens receive chemical cues from eggs already parasitized by other females and avoid ovipositing on them. The discrimination has been referred to as “perfect” by Salt, and further experimental evidence makes its adaptive significance clear: when host eggs were superparasitized in laboratory studies, the resulting young were inviable or at least abnormal.
Recognition of Group and Rank
The use of odors to recognize members of a group is a universal feature in the social insects. In most, but not all species, workers are able to discriminate nest mates from aliens. Experiments by Nixon and Ribbands (1952) show that such recognition scents in the honeybee are derived at least in part from diet, and Lange (1960) has proved that both diet and the chemical nature of the nest wall can contribute to colony odor in the ant Formica polyctena. In addition to such extrinsic components of the epicuticular odor, there are genetically determined components which operate to allow workers to discriminate members of alien species and probably, to some extent, of alien colonies as well. Nothing is known concerning the chemistry of the colony odors. Until such information is forthcoming, it is futile to speculate on the relative contributions of the genetic and phenotypic variances within and among species.
Recent discoveries have emphasized the role of odors in group recognition and organization in some of the mammals. Individuals of the phalanger Petaurus papuanus are able to discriminate odors at the specific, group, and individual level. In addition to marking their territories with their own distinctive secretions from several glandular sources, males also use specialized behavior to smear frontal gland secretion over the bodies of their partners (Schultze-Westrum, 1965); see Table 3. Mykytowycz (1965) found that the male rabbits contribute most of the submandibular gland secretion to territorial marking. In high-density experimental populations the total size and histological development of the gland is correlated with social status.
When houseflies (Musea domestica) visit a bait, they contribute to it a “fly-factor” substance that increases its attractiveness to other flies (Barnhart and Chadwick, 1953). Beyond this primitive form of chemical recruitment, several highly evolved systems can be found in the insects. Certain species of ants, termites, and meliponid bees use odor trails regularly as part of their foraging behavior. Among these the behavior of the fire ant Solenopsis saevissima has been most intensively investigated (Wilson, 1962; Bossert and Wilson, 1963; Walsh et al., 1965). Workers of this species forage singly while visually orienting with respect to the nest. On finding a food particle too large to carry home, an individual returns nestward while laying down minute traces of secretion in broken streaks. The pheromone originates from Dufours gland in the abdomen and is dispersed at the estimated rate of only at most 6 X 1015 molecules/cm. The substance is a potent attractant. When it is blown as a gas into a nest interior, the ants inside move toward its source. A majority of the workers, and on occasion even the queen, can be attracted in this way. When drawn out in artificial liquid trails from an applicator, Dufours gland secretion causes behavior indistinguishable from natural trail-following. The pheromone is relatively volatile: a natural single trail evaporates within a minute or two to below threshold density, depending on the absorptive power of the substrate. This means that a trail-laying worker can pinpoint a food source only within a meter or less of the nest, since the trail fades out before other workers can follow it for greater distances. Rapid fade-out has the obvious advantage, however, of reducing “noise” in the trail system around the nest. The active space is a semiellipsoid which maintains a diameter of about two centimeters for most of its length and approaches the maximum density gradient at the fade-out point. Both features contribute to the orienting power of the pheromone. The Q/K ratio is about unity, the lowest yet recorded for a pheromone system, indicating a relatively high behavioral threshold. Trail substances of ants tend to be species-specific in action, thus affording privacy in communication among species whose foraging territories overlap (Wilson, 1962; unpublished observations on the Attini by Margaret Dix, personal communication).
Recruitment leading to cooperation is not limited to the social insects. The first-ins tar larvae of the sawfly Neodiprion pratti have difficulty biting into the tough skin of jack pine needles, their preferred food. When one does break through to the inner tissues, volatile attractants are released from the plant as well as from the saliva of the successful larva. Other larvae soon congregate in the vicinity of the breakthrough, and the needle is then efficiently opened up as a result of group activity (Ghent, 1960).
