VOL. 28, NO. 4 December, I95 3

A CRITIQUE OF KONRAD LORENZ'S THEORY OF INSTINCTIVE BEHAVIOR BY DANIEL S. LEHRMAN The American Museum of Natural History and Rutgers University

EGINNING about 1931, Konrad Lorenz, the psychologists who constitute the majority of with his students and collaborators American students of animal behavior (Schneirla, (notably N. Tinbergen), has published 1945). numerous behavioral and theoretical In recent years Lorenz's theories have at- Bpapers on problems of instinct and in- tracted more and more attention in the United nate behavior which have had a widespread in- States as well, partly because of a developing in- fluence on many groups of scientific workers terest in animal behavior among American (Lorenz, 1931, 1932, 1935, 1937a; Lorenz and zoologists and ecologists, and partly through the Tinbergen, 1938; Lorenz, 1939; Tinbergen, 1939; receptive audience provided for Lorenz and his Lorenz, 1940, 1941; Tinbergen, 1942, 1948a, colleague, Tinbergen, by American ornithologists. 1950; Lorenz, 1950; Tinbergen, 1951). Lorenz's The ornithologists were interested from the start, influence is indicated in the founding of the especially because a great part of the material on Zeitschrqt juv Tierpsychologie in 1937 and in its which Lorenz based his system came from studies subsequent development, and also in the journal of bird behavior, but the range of interest in Behaviour, established in 1948 under the editor- America has widened considerably. Lorenz and his ship of an international board headed by Tin- theories were recently the subject of some discus- bergen. sion at a conference in New York at which Lorenz's theory of instinctive and innate be- zoologists and comparative psychologists were havior has attracted the interest of many investi- both represented (Riess, 1949), and are promi- gators, partly because of its diagrammatic sim- nently represented in the recent symposium on plicity, partly because of its extensive use of animal behavior of the Society of Experimental neurophysiological concepts, and partly because Biologists (Armstrong, 1950; Baerends, 1950; Lorenz deals with behavior patterns drawn from Hartley, 1950; Koehler, 1950; Lorenz, 1950; the life cycle of the animals discussed, rather than Tinbergen, 1950), and extensively used in several with the laboratory situations most often found chapters of a recent American handbook of ex- in American comparative psychology. These perimental psychology which will be a standard factors go far toward accounting for the great sourcebook for some years to come (Beach, 1951a; attention paid to the theory in Europe, where Miller, 1951; Nissen, 1951). most students of animal behavior are zoologists, Because Lorenz's ideas have gained wide atten- physiologists, zoo curators or naturalists, unlike tion, and in particular because a critical discussion 338 THE QUARTERLY REVIEW OF BIOLOGY

of these matters should bring usefully into review toward the egg, frequently continues the sagittal Lorenz's manner of dealing with basic problems movement to completion much as though an egg in the comparative study of behavior, a recon- were present. The longer one waits after "ex- sideration of Lorenz's system and school seems haustion," the easier it is to reelicit the act. very desirable at this time. In connection with certain other behavior pat- terns, Lorenz (1937b) has noted that, after long resting intervals, the animal may perform a com- We may best represent the general character- plete act without any external stimulus. This istics of the theory under discussion in terms of a performance "in a vacuum" is regarded by Lorenz case analyzed by Lorenz and Tinbergen (1938). as the extreme case of the lowering of the threshold The many subsequent references to this case and of elicitation after long non-exercise of the act. He the proffered analysis by these authors and their calls it "Leerlaufreaktion," or "going off in a colleagues leave no doubt that the case and its vacuum" [translated by Tinbergen (1942) as treatment may stand as representative. ''vacuum activities"]. The sagittal movement thus has the following Egg-rolling i~zthe Gray Goose characteristics: (1) it displays a reaction-specific When a gray goose, sitting on its nest, sees an exhaustibility; (2) although released by stimuli egg that has rolled out of the nest, it reacts in a coming from the egg, once released it remains characteristic fashion. It extends its head toward constant in form regardless of variations in stimu- the egg and then, keeping its head and neck pointed lation from the environment and even of the pres- toward the egg and its eyes fixed upon it, stands ence or absence of the original releasing stimulus; up and slowly steps forward to stand on the rim and (3). . the threshold for elicitation falls continu- of the nest. Next the goose bends its neck down- ously during non-exercise of the act. ward and forward so that the egg rests against the The movement in the sagittal plane, distin- underside of the bill. It then proceeds to roll the guished from lateral deviations, is a typical "in- egg back into the nest by shoving it back between stinctive act" (Erbkoordination) in