Ethology BIBLIOGRAPHY Ethology has existed as a concept since 1762 when it was defined in France as the study of animal

behavior. In this sense it carries the same meaning as the Greek word ethos, from which the modern term ethology is derived. However, a separate meaning of the word ethology, related to the term ethics, has been used in the Anglo-Saxon literature to define the science of character. The founder of modern ethology is Konrad Z. Lorenz, physician, zoologist, and comparative anatomist. By systematic application of biological research methods to the analysis of animal behavior, he provided the initial impetus in the 1930s. The first modern ethology textbook, The Study of Instinct, was written by Nikolaas Tinbergen in 1951, and E. H. Hess (1962) and Eibl-Eibesfeldt (1966) recently produced summaries of the modern concepts of behavior. The observations of a number of pioneers, including Spalding (1873), Darwin (1872), Whitman (1898), Altum (1868), Heinroth (1911), and Craig (1918), awakened scientific interest in animal behavior, and ethology came to be considered an independent branch of zoology around 1910. As with every young science, ethology inevitably suffers, on the one hand, from the incorporation of concepts whose meaning has oscillated or has already become too specialized (such as instinct) and, on the other hand, from the application of provisional concepts, which may alter in meaning with advances in knowledge, to contemporary working hypotheses. The term ethology is now attached to the scientific investigation of the behavior of animals and of some aspects of human behavior. Pronouncements about inaccessible psychic phenomena are avoided; the term animal psychology is still occasionally used but on purely historical grounds. Ethology is concerned with the investigation of animals, whether these be single cellseither as individual protozoans or as parts of metazoans or more complex animal structures, that is, individuals, groups, or so-called animal colonies (e.g., ants, bees, and termites). The behavior of an animal is equated with changes brought about by effectors (e.g., movements, sounds, scent production, color changes). Such effector responses are temporal events. For this reason only effector responses which repeatedly and identifiably occur are open to scientific analysis; they are then termed fixed action patterns. It is important to note that these temporal events can be recorded by tape and film, except in the case of chemical or tactile signals. The locomotion of an animal can be subdivided into the movement of the extremities, of antagonistic muscle groups, of single muscles, and ultimately of the muscle fibers. The smallest identifiable effector components, occurring either singly or in combination with other components, are chosen as the units of ethological study. The aim of ethology is to explain both phylogenetically and physiologically the functional relationships of all factors involved in behavior. This is evident in the modern definition of instinct suggested by Tinbergen: a hierarchically organized nervous mechanism which is susceptible to certain priming, releasing and directing impulses of internal as well as of external origin, and which responds to these impulses by coordinated movements that contribute to the maintenance of the individual and the species (1951, p. 112). The touchstone for ethological hypotheses is the reliable prediction of the behavior of a living system in any given situation. Ethologists are zoologists; they are thus interested in the biology of a species, and their prime interest is behavior as it occurs under natural conditions. The ethologist always begins by compiling an action catalogue, or ethogram of the species in question, that is, as complete a description as possible of the behavior throughout the animals life cycle. This simply describes what the animal does, not why it does it. The various behavior patterns are then classified and compared with those of other species, especially with closely related species. It is important that the animals should be observed in their natural habitats or in surroundings which closely resemble them. Additional observations in captivity are often necessary. A very useful expedient, first known to have been practiced by Baron Ferdinand Adam von Pernau in 1702, is to rear the animals to be both tame and unconfined.