A similar and even more dramatic form of communication is employed by adults of scolytid bark beetles. When individual males of Ips confusus make a penetration into the phloem-cambial tissue of a host tree, they release a volatile substance from their hind gut that is attractive to both males and females. Other individuals exploring in the vicinity are drawn to the penetration gallery, producing the massattack phenomenon well known to forest entomologists (Wood and Bushing, 1963, and contained references). Recently Silverstein et al. (1966) have identified the pheromone as a mixture consisting principally of three alcohols: compound I, (—) -2-methyl-6-methylene-7-octen-4-ol; compound II, (+)-cis-verbenol; and compound III, (+)-2-methyl-6-methylene-2,7-octadien-4-ol. The beetles respond to a combination of compound I with either compound II or compound III. In the genus Dendroctenus it is the female that produces the pheromone (McMullen and Atkins, 1962). Pheromone recruitment also appears to occur in other scolytid genera, such as Scolytus and Xyleborus, and there is an excellent chance that it will be found in the related families Platypodidae and Bostrichidae.
This category includes a variety of specific responses related only by the threatening nature of the stimuli that induced them. Tadpoles of the toad genus Bufo and some fishes scatter from the vicinity of epidermal alarm substances (Schutz, 1956). Ants always react to alarm substances by excited circular or zigzag motion. In the immediate nest vicinity the workers of most species posture aggressively and run toward the stimulus source, but those of other species characteristically flee. A few actually attack the pheromone source, even if it is a sister worker. Aggressiveness increases with the size and health of the colony. As a rule, ant alarm substances are attractants at low concentrations and release aggressive or retreat behavior at higher concentrations (Wilson, 1965a).
Termites lay chemical trails to the area of the threatening stimulus, whether it is an intruding insect or a breach in the nest wall. In the lower termites, trail-laying appears to serve this function exclusively; only in higher evolutionary lines has it come to serve in recruitment to food sources as well (Stuart, 1963). When honeybees sting an enemy, they emit isoamyl acetate, probably in combination with other compounds, from glands near the base of the sting; this substance is in part responsible for the well-known tendency of other bees to sting repeatedly in the same area (Boch et al., 1962; Ghent and Gary, 1962).
Pinenes manufactured in the cephalic glands of nasute soldiers of the termite genus Nasutitermes probably serve both in colony defense and communication of alarm (Moore, 1964). Alarm substances in ants have so far proved to be mostly terpenes with molecular weights between 100 and 200 (Table 4). At least a few of these substances, e.g., citronellal in Acanthomyops claviger, serve as both defense secretions and pheromones. An intimate relation between defensive behavior and chemical communication thus exists in the social insects. Which function came first in the evolution of these insects remains uncertain. Since the employment of exocrine secretions in defense is widespread in both the nonsocial and social insects (Roth and Eisner, 1962), the communicative function is probably secondary.
In the ants the alarm substances are not species-specific or even subfamily-specific in action. Some terpenes are even manufactured by different species, e.g., citral both by species of Atta and by Acanthomyops claviger. (Citral is also produced by the honeybee, where it acts as a simple attractant.) In sixteen species of snails studied by Snyder (1965) the alarm substances are species-specific. These “pheromones” are released when bodies of snails are crushed, a quite different situation from alarm communication in social insects, where the substances are manufactured in special exocrine glands and released as part of the behavioral repertoire of unharmed individuals. The snail substances have not been identified. Quite likely they are normal tissue or blood constituents whose employment as pheromones is secondary.
Bossert and Wilson (1963) have studied the physical features of alarm communication by mandibular gland secretion in the ant Pogonomyrmex badius. The Q/K ratio for the entire contents of the gland reservoirs of one worker is of the order of a thousand, a value intermediate in magnitude between the Q/K values estimated for the fire ant odor trail and the moth sex attractants. The total capacity of the gland reservoirs is about 105 ml, or 1015 to 1016 molecules. In still air the active space, within which the pheromone (or pheromones) acts as a simple attractant, expands to its maximum radius of about six centimeters in thirteen seconds, then shrinks and finally fades out in about thirty-five seconds. Inside this space there is an inner space of higher concentration that induces true alarm behavior. The true alarm space expands to a radius of three centimeters and fades out in about eight seconds. The two concentric spaces are smaller and their fade-out quicker when the pheromone is released in turbulent air or in spurts rather than all at once. The system seems very well designed for the intuitive “goals” of an alarm system in ant colonies; that is, single disturbed workers can spread swift alarm over a short distance during a short interval of time. If the danger is local, the signal quickly fades and the bulk of the colony is undisturbed. If it is persistent, increasing members of workers become involved and the signaling escalates. When the danger stimuli cease, the pheromone signals quickly fade.