Lorenz's its legs, using the underside of the bill. At the system. This "instinctive act," of course, is only a same time that this movement of the head and part of the total behavior pattern of egg-rolling. neck is taking place in the sagittal plane, the goose However, Lorenz maintains that every "in- performs side-to-side movements of the head stinctive" behavior pattern has as its focus such an which have the effect of balancing the egg against "instinctive act" or "consummatory act" (Craig, the under-side of the bill. 1918), the performance of which serves as goal for much of the rest of the pattern. The instinctive act To Lorenz, the instinctive act is a rigidly stereo- The egg-rolling movement in the sagittal plane typed innate movement or movement pattern, may be considered first, without reference to based on the activity of a specific coordinating whatever side-to-side movements may occur, center in the central nervous system. In this co- since these two types of movement are distin- ordinating center, there is a continuous accumula- guished very sharply in the theory. tion of excitation or energy specific for the act. Lorenz and Tinbergen found that the goose's When the animal comes into the appropriate ex- tendency to perform the sagittal movement can be ternal situation for the performance of the act, "exhausted" by repeated elicitation, even though stimuli provided by that situation release the observations indicated that the muscles involved energy, the instinctive act is performed, and some evidently are not themselves fatigued. The authors or all of the excitation is used up. The center therefore concluded that what is exhausted is a specific for the act thus is able to coordinate the central neural mechanism. instinctive act completely independently of the The form of the sagittal movement is always receptors, so that once the act is released (i.e., much the same, regardless of variations in the elicited) its performance occurs in complete form, shape of the egg-object or irregularities in the path coordinated by impulses from the center and with- over which the egg is rolled. Furthermore, when the out any chain-reflex character. The function of the egg rolls away from its bill, the goose, instead of stimulus is to release or elicit the act. Once re- stopping the sagittal movement and reaching out leased, the act no longer depends for its form on ON LORENZ'S THEORY OF INSTINCTIVE BEHAVIOR 339 anything outside the central nervous system. is called the "innate releasing pattern." According When the animal happens not to be in the appropri- to Lorenz and Tinbergen (1938), "The innate re- ate stimulus-situation, this reaction-specific energy leasing pattern . . . [is] . . . the innately-determined is presumed to be accumulated, or dammed up. readiness of an animal to respond to a particular Also, the greater the amount of reaction-specific combination of external stimuli with a particular energy which has accumulated, the more easily behavior. It thus consists of an innate receptoral may the act be elicited and the more complete correlate of a combination of stimuli which, despite will be its form when elicited. its relative simplicity, characterizes a certain bio- This picture is regarded by Lorenz as a repre- logically-significant situation sufficiently uniquely sentation of the neurophysiological basis of the so that the animal will not normally perform the above-described functional characteristics of the appropriate reaction except in that situation." instinctive act. In particular, accumulation of The view, then, is that the innate releasing energy in a neural center capable of determining mechanism holds the instinctive act under in- the form and order of performance of the various hibition until there appears a specific innate re- movements of the act, independently of the re- leasing pattern of stimuli capable of switching off ceptors (except for a trigger-like elicitation), is the inhibition and "triggering" an outflow of postulated to explain the reaction-specific ex- impulses from the instinctive center to the periph- haustibility (using up of the specific energy), the eral effectors. presumed independence of the form of the act from Also, the higher the level reached by the ac- concurrent external stimulation (reaction-speci- cumulation of reaction-specific energy in the center, ficity of the energy), and the lowering of the the more difficult is presumed to be any inhibition threshold during a non-exercise interval (i.e., of the act by the releasing mechanism. Conse- when an accumulation of reaction-specific ex- quently, the less completely does any stimulus- citation is presumed to occur). combination need to fit the innate releasing pattern as a whole to elicit the act. This explanation is The innate releasing mechanism offered by Lorenz for the fact that the instinctive act is easier to elicit, the more time has elapsed If energy specific for the instinctive act can since it was last performed. The Leerlaufreaktion accumulate continuously in the neural center thus is a breaking of reaction-specific energy specific for that act, why is the act not continuously through the inhibiting