Learning, maturation, and genetics. Although learning is considered to be very important in animal behavior, the first concern of the ethologist is with behavior patterns typically performed by all animals of a species, because it is necessary for him to know the basic predetermined responses before proceeding to study changes brought about by learning. This is important, since not every change of form or effectiveness of a given behavior pattern occurring during the life of an individual involves learning in the form of acquisition by experience. As early as 1760 a professor in Hamburg, Hermann Samuel Reimarus, discovered the phenomenon of maturation of instincts and pointed out the difference between innate and acquired skills. The innate skills, for example, the collection of food or the performance and understanding of the dance language in bees, are present from the time of birth or of hatching from the egg or pupa. Without involving a definition of learning, the problem can be formulated as follows: the majority of behavior patterns in most animals are adapted (adjusted) to special situations in their respective environments. Since this fact cannot be explained as a chance phenomenon and since it is not a self-evident phenomenon, an explanation must be provided. In order to behave adaptively, the animal must have at its disposal information about the environment. This information can be stored either in the chromosomes or in memory; that is, it can be either innate or acquired. In complex behavior patterns, there is often an interaction between innate and acquired elements. However, although we know of perceptual and motor skills in which learning plays no part, it is impossible to postulate a completely learned element of behavior that is not based on genetically determined and, therefore, delimited capabilities. Further, no one has so far been able to demonstrate the infinite modifiability of any arbitrarily chosen, innately determined element of behavior or the possibility that learning could be the function of a nonorganized aggregate of neural elements. In learning, the fact that the organism selects good and not bad behavioral responses or stimuli logically implies a built-in mechanism which is able to direct learning toward survival value. A particularly good method for distinguishing between the learned and innate elements of behavior is contained in the deprivation experiment: the animal, usually isolated from members of its own species, is deprived of certain experiences and later tested in the situation to which the behavior pattern in question is normally adapted. As a control a normal animal must be tested in the same situation. (This is one of several safety precautions which are necessary in the evaluation of the deprivation experiment.) The majority of behavior patterns do not follow the all-or-none law but can occur at varying intensities. The lowest intensities, where it is just possible to recognize which pattern has been activated, are referred to as intention movements. The intensity with which a behavior pattern is performed depends upon both internal and external factors. The appearance of a particular fixed action pattern in animals isolated from their own species is clear evidence of genetic fixity. It is a constant characteristic of the species concerned and is based upon a specific central nervous mechanism that is inherited just as are morphological and physiological characteristics. (This had already been stated by the English naturalist Spalding in 1873.) A particularly good example is provided by many bird songs which develop into the species-specific pattern even in completely isolated animals. Research into the genetic basis of behavior patterns is developing as an important part of ethology. For example, crossing two duck species which differ in their courtship behavior can give rise to hybrids exhibiting courtship motor sequences not evident in any known species of duck or sometimes to hybrids possessing behavior patterns absent in both parent species but present in some presumed ancestral type (Wall 1963). However, it is still not clear what changes in the complex physiological basis of such behavior patterns are responsible for these differences. Dilger investigated the carrying of nest material in F1hybrids of Agapornis roseicollis parr at (male) x A. per sonata fischeri (female). Both parent species cut strips of paper or leaves and carry them to the nest. Females of the first species carry the strips in their bills, while females of the second species tuck the strips under special feathers on the lower back. The hybrids attempted to perform this latter pattern but failed for various reasons. For example, some were unable to let go of the strip of paper and tried to carry the strips in the bill and under the feathers at the same time. Within two years the behavior of the hybrids improved through learning, but they continued to perform ineffective tucking movements (Dilger 1962).