THE INCREASE OF INFORMATION
There are several conceivable ways in which chemical systems can be adjusted to enhance the specificity of signals or increase the rate of information transfer. As our knowledge of real cases grows, it is clear that various animal species have actually evolved a rich diversity of these modifications.
Adjustment of Fading Time
The interval between discharge of the pheromone and the total fade-out of its active space can be adjusted, in the course of evolution, by altering the Q/K ratio. The rate of information transfer can be increased by lowering the emission rate Q or raising the threshold concentration K, or both. This adjustment achieves a shorter fade-out time and permits signals to be more sharply pinpointed in time and space. A lower Q/K ratio characterizes both alarm and trail systems.
In the case of ingested pheromones, the duration of the signal can be shortened by enzymatic deactivation of the molecules. When Johnston et al. (1965) traced the metabolism of radioactive trans-9-keto-2-decenoic acid fed to worker honeybees, they found that within seventy-two hours more than 95 per cent of the pheromone had been converted into inactive substances consisting principally of 9-ketodecanoic acid, 9-hydroxydecanoic acid, and 9-hydroxy-2decenoic acid.
Expansion of the Active Space
By increasing Q or decreasing K, the size of the active space can be expanded. This, of course, is what has occurred in the case of sex attractants and the scolytid recruitment substances. If the pheromone is expelled downwind, only a relatively small amount is required, since orientation can be achieved by anemotaxis rather than osmotropotaxis. As a consequence Q can be kept small. The rate of information transfer is kept down, in the sense that signals cannot be turned on or off as rapidly; but the total amount of information eventually transmitted is increased, since a very small target can be pinpointed within a very large space. Also, the use of wind immensely increases the rate of signal transfer over what it would be otherwise, while the substitution of anemotaxis allows Q to be minimized in the first place.
Temporal Patterning of Single Pheromones
Frequency and amplitude modulation is feasible in a steady, moderate wind or over distances of a few centimeters in still air (Bossert, 1966). Although no cases are yet known of this phenomenon in nature, it could easily have been overlooked by past investigators.
Use of Multiple Exocrine Glands
This elementary device has been employed in several animal groups, most conspicuously in the mammals and insects. Kullenberg (1956) found that females of many of the aculeate hymenopteran species release simple attractants from the head and sexual excitants from the abdomen. The highest known development of a multiple pheromone system is in the honey bee and ants. In Table 4 some of the glandular sources and functions of pheromones in these insects are given. Of special interest are glands that have evolved de novo to serve a communicative function. These include Koschevnikov’s and Nassanoffs gland in the honey bee, Pavan’s gland in the dolichoderine ants, and the sternal gland of termites. The exploration of the chemical “codes” of the various groups of social insects is still in the earliest stage.
Medleys of Pheromones
Different substances with different meanings can be generated by the same gland. A minimum of thirty-two compounds have been detected in the heads of honey bee queens, including methyl 9-ketodecanoate, methyl 9-keto-2-decenoate, nonoic acid, decenoic acid, 2decenoic acid, 9-ketodecanoic acid, 9-hydroxy-2-decenoic acid, 10hydroxy-2-decenoic acid, and 9-keto-2-decenoic acid, and others (Callow et al., 1964). Most or all are present in the mandibular gland secretion. The biological significance of most of these substances is still unknown. Some are undoubtedly precursors to pheromones, but at least two are known pheromones with contrasting effects: the 9-ketodecenoic acid is the inhibitory pheromone already mentioned, and the 9-hydroxydecenoic acid causes clustering and stabilization of worker swarms (Butler et al., 1964). A rich mixture of chemicals is also found in the castoreum of the beaver. About forty-five substances have been identified, including a surprisingly diverse array of alcohols, phenols, ketones, organic acids, and esters, as well as salicylaldehyde and castoramine (C15H23O2N) (Lederer, 1950). Although a behavioral function has not yet been demonstrated, it is likely that a conscious testing of the idea will reveal some of the substances to be pheromones.