barrier, when such energy performed? Tinbergen (1948a) concludes that each reaches a very high level. coordinating center is normally held under in- hibition by another center which functions to The taxis block impulses from the coordinating center, save under specific conditions of external stimulation. Side-to-side movements of the head, by which This postulated inhibiting center is called the the goose keeps the egg balanced against the under- "innate releasing mechanism." The effect of an side of the bill, unlike the sagittal movement, lack external stimulus which elicits an instinctive act the character of centrally-preformed movement is to release the instinctive center from this patterns. The side-to-side movements are believed inhibition. to be elicited independently by contact stimulation For example, the sagittal component of the of the underside of the bill by the egg being rolled. goose's egg-rolling movement is not performed con- Whenever the egg rolls off center, a bill movement tinuously, even though energy specific for it is toward the side of the deviation restores the egg to being produced continuously in the central nervous the path. If the egg happens to roll free so that the system. The movement is only performed in a bird may continue the sagittal movement without particular stimulus-situation: i.e., when a smooth- any egg, there are no accompanying side-to-side outlined hard object is present near the nest. (In movements. When the bird is permitted to roll a non-experimental situations, such an object in cylinder, there are usually no side-to-side move- such a place will almost always be an egg that has ments, since a cylinder is unlikely to roll from side rolled out of the nest.) This combination of stimuli, to side in its path. which is considered capable of releasing the par- Thus the side-to-side movement is not only ticular instinctive act from the inhibition under elicited by external stimuli like the sagittal move- which it is held by the innate releasing mechanism, ment, but is also continuously wiented with respect 340 THE QUARTERLY REVIEW OF BIOLOGY

to external stimuli while being performed. In this up a piece of food in its paws, bite at it, and chew respect lateral deviations differ ' fundamentally it. Now, the whole sequence of behavior in this from the instinctive act, the form of which is hypothetical example would be regarded by Lorenz determined centrally so that the external stimulus as appetitive to the instinctive act of chewing. For acts as a trigger only. him, turning toward the corner is a taxis, the Movements like the side-to-side movements, walking is an instinctive act, picking up the food which are continuously oriented to stimuli during might be an instinct, turning the head toward the their performance, are called by Lorenz orienting food held in the paws is a taxis, and the chewing movements, or taxes. A taxis may occur simul- an instinct. Which corner the rat turns toward taneously with an instinctive act (as in the case of depends on his past experience-the taxis is thus the goose's egg-rolling), or may occur interspersed partly learned. The turning of the head toward with instincts in a behavior-chain. the food in the paws, however, might be innate. The stimuli releasing the act (innate releasing [Tinbergen (1942) points out that some taxes may pattern for the instinctive act), according to be learned, others innate. According to Lorenz's Tinbergen (Tinbergen and Kuenen, 1939; Tin- system, however, all instinctive acts (as distinct bergen, 1942) are not necessarily the same as those from taxes) are innate.] guiding it (i.e., the innate receptor pattern for the For Lorenz, the whole complex of behavior in taxis). In the case of egg-rolling, for example, the this example, involving instincts and taxes, learned instinctive act is released by a combination of and innate elements, has at its core the act of visual stimuli and tactual stimuli related to the chewing, and is motivated by the excitation set up hardness of the egg, as felt by the tip of the bill in by the neural center for chewing. The appetitive tapping. The (side-to-side) taxis, on the other behavior continues until the instinctive act is hand, is released by tactual stimuli on the under- performed, and the specific energy is thus used up. side of the bill. It is important to note that according to Lorenz the goal of the appetitive behavior is the per- Appetitive behavior formance of the act, not its biological result. That is The first part of the goose's reaction to the egg to say, in our hypothetical case, the need of which outside the nest is a stretching of the neck forward the appetitive restlessness is an expression is re- and downward, toward the egg. This act, according duced not by the introduction of food into the to Lorenz, has a different character from the in- stomach, but by the act of chewing. This is shown stinctive act itself. It serves the purpose of getting more clearly in the case of instinctive acts like the animal into the particular situation in which a courtship displays of birds, which form the goal of specific instinctive act can be released. The act appetitive behavior (moving toward the female, thus is truly goal-directed, according to Lorenz, orienting to the female, etc.) and which according who terms it "appetitive behavior" (Appetenzver- to Lorenz are subject to Leerlaufreaktionen even halten). He regards all goal-directed behavior as though they do not (like chewing and swallowing) appetitive, in the sense that such acts are directed result in the satisfaction of an a$parent peripheral toward getting the animal into a situation in which tissue need. some instinctive act can be released. For him, PROBLEMS RAISED BY INSTINCT THEORIES appetitive behavior can be of enormous com- plexity, involving instincts, taxes, and learned Even this brief summary brings to light several behavior of various kinds. Such behavior normally questions which ought to be critically examined occurs when the level of excitation in the central with reference to the theory. These are questions, nervous system for any instinctive act becomes furthermore, which apply to instinct theories in high enough. This causes the animal to become general. Among them are: (1) the problem of restless and active. What specific kind of activity "innateness" and the maturation of behavior; (2) may occur depends on the kind of animal, and on the problem of levels of organization in an organ- which instinct is the source of the appetitive ism; (3) the nature of evolutionary levels of be- restlessness. For example, a rat set into activity by havioral organization, and the use of the com- a high level of energy specific for the instinctive parative method in studying them; and (4) the act of eating (i.e., he wants to eat) may turn manner in which physiological concepts may be toward a corner of the cage, walk toward it, pick properly used in behavior analysis. There follows ON LORENZ'S THEORY OF INSTINCTIVE BEHAVIOR 341 an evaluation of Lorenz's theory in terms of these sarily imply that Lorenz's interpretation of these general problems. behavior patterns as "innate" offers genuine aid to a scientific understanding of their origin and of the "Innateness" of behavior mechanisms underlying them. The problem In order to examine the soundness of the con- cept of "innateness" in the analysis of behavior, it Lorenz and Tinbergen consistently speak of be- will be instructive to start with a consideration of havior as being "innate" or "inherited" as though one or two behavior patterns which have already these words surely referred to a definable, definite, been analyzed to some extent. and delimited category of behavior. It would be impossible to overestimate the heuristic value Pecking in the chick which they imply for the concepts "innate" and "not-innate." Perhaps the most effective way to Domestic chicks characteristically begin to peck throw light on the "instinct" problem is to con- at objects, including food grains, soon after hatch- sider carefully just what it means to say that a ing (Shepard and Breed, 1913; Bird, 1925; Cruze, mode of behavior is innate, and how much insight 1935; and others). The pecking behavior consists of this kind of statement gives into the origin and at least three highly stereotyped components: nature of the behavior. head lunging, bill opening and closing, and swal- Tinbergen (1942), closely following Lorenz, lowing. They are ordinarily coordinated into a speaks of instinctive acts as "highly stereotyped, single resultant act of lunging at the grain while coordinated movements, the neuromotor apparatus opening the bill, followed by swallowing when the of which belongs, in its complete form, to the grain is picked up. This coordination is present to hereditary constitution of the animal." Lorenz some extent soon after hatching, and improves (1939) speaks of characteristics of behavior which later (even, to a slight extent, if the chick is pre- are "hereditary, individually fixed, and thus open vented from practicing). to evolutionary analysis." Lorenz (1935) also This pecking is stereotyped, characteristic of the refers to perceptual patterns ("releasers") which species, appears in isolated chicks, is present at the are presumed to be innate because they elicit time of hatching, and shows some improvement in "instinctive" behavior the jrst time they are pre- the absence of specific practice. Obviously, it sented to the animal. He also refers to those motor qualifies as an "innate" behavior, in the sense patterns as innate which occur for the first time used by Lorenz and Tinbergen. when the proper stimuli are presented. Lorenz's Kuo (1932a-d) has studied the embryonic de- student Grohmann (1938), as well as Tinbergen velopment of the chick in a way which throws con- and Kuenen (1939), speak of behavior as being siderable light on the origin of this "innate" be- innately determined because it matures instead of havior. As early as three days of embryonic age, developing through learning. the neck is passively bent when the heartbeat It is thus apparent that Lorenz and Tinbergen, causes the head (which rests on the thorax) to rise by "innate" behavior, mean behavior which is and fall. The head is stimulated tactually by the hereditarily determined, which is part of the origi- yolk sac, which is