The genetic fixity of elements of behavior and the fact that they are nearly always to be found in more than one species prove their taxonomic value. In fact they are often characteristic of genera, families, or even higher taxonomic categories. For this reason it is possible to employ behavior patterns in the investigation of the relationships between animals. Indeed, Whitman (1898) and Heinroth (1911) investigated the behavior of doves and ducks respectively in the hope of finding characteristics useful for a more systematic analysis of their interrelatedness. In some grasshopper and toad species, speciesspecific calls or songs are used for species recognition; thus they represent barriers to interspecific reproduction. On the other hand, it is possible to reconstruct the phylogeny of behavior patterns on the basis of variations in the form of the same basic pattern between closely related species, as was pointed out by Darwin in The Expression of the Emotions in Man and Animals (1872). Exactly the same method is used in comparative anatomy and morphology. Although no behavioral fossils exist, more transitional forms exist between different behavioral types than is the case with morphological characteristics; behavioral characteristics occur repeatedly and at different intensities, while a leg is formed only once. The individual elements of various behavior patterns are, however, more open to formation of novel combinations (e.g., in contrast to the bones of the skull in vertebrates). For this reason, the phylogeny of behavior patterns must be based on the simplest possible elements. We know that no behavioral characteristic is dependent upon only one gene, that each hereditary component affects several characteristics, and that there are not two separate sets of hereditary material governing body construction and behavioral features. The interaction of a behavior pattern with its effector organ is thus just as labile as the coadaptation of several functionally correlated organs. Evolution and selection. Reimarus pointed out that in many instances behavior patterns adapted to the use of certain organs are performed, at times, even before these organs are developed. Apart from differences in speed of development of behavior patterns and their effector organs, it is also possible that one survives when the other is lost; a cerambycid beetle will continue to preen its antennas after they have been removed by dissection, mutant fruit flies (Drosophila) with no wings still perform the wingpreening movements typical of the species, stump-tailed monkeys, when they run along a branch, still show the balancing movements once effective in their long-tailed ancestors. Such historical carry-overs can also be observed in behavior patterns originally adapted to certain environmental conditions which have since changed. Ground-breeding birds regularly use the beak to perform specific behavior patterns for rolling the eggs back into the nest if they are found outside. Treebreeding birds do not show this, since the eggs that fall disappear. However, Poulsen (1953) was able to show that some birds which have recently evolved from ground-living stock to a tree-living habit still exhibit egg-rolling patterns, while some recently evolved ground-breeding birds lack this pattern. There are other examples which show that fixed action patterns can be extremely conservative in the evolution of a species. On the other hand, closely related species occupying the same area exhibit rapid phylogenetic changes in the sexual behavior patterns serving for sexual recognition. In fact, in some such species greater differences in species recognition signals are seen where two species occur together than where either species occurs in isolation. This phenomenon has been called character displacement. Intraspecific signals usually undergo selection for better recognition (to avoid misunderstandings) and tend to become more and more conspicuous and outlandish. The behaviors involved are performed more conspicuously and are emphasized by morphological characteristics of color, form, or odor. In addition such behaviors are often rhythmically repeated. This fixed patterning often ceases to show different degrees of intensity. The level of motivation is no longer expressed in the intensity of the behavior but in how often it is repeated at one and the same fixed intensity (Morris 1957), much as the urgency of a telephone call is indicated by how often the bell rings and not how loud it rings. Finally such a recently formed fixed action pattern may become motivationally autonomous of the situation in which it was originally aroused by a process that Tinbergen has called emancipation. These and other changes in signal behavior patterns, leading to improved communication between signal sender and signal receiver, are referred to as ritualization. We still know very little regarding the physiological and neuroanatomical basis of both nonritualized and ritualized behavior patterns. [see COMMUNICATION, ANIMAL.]