Change of Meaning Through Change of Context
trans-9-keto-2-decenoic acid serves as a caste-inhibitory pheromone inside the honey bee nest and as the primary sex attractant during the nuptial flight. The Dufours gland secretion of the fire ant Solenopsis saevissima is an attractant that is effective on members of all castes during most of their adult lives. Under different circumstances it serves variously to recruit workers to new food sources, to organize colony emigration, and—in conjunction with a volatile cephalic secretion—to cause oriented alarm behavior.
Variation in Concentration and Duration
Workers of the ant Pogonomyrmex badius react to low concentrations of mandibular gland secretion by simple positive Chemotaxis and to higher concentrations by typical aggressive alarm behavior. When exposed to high concentrations for more than a minute or two, many individuals switch from alarm to digging behavior.
New Meanings from Combinations
There are a few examples of pheromones acquiring additional or even different meanings when presented in combination. When released near fire ant workers, cephalic and Dufour’s gland secretions cause alarm behavior and attraction, respectively; when expelled simultaneously by a highly excited worker, they cause oriented alarm behavior. Honey bee workers confined closely with queens for hours acquire scents from her which, evidently in combination with their own worker-recognition scent, cause them to be attacked by nest mates (Morse and Gary, 1961).
CHEMICAL COMMUNICATION AMONG SPECIES
So far in this chapter we have considered only communication within species. The word “pheromone” is defined as a chemical employed as a signal among the members of the same species, a usage that reflects our past confidence that intraspecific communication embraces a reasonably discrete set of phenomena. Yet it is becoming clear that many if not all of the same phenomena are displayed in various instances of communication among species. Symbiosis, host selection, parasitism, predation, and interspecific competition are all relationships which, under certain circumstances, require a large amount of chemical information. It is true that most of these relationships differ from pheromone communication in that ordinarily only the receiving species alters its behavior to take advantage of the presence of the chemical. The “signaling” species—the host species, to be more precise—is most often under pressure to reduce its signal if it can rather than improve it. Perhaps only in cases of aposematic (threatening) behavior and mutualistic symbiosis is it to the advantage of the sender to emit a sharp, unmistakable taste or odor.
The literature touching on chemical communication among species is already enormous, and it would be wholly out of place to attempt to review it in a book on animal communication. I will limit treatment of it here to a brief description of three remarkable examples.
Trail Sharing in Azteca and Camponotus Ants
Camponotus beebei is a relatively scarce species found in close association with the dominant arboreal species Azteca chartifex in Trinidad. During the day, the Camponotus workers follow the Azteca odor trails down the tree trunks to the foraging grounds, avoiding the smaller, more numerous Azteca workers on the trails with swift sidestepping movements. On occasion the Camponotus workers lay odor trails of their own on top of the Azteca trails; the Camponotus but not the Azteca are able to follow these superimposed trails. Thus the Camponotus exploit the Azteca communication system without sharing their own (Wilson, 1965b).
Predator Recognition in Mosquito Fish
When a strange fish enters the vicinity of North American mosquito fish (Gambusia partruelis), the latter swim cautiously toward it while orienting visually to avoid the “attack cone” or conic space in front of the head of the intruder. If the odor of a pickerel (Esox sp.) is detected, the Gambusia undergo rapidly the following changes: the eyes darken; conspicuous suborbital pigment bands appear; the fins are held more erect; the bodies become straighter and the postures more rigid; the caudal sweep of the tail fins is diminished; and the fish tend to move upward and remain near the surface film. While they remain in this alerted state, the Gambusia are easily induced to give the jump response, an erratic skipping along and above the water surface; and they are less likely to swim downward when a shadow passes over. Of all the fish species that commonly occur with Gambusia, only the members of Esox produce an odor that induces this complex set of responses. Esox are among the principal predators of Gambusia, and the responses are ideally suited to the avoidance of the larger fish (George, 1960).
Cross-Species Embryonic Induction in Rotifers
Among the rotifers, species of Asplancha are common predators of Brachionus calyciflorus. Asplancha releases a water-soluble, nodialyzable factor that causes eggs of B. calyciflorus to develop into individuals with a pair of long, movable spines possessed by neither their mothers nor unexposed control individuals of the same generation. The spines protect the B. calyciflorus individuals from Asplancha predation (Gilbert, 1966).