moved mechanically by amnion nal constitution of the animal, which arises quite contractions synchronized with the heartbeats independently of the animal's experience and which cause head movement. Beginning about one environment, and which is distinct from acquired day later, the head first bends actively in response or learned behavior. to tactual stimulation. At about this time, too, the It is also apparent, explicitly or implicitly, that bill begins to open and close when the bird nods- Lorenz and Tinbergen regard as the major criteria according to Kuo, apparently through nervous of innateness that: (1) the behavior be stereotyped excitation furnished by the head movements and constant in form; (2) it be characteristic of the through irradiation in the still-incomplete nervous species; (3) it appear in animals which have been system. Bill-opening and closing become inde- raised in isolation from others; and (4) it develop pendent of head-activity only somewhat later. fully-formed in animals which have been prevented After about 8 or 9 days, fluid forced into the throat from practicing it. by the bill and head movements causes swallowing. Undoubtedly, there are behavior patterns which On the twelfth day, bill-opening always follows meet these criteria. Even so, this does not neces- head-movement. 342 THE QUARTERLY REVIEW OF BIOLOGY

In the light of Kuo's studies the "innateness" of all normal rats; they occur in rats which have been the chick's pecking takes on a different character raised in isolation; and they occur with no evidence from that suggested by the concept of a unitary, of previous practice, since both are performed innate item of behavior. Kuo's observations well by primiparous rats (retrieving may take place strongly suggest several interpretations of the for the first time only a few minutes after the birth developnzent of pecking (which, of course, are sub- of the first litter of a rat raised in isolation). Both ject to further clarification). For example, the behavior patterns therefore appear to satisfy the head-lunge arises from the passive head-bending criteria of "innateness" (Wiesner and Sheard, which occurs contiguously with tactual stimulation 1933). of the head while the nervous control of the Riess (pers. corn.), however, raised rats in iso- muscles is being established. By the time of hatch- lation, at the same time preventing them from ing, head-lunging in response to tactual stimulation ever manipulating or carrying any objects. The is very well established (in fact, it plays a major floor of the living cage was of netting so that feces role in the hatching process). dropped down out of reach. All food was powdered, The genesis of head-lunging to visual stimulation so that the rats never carried food pellets. When in the chick has not been analyzed. In Amblystoma, mature, these rats were placed in regular breeding however, Coghill (1929) has shown that a closely cages. They bred, but did not build normal nests or analogous shift from tactual to visual control is a retrieve their young normally. They scattered consequence of the establishment of certain nesting material all over the floor of the cage, and anatomical relationships between the optic nerve similarly moved the young from place to place and the brain region which earlier mediated the without collecting them at a nest-place. lunging response to tactual stimulation, so that Female rats do a great deal of licking of their visual stimuli come to elicit responses established own genitalia, particularly during pregnancy during a period of purely tactual sensitivity. If a (Wiesner and Sheard, 1933). This increased licking similar situation obtains in the chick, we would be during pregnancy has several probable bases, the dealing with a case of intersensory equivalence, relative importance of which is not yet known. The in which visual stimuli, because of the anatomical increased need of the pregnant rat for potassium relationships between the visual and tactual salts (Heppel and Schmidt, 1938) probably ac- regions of the brain, became equivalent to tactual counts in part for the increased licking of the stimuli, which in turn became effective through an salty body fluids as does the increased irritability already analyzed process of development, which of the genital organs themselves. Birch (pers. com.) involved conditioning at a very early age (Maier has suggested that this genital licking may play and Schneirla, 1935). an important role in the development of licking The originally diffuse connection between head- and retrieving of the young. He is raising female lunge and bill-opening appears to be strengthened rats fitted from an early age with collars made of by the repeated elicitation of lunging and billing rubber discs, so worn that the rat is effectively by tactual stimulation by the yolk sac. The re- prevented from licking its genitalia. Present peated elicitation of swallowing by the pressure of indications, based on limited data, are that rats so amniotic fluid following bill-opening probably is raised eat a high percentage of their young, that important in the establishment of the post- the young in the nest may be found