It is commonplace to say that no animal performs the behavior patterns in its repertoire in random order. An animal responds to signals according to set principles. It is the task of the behavioral physiologist to analyze this phenomenon. This task involves large-scale studies of sensory physiology, since the animal receives the stimuli with its sense organs; of hormone physiology, since hormones can decide whether the sight of the female elicits courtship by the male, and so on; and of the physiology of the central nervous system, which is responsible for the analysis of the stimuli and for the coordination of the requisite behavior patterns. Releasing mechanisms. The carnivorous water beetle Dysticus marginalis does not react to the sight of prey (e.g., a tadpole in a glass tube), although it has perfectly developed eyes, but it does react to the chemical stimuli emanating from the prey. If some prey-extract solution is added to the water, the beetle will clasp even inanimate objects immersed in the water. A male robin will attack a bundle of red feathers but not a perfect dummy of a male lacking the characteristic red breast. Such examples show that animals respond with quite specific reactions to quite specific stimuli among the many perceived from the environment. These relevant stimuli are called sign stimuli, or releasers. Sign stimuli act upon specific functional units of the central nervous system, the so-called releasing mechanisms. The specific properties of these units may likewise be either genetically determined, in which case they are termed innate releasing mechanisms (IRM), or partially determined by learning. The releasing mechanism filters out the sign stimuli and thereby triggers off specific behavior patterns. Some behavior patterns can be elicited by more than one stimulus (e.g., an odor or a vibration). The vigor of the reaction generally depends upon the strength of the stimulus, and heterogeneous stimuli may summate (the same intensity of a reaction may be shown toward a strong odor or a strong vibration or a weak odor together with a weak vibration). Sometimes it is possible to present the animal with an abnormally strong stimulus and obtain a response stronger than that released by the naturally occurring stimulus; Magnus (1958) has shown that the males of the silver-washed fritillary butterfly react with courtship toward the orange and black color pattern of the females fluttering wings. By placing orange and dark stripes on a cylinder and rotating it, he proved that more rapid color-dark alternation than the rate characteristic of the female was more effective in eliciting the males reaction. The greater the speed of rotation of the cylinder, the greater were the courtship responses, right up to the physiologically demonstrated flickerfusion frequency for the species concerned. [see SEXUAL BEHAVIOR, article on ANIMAL SEXUAL BEHAVIOR.] This susceptibility of animals tosupernormal releasers provides us with an insight into the reason for the development of bizarre morphological signal structures such as the feathers of the peacock. Some parasitic birds even capitalize on this phenomenon when their young are larger and more babyish than, and therefore preferred to, the hosts own young. Motivation and drives. It has been observed that one individual will sometimes respond to a weak stimulus with a strong response, while at other times respond only at the same intensity or not at all to a much stronger releasing stimulus. It is therefore necessary to measure independently the specific readiness of the animal to react, apart from the strength of the stimulus. The strength of a reaction often decreases sharply with repeated equivalent stimulation, as is the case with escape attempts in aquarium fish in response to tapping on the glass pane or with gaping in young birds when the nest is lightly shaken. The readiness of an animal to perform certain patterns exhibits extensive and independent variation; an animal which is not prepared to eat may nevertheless exhibit readiness to flee. The readiness to perform a certain pattern is referred to as the motivation. Motivation (e.g., in hunger) often increases with the time interval from the last elicita-tion of the type of behavior concerned (damming effect, an effect which is related to the corresponding stimulus threshold). In the extreme case the action pattern can occur without any evident external elicitationas vacuum activity. However, an animal with high motivation to perform specific behavior patterns (where the drive is under restraint) usually performs certain behavior patterns suitable for attainment of a stimulus situation appropriate to the motivated patterns. In simple terms, the animal searches. Craig (1918) observed the occurrence of restlessness, varied movements, and searching behavior as symptoms of a physiological state of appetite for specific stimuli and labeled such behavior appetitive behavior, as distinct fromconsummatory behavior,

which lowers the degree of motivation when performed and leads to a state of satisfaction. It is important to note that the animal does not attempt to achieve the biological effect associated with the consummatory act but merely the performance of the consummatory behavior itself. The state of satisfaction can also be achieved by abreaction in response to models. In the simplest case, appetitive behavior consists of undirected locomotion, but many animals (especially higher-developed forms) learn from experience and modify the appetitive behavior, so that it more rapidly leads to success. They learn when, where, and how they can attain the releasing situation. Briefly, appetitive behavior is typically variable (plastic), whereas the consummatory act is relatively fixed (stereotyped). Motivational analysis attempts to demonstrate how many behavior patterns are dependent upon the same motivational source and how many partially or completely independent motivational centers are present in a given animal species. It is taken as axiomatic that there are fewer independent sources of motivation than observably distinct behavior patterns and that behavior patterns which