These striking instances of chemical communication among species illustrate some of the richness and diversity of chemically mediated responses that have been evolved.
Barnhart, C. S., and L. E. Chadwick, 1953. A fly factor in attractant studies. Science, 117:104-105.
Barth, R., 1937. Herkunft, Wirkung und Eigenschaften des weiblichen Sexualduftstoffes einiger Pyraliden. Zool. Jahrb. Allgem. Zool. Physiol., 58:297-329.
Barth, R. H., 1962. Comparative and experimental studies on mating behavior in cockroaches. Ph.D. thesis, Harvard University, Cambridge, Mass.
Berger, R. S. 1966. Isolation, identification, and synthesis of the sex attractant of the cabbage looper, Trichoplusia ni. Ann. Entomol. Soc. Am., 59:767-771.
Boch, R., and D. A. Shearer, 1963. Production of geraniol by honey bees of various ages. J. Inst. Physiol., 9:431-434.
——, D. A. Shearer, and B. C. Stone, 1962. Identification of iso-amyl acetate as an active component in the sting pheromone of the honey bee. Nature, 195:1018-1020.
Boeckh, J., K.-E. Kaissling, and D. Schneider, 1965. Insect olfactory receptors. Cold Spring Harbor Symp. Quant. Biol., 30:263-280.
Bornemissza, G. F., 1964. Sex attractant of male scorpion flies. Nature, 203:786-787.
——, 1966. Specificity of male sex attractants in some Australian scorpion flies. Nature, 209:732-733.
Bossert, W. H., 1967. Temporal patterning in olfactory communication. J. Theoret. Biol, (in press).
——, and E. O. Wilson, 1963. The analysis of olfactory communication among animals. J. Theoret. Biol., 5:443-469.
Bretschneider, L. H., and J. J. D. DeWit, 1947. Sexual Endocrinology of Non-Mammalian Vertebrates. Amsterdam, Elsevier.
Butenandt, A., and E. Hecker, 1961. Synthese des Bombykols, des Sexual-Lockstoffes des Seidenspinners, und seiner geometrischen Isomeren. Angew. Chem., 73:349-353.
Butler, C. G., 1964. Pheromones in sexual processes in insects. Symp. Roy. Entomol. Soc. London, 2:66-77.
——, R. K. Callow, and N. C. Johnston, 1961. The isolation and synthesis of queen substance, 9-oxodec-/rans-2-enoic acid, a honeybee pheromone. Proc. Roy. Soc. (London), B 155:417-432.
——, R. K. Callow, and J. R. Chapman, 1964. 9-Hydroxydec-trans-2-enoic acid, a pheromone stabilizing honeybee swarms. Nature, 201:733.
Callow, R. K., J. R. Chapman, and P. N. Paton, 1964. Pheromones of the honeybee: Chemical studies of the mandibular gland secretion of the queen. J. Apicult. Res., 3:77-89.
Casida, J. E., H. C. Coppel, and T. Watanabe, 1963. Purification and potency of the sex attractant from the introduced pine sawfly, Diprion similis. J. Econ. Entomol., 56:18-24.
Crane, J., 1949. Comparative biology of salticid spiders at Rancho Grande, Venezuela. Part IV. An analysis of display. Zoologica, 34:159-214.
Day, A. C., and M. C. Whiting, 1964. The structure of the sex-attractant of the American cockroach, ?roc. Chem. Soc., 1964:368.
Eisner, T., and F. C. Kafatos, 1962. Defense mechanisms of arthropods. X. A pheromone promoting aggregation in an aposematic distasteful insect. Psyche, 69:53-61.
Engelhardt, M., H. Rapoport, and A. Sokoloff, 1965. Odorous secretion of normal and mutant Tribolium confusum. Science, 150:632-633.
Gary, N. E., 1962. Chemical mating attractants in the queen honey bee. Science, 136:773-774.
Gaston, L. K., and H. H. Shorey, 1964. Sex pheromones of noctuid moths. IV. An apparatus for bioassaying the pheromones of six species. Ann. Entomol. Soc. Am., 57:779-780.