under any part hatching integration of bill-opening and of the female instead of concentrated posteriorly swallowing. as with normal mother rats, and that retrieving does not occur. Maternal behavior in the rat These considerations raise some questions con- Another example of behavior appearing to fulfil cerning nativistic interpretations of nest-building the criteria of "innateness" may be found in the and retrieving in the rat, and concerning the mean- maternal behavior of the rat. ing of the criteria of "innateness." To begin with, Pregnant female rats build nests by piling up it is apparent that practice in carrying food pellets strips of paper or other material. Mother rats will is partly equivalent, for the development of nest- "retrieve" their pups to the nest by picking them building and retrieving, to practice in carrying up in the mouth and carrying them back to the nesting-material, and in carrying the young. nest. Nest-building and retrieving both occur in Kinder (1927) has shown that nest-building ac- ON LORENZ'S THEORY OF INSTINCTIVE BEHAVIOR 343 tivity is inversely correlated with environmental chick inside the egg might be regarded as the ideal temperature, and that it can be stopped by raising isolation experiment. the temperature sufficiently. This finding, together It must be realized that an animal raised in with Riess's experiment, suggests that the nest- isolation from fellow-members of his species is not building activity arises from ordinary food (and necessarily isolated from the eject of processes and other object) manipulation and collection under events which contribute to the developnent of any conditions where the accumulation of certain types particular behavior pattern. The important question of manipulated material leads to immediate satis- is not "Is the animal isolated?" but "From what is faction of one of the animal's needs (warmth). the animal isolated?" The isolation experiment, if The fact that the rat is generally more active at the conditions are well analyzed, provides at best a lower temperatures (Browman, 1943; Morgan, negative indication that certain specified environ- 1947) also contributes to the probability that nest- mental factors probably are not directly involved building activity will develop. In addition, the rat in the genesis of a particular behavior. However, normally tends to stay close to the walls of its the isolation experiment by its very nature does not cage, and thus to spend much time in corners. This give a positive indication that behavior is "innate" facilitates the collection of nesting material into or indeed any information at all about what the one corner of the cage, and the later retrieving of process of development of the behavior really con- the young to that corner. Patrick and Laughlin sisted of. The example of the nest-building and (1934) have shown that rats raised in an environ- retrieving by rats which are isolated from other ment without opaque walls do not develop this rats but not from their food pellets or from their "universal" tendency of rats to walk close to the own genitalia illustrates the danger of assuming wall. Birch's experiment suggests that the rat's "innateness" merely because a particular hypothe- experience in licking its own genitalia helps to sis about learning seems to be disproved. This is establish retrieving as a response to the young, as what is consistently done by Tinbergen, as, for does its experience in carrying food and nesting example, when he says (1942) of certain behavior material. patterns of the three-spined stickleback: "The re- leasing mechanisms of these reactions are all in- Maturation-vs.-learning, or development? nate. A male that was reared in isolation . . .was The isolation experiment tested with models before it had ever seen another These studies suggest some second thoughts on stickleback. The . . . [stimuli] . . . had the same re- the nature of the "isolation experiment." It is leaser functions as in the experiments with normal males." Such isolation is by no means a final or obvious that by the criteria used by Lorenz and other instinct theorists, pecking in the chick and complete control on possible effects from experi- ence. For example, is the "isolated" fish unin- nest-building and retrieving in the rat are not fluenced by its own reflection from a water film or "learned" behavior. They fulfil all criteria of "innateness," i.e., of behavior which develops glass wall? Is the animal's experience with human handlers, food objects, etc., really irrelevant? without opportunity for practice or imitation. Yet, in each case, analysis of the developmental process Similarly, Howells and Vine (1940) have re- ported that chicks raised in mixed flocks of two involved shows that the behavior patterns con- varieties, when tested in a Y-maze, learn to go to cerned are not unitary, autonomously developing chicks of their own variety more readily than to things, but rather that they emerge ontogenetically those of the other variety. They concluded that the in complex ways from the previously developed "learning is accelerated or retarded because of organization of the organism in a given