regularly occur within short intervals from one another are thus commonly motivatedthe motivational state of the animal oscillates more slowly than the alternation of behavior patterns. Further, it is known that behavior patterns exist which are characteristic for specific conflict situations; in the conflict between attack and flight these are represented by threat behavior. Such patterns are certainly motivated from different sources, which may vary independently from one another. It is not known from the outset how many of the behavior patterns observed have mixed motivation, but for the purposes of analysis it is assumed that it is the minimum possible. Wiepkema, a Dutch ethologist, carried out the following model experiment (1961) with the European bitterling: First he recorded the occurrence of the behavior patterns which he had identified for this species (ramming the flank of a conspecific with the head, scouring of the substrate, tail-beating, swimming before the female, etc.) over a long period of time. In this process typical locomotory sequences are found to occur regularly, while some behavior patterns are seen to be mutually exclusive within a given time interval. Wiepkema computed the minimum number of independent variables (i.e., motives) necessary to account for the observed distribution of action patterns. Mixed motivation was taken into account, but it was assumed that one given motive was predominant in each case. For the reproductive period of the bitterling it was found that three independent motivational sources are necessary and that each source governs a group of motor patterns which are totally or predominantly dependent upon the source concerned. These groups are comprised of (1) behavior patterns directed at the rival, objectively termed fight, (2) behavior patterns directed away from the rival, termed flight, and (3) the patterns carried out in combination with the female in association with spawning. Accordingly it is possible to refer to the predominant motivation in each case as (1) fight drive, (2) flight drive, and (3) sex drive. Some behavior patterns lie between these groups, however, and are thus more or less equally dependent upon more than one motive. For example, spreading of the fins combined with an undulating movement of the entire body, which we refer to as threat, is motivated by both fight and flight drives; a specific courtship pattern is motivated by both fight and sex drives. Using factor analysis it is even possible to rank the action patterns according to a scale of ratios between different drives. [see DRIVES; MOTIVATION.] Various phenomena occur in conflict situations; the animal may combine two behavior patterns (e.g., warding off a rival and eating), it may oscillate between the intention movements of different action patterns (e.g., oscillation between motions toward attack and flight without actually attacking or fleeing), it may exhibit abreaction of an inhibited behavior pattern by transferring the direction to a neutral object (e.g., gulls which do not dare to attack a stronger rival may tear and pluck at tufts of grass), or it may exhibit a behavior pattern which does not belong to either of the motivational sources directly involved in the conflict. This last pattern is referred to as displacement activity. For example, domestic cocks will start to eat when they are involved in a conflict between attack and flight, while avocets will assume a sleeping posture. The physiological foundations of displacement activities have been investigated only in a few cases and appear to vary from case to case. Some behavior patterns may be dependent upon the same releasing stimuli as well as upon the same motivational sources. [see CONFLICT, article on PSYCHOLOGICAL ASPECTS.]

Sequential and hierarchical organization. The fact that some elements of behavior can give rise to conflict at corresponding integrational levels, while others are mutually exclusive, indicates that groups of elements are governed by superior systems which can similarly show mutual interference, promotion, inhibition, or exclusion. In this way, we arrive at the concept of a hierarchical system of dominant and subordinate drives. The same concept emerges from the comparison of releasing situations and appetitive behavior. A hungry squirrel (1) climbs (2) trees (3) looking for cones; when motivated to build a nest, it (1) climbs (2) trees (3) looking for twigs. Thus different motivations may employ the same lower motor and orientation components. The latter are called taxes, a taxis being defined as orientation of the whole body or parts of it with respect to the source of stimulation. Further, the distinction between appetitive behavior and consummatory behavior is a relative one. Normally, certain appetitive behavior leads to a stimulus situation which initiates another, more specific, appetitive behavior. This fact has been carefully worked out by Baerends (1941) and Tinbergen (1951). The three-spined stickleback is brought into reproductive motivation by the gradual increase in day length in spring and begins migration inland into shallow freshwater habitats. This factor, together with the rise in water temperature and the visual stimulation of heavily vegetated sites, is a releasing mechanism for the establishment of a suitable territory by the males. A territory is necessary for the male to acquire its characteristic red belly. Only then does it begin to react to particular stimuli which previously had no effect. The male will build a nest with suitable material, fight against rival males (where the