George, C. J., 1960. Behavioral interaction of the pickerel (Esox niger Le Sueur and Esox americanus Le Sueur) and the mosquito-fish [Gambusia patruelis (Baird and Girard)]. Ph.D. thesis. Harvard University, Cambridge, Mass.
Ghent, A. W., 1960. A study of the group-feeding behaviour of larvae of the jack pine sawfly, Neodiprion pratti banksianae Roh. Behaviour, 16:110-148.
——, and N. E. Gary, 1962. A chemical alarm releaser in honey bee stings (Apis mellifera L.). Psyche, 69:1-6.
Gilbert, J. J., 1966. Rotifer ecology and embryological induction. Science, 151: 1234-1237.
Haas, A., 1952. Die Mandibeldrüse als Duftorgan bei einigen Hymenopteren. Naturwiss., 39:484.
Haidane, J. B. S., 1955. Animal communication and the origin of human language. Sci. Progr., 43(171):385-401.
Happ, G. M., and J. Meinwald, 1965. Biosynthesis of arthropod secretions. I. Monoterpene synthesis in an ant (Acanthomyops claviger). J. Am. Chem. Soc., 87:2507-2508.
Jacobson, M., 1965. Insect Sex Attractants. New York, Wiley (Interscience).
——, M. Beroza, and W. A. Jones, 1960. Isolation, identification and synthesis of the sex attractant of gypsy moth. Science, 132:1011-1012.
Johnston, N. C., J. H. Law, and N. Weaver, 1965. Metabolism of 9-ketodec2-enoic acid by worker honeybees (Apis mellifera L.). Biochemistry, 4:1615— 1621.
Jones, W. A., M. Jacobson, and D. F. Martin. 1966. Sex attractant of the pink bollworm moth: isolation, identification, and synthesis. Science, 152:1516-1517.
Karlson, P., and A. Butenandt, 1959. Pheromones (ectohormones) in insects. Ann. Rev. Entomol., 4:39-58.
Kullenberg, B., 1956. Field experiments with chemical sexual attractants on aculeate Hymenoptera males. Zool. Bidrag. Uppsala, 31:253-354.
Lange, R., 1960. Uber die Futterweitergabe zwischen Angehörigen verscheidener Waldameisen. Z. Tierpsychol., 17:389-401.
Law, J. H., E. O. Wilson, and J. A. McCloskey, 1965. Biochemical polymorphism in ants. Science, 149:544-546.
Lederer, E., 1950. Odeurs et parfums des animaux. Fortschr. Chem. Org. Natur-Stoffe, 6:87-153.
Lindauer, M., and H. Martin, 1963. Uber die Orientierung der Biene im Duftfeld. Naturwiss., 50:509-514.
Loconti, J. D., and L. M. Roth, 1953. Composition of the odorous secretion of Tribolium castaneum. Ann. Entomol. Soc. Am., 46:281-289.
McMullen, L. H., and M. D. Atkins, 1962. On the flight and host selection of the Douglas-fir beetle, Dendroctonus pseudotsugae Hopk. (Coleoptera: Scolytidae). Can. Entomol, 94:1309-1325.
Meinwald, J., Y. C. Meinwald, J. W. Wheeler, T. Eisner, and L. P. Brower, 1966. Major components in the exocrine secretion of a male butterfly (Lycorea). Science, 151:583-585.
Moore, B. P., 1964. Volatile terpenes from Nasutitermes soldiers (Isoptera, Termitidae). J. Inst. Physiol., 10:371-375.
Morse, R. A., and N. E. Gary, 1961. Colony response to worker bees confined with queens (Apis mellifera L.). Bee World, 42:197-199.
Mykytowycz, R., 1965. Further observations on the territorial function and histology of the subcutaneous (chin) glands in the rabbit, Oryctolagus cuniculus (L.). Animal Behaviour, 13:400-412.
Naylor, A. F., 1959. An experimental analysis of dispersal in the flour beetle, Tribolium confusum. Ecology, 40:453-465.
Nixon, H. L., and C. R. Ribbands, 1952. Food transmission within the honeybee community. Proc. Roy. Soc. (London), B 140:43-50.
Parkes, A. S., and H. M. Bruce, 1961. Olfactory stimuli in mammalian reproduction. Science, 134:1049-1054.
Roth, L. M., and T. Eisner, 1962. Chemical defenses of arthropods. Ann. Rev. Entomol., 7:107-136.
Ryan, E. P., 1966. Pheromone: evidence in a decapod crustacean. Science, 151: 340-341.
Salt, G., 1936. Experimental studies in insect parasitism. IV. The effect of superparasitism on populations of Trichogramma evanescens. J. Exptl. Biol., 13: 363-375.
Schneider, D., 1962. Electrophysiological investigation on the olfactory specificity of sexual attracting substances in different species of moths. J. Inst. Physiol., 8:15-30.
Schultze-Westrum, T., 1965. Innerartliche Verständigung durch Düfte beim Gleitbeutler Petaurus breviceps papuanus Thomas (Marsupialia, Phalangendae). Z. Vergleich. Physiol, 50:151-220.
Schutz, F., 1956. Vergleichende Untersuchungen über die Schreckreaktion bei Fischen und deren Verbreitung. Z. Vergleich. Physiol., 38:84-135.
Schwinck, I., 1955. Freilandversuche zur Frage der Artspezifität des weiblichen Sexualduftstoffes der Nonne (Lymantria monacha L.) und des Schwammspinners (Lymantria dispar L.). Z. Angew. Entomol., 37:349-357.
Shorey, H. H., L. K. Gaston, and J. S. Roberts, 1965. Sex pheromones of noctuid moths. VI. Absence of behavioral specificity for the female sex pheromones of Trichoplusia ni versus Autographa californica, and Heliothus zea versus H. virescens (Lepidoptera: Noctuidae). Ann. Entomol. Soc. Am., 58:600-603.
——, and L. K. Gaston, 1965. Sex pheromones of noctuid moths. VII. Quantitative aspects of the production and release of pheromone by females of Trichoplusia ni (Lepidoptera: Noctuidae). Ann. Entomol. Soc. Am., 58: 604-611.
Silverstein, R. M., J. O. Rodin, and D. L. Wood, 1966. Sex attractants in frass produced by male Ips confusus in ponderosa pine. Science, 154:509-510.
Snyder, N., 1965. An alarm reaction of aquatic gastropods to intraspecific extract. Ph.D. thesis, Cornell Univ., Ithaca, N.Y.
Soo Hoo, C. F., and R. J. Roberts, 1965. Sex attraction in Rhopaea (Coleoptera: Scarabaeidae). Nature, 205:724-725.
Stein, G., 1963. Über den Sexuallockstoff von Hummelmännchen. Naturwiss., 50:305.
Steinbrecht, R. A., 1964. Die Abhängigkeit der Lockwirkung des Sexualduftorgans weiblicher Seidenspinner (Bombyx mori) von Alter und Kopulation. Z. Vergleich. Physiol., 48:341-356.
Stuart, A. M., 1963. Origin of the trail in the termites Nasutitermes corniger (Motschulsky) and Zootermopsis nevadensis (Hagen), Isoptera. Physiol. Zool., 36:69-84.
Verron, H., 1963. Role des stimuli chimiques dans l’attraction sociale chez Calotermes flavicollis (Fabr.). Chap. III. La perception olfactive chez Calotermes flavicollis. Insectes Sociaux, 10:297-336.
Walsh, C. T., J. H. Law, and E. O. Wilson, 1965. Purification of the fire ant trail substance. Nature, 207:320-321.
Wilson, E. O., 1962. Chemical communication among workers of the fire ant Solenopsis saevissima (Fr. Smith). Animal Behaviour, 10:134-164.
——, 1965a. Chemical communication in the social insects. Science, 149: 1064-1071.
——, 1965b. Trail sharing in ants. Psyche, 72:2-7.
——, and W. H. Bossert, 1963. Chemical communication among animals. Recent Progr. Hormone Res., 19:673-716.
Wood, D. L., and R. W. Bushing, 1963. The olfactory response of Ips confusus (LeConte) (Coleoptera: Scolytidae) to the secondary attraction in the laboratory. Can. Entomol., 95:1066-1078.
Wright, R. H., 1966a. Why is an odour? Nature, 209:551-554.
——, 1966b. Odour and molecular vibration. Nature, 209:571-573.