releasing stimulus is the red belly of the male intruding into his territory), and court passing females, which present their silvery, swollen, egg-filled bellies to the male in a characteristic manner. Thus, the stimuli emanating from a territory will activate the fighting, building, and mating drives, which must then be elicited by special releasers. Fighting itself consists of a number of behavior patterns (chasing, threatening, tail-beating, biting), each dependent upon still further, highly specific stimuli emanating from the intruders behavior. The behavioral sequences of male and female form an alternating chain of reactions, each action of one partner releasing the following appropriate reaction of the other partner until the female spawns and the male fertilizes the eggs. The act of fertilization initiates brood care in the male; he now fans fresh water onto the eggs and continues to drive off rivals but does not exhibit further courtship until the young hatch. It is thus clear that there are chains of behavioral tendencies connected at higher and lower levels of integration and that these different levels are organized into a hierarchical system. The advantage of a hierarchy, as opposed to a stereotyped series of single fixed actions, lies in its adaptability to unpredictable sequences of events. Neurophysiological aspects. It seems evident that some structural organization must exist within the central nervous system, paralleling the observed organization of behavioral responsesin particular the hierarchical organization. Neurophysiological investigation of fixed action patterns has therefore become an important branch of ethological research. Extending the earlier experiments of W. R. Hess (1949), Hoist and Saint Paul (1959) demonstrated the existence of structural hierarchical organization of the mechanisms underlying behavior by electrically stimulating specific areas of the brain in chickens. Wellcoordinated, complete sequences of movements identical with those observed in normal behavior were elicited. All these sequences of behavior were composed of single actions, each of which could be obtained in isolation by stimulation of specific brain areas. Holst and Saint Paul combined brain stimulation with the normal releasing stimuli, electrically changed the mood (motivational state) of the animal and studied artificial conflict between drives by producing interaction of different behavior patterns with simultaneous elicitation. Neurological research has substantiated the conclusion, derived from ethological field studies, that the coordination of many locomotive patterns arises from impulses generated in the central nervous system. Potent support has also been provided for Lorenz hypothesis postulating constant production of actionspecific potentials in the central nervous system. Habituation and imprinting. Most of the original schemata postulated for the functional structure of behavior have been shown to be simplifications, though correctly describing special cases. In order to arrive at a generalized schema, it is still necessary to modify repeatedly such hypothetical schemata.

Even the hierarchical system must be altered to a multidimensional network. The concept of habituation likewise increases in complexity with time. A behavior pattern repeatedly elicited by a particular sign-stimulus will cease to occur after a given time. The sense organ can nevertheless be demonstrated to be fully capable of functioning, and it is not even necessary to presume that the motivation of the animal to perform this pattern is entirely extinguished; it is often possible to elicit the same behavior pattern immediately afterward with another sign-stimulus. In such cases, it is necessary to assume that central cut-off systems are involved. Such systems are capable of very complex functions; mechanisms of this type in turkeys extinguish flight behavior in response to all relatively slow-flying objects which occur frequently, and the adult animal flees only in response to uncommon flying objects. In fact, the most infrequently occurring objects which (relative to their own size) are slow-flying are birds of prey. The adult turkey thus shows a well-adapted flight reaction in response to predatory birdsbut also to advertising balloons. It is a question of definition whether or not one refers to this effect as learning; it takes place without marked exogenous reward or punishment. The same is true of imprinting, which was described by Spalding as early as 1873. In imprinting, a specific reaction of the animal (which need not be functional at the time of imprinting) becomes attached to an object which later functions as the releasing agent. This occurs within a limited sensitive period, usually at a young age. In contrast to learning by association, there is no reward (as in the previous example), and even punishing stimuli have a reinforcing effect. If ducks or doves are reared by other species they will later show a pairing preference toward a partner belonging to the foster species, even when a conspecific partner is available. [seeIMPRINTING.] Phenomena similar to imprinting have been discovered in many fields. Some juvenile birds learn the song entirely from the father and will learn the song of a foster-father even in the midst of conspecifics. Since later offspring learn from these birds, the possibility of speech dialects arises. In different mammal species there have been cases of traditions which largely concern food preferences or forms of food acquisition, although traditions may also arise in the avoidance of enemies.