setting. . . . What, then is wrong with the implication of the the directive influence of innate factors." In this case, Schneirla (1946) suggests that the effect of "isolation experiment," that behavior developed in the chick's experience with its own chirping during isolation may be considered "innate" if the animal feeding has not been adequately considered as a did not practice it specifically? Lorenz repeatedly refers to behavior as being source of differential learning previous to the ex- innate because it is displayed by animals raised in periment. This criticism may also be made of a isolation. The raising of rats in isolation, and their similar study by Schoolland (1942) using chicks subsequent testing for nesting behavior, is typical and ducklings. of isolation experiments. The development of the Even more fundamental is the question of what 344 THE QUARTERLY REVIEW OF BIOLOGY is meant by "maturation." We may ask whether know much more about the course of establishment experiments based on the assumption of an abso- of the connections between the chick's head-lunge lute dichotomy between maturation and learning and bill-opening, and between bill-opening and ever really tell us what is maturing, or how it is swallowing. This does not mean that we expect to maturing? When the question is examined in terms establish which of the components is learned and of developmental processes and relationships, which matured, or "how much" each is learned and rather than in terms of preconceived categories, how much matured. The effects of learning and the maturation-versus-learning formulation of the of structural factors differ, not only from com- problem is more or less dissipated. For example, in ponent to component of the pattern, but also from the rat nest-building probably does not mature developmental stage to developmental stage. What autonomously-and it is not learned. It is not is required is a continuation of the careful analysis "nest-building" which is learned. Nest-building of the characteristics of each developmental stage, develops in certain situations through a develop- and of the transition from each stage to the next. mental process in which at each stage there is an Our scepticism regarding the heuristic value of identifiable interaction between the environment the concept of "maturation" should not be inter- and organic processes, and within the organism; preted as ignorance or denial of the fact that the this interaction is based on the preceding stage of physical growth of varied structures plays an development and gives rise to the succeeding stage. important role in the development of most of the These interactions are present from the earliest kinds of behavior patterns under discussion in the (zygote) stage. Learning may emerge as a factor in present paper. Our objection is to the interpretation the animal's behavior even at early embryonic of the role of this growth that is implied in the stages, as pointed out by Carmichael (1936). notion that the behavior (or a specific physiological Pecking in the chick is also an emergent-an substrate for it) is "maturing." For example, the integration of head, bill, and throat components, post-hatching improvement in pecking ability of each of which has its own developmental history. chicks is very probably due in part to an increase This integration is already partially established by in strength of the leg muscles and to an increase the time of hatching, providing a clear example of in balance and stability of the standing chick, "innate" behavior in which the statement "It is which results partly from this strengthening of the innate" adds nothing to an understanding of the legs and partly from the development of equi- developmental process involved. The statement librium responses (Cruze, 1935). Now, isolation or that "pecking" is innate, or that it "matures," prevention-of-practice experiments would lead to leads us away from any attempt to analyze its the conclusion that this part of the improvement specific origins. The assumption that pecking was due to "maturation." Of course it is partly due grows as a pecking pattern discourages examination to growth processes, but what is growing is not of the embryological processes leading to pecking. peckilzg ability, just as, when the skin temperature The elements out of whose interaction pecking receptors of the rat develop, what is growing is not nest-building activity, or anything isomorphic emerges are not originally a unitary pattern; they with it. The use of the categories "maturation- become related as a consequence of their positions vs.-learning" as explanatory aids usually gives a in the organization of the embryonic chick. The false impression of unity and directedness in the understanding provided by KUO'S observations growth of the behavior pattern, when actually the owes nothing to the "maturation-versus-learning" behavior pattern is not primarily unitary, nor does formulation. development proceed in a straight line toward the Observations such as these suggest many new completion of the pattern. problems the relevance of which is not apparent It is apparent that the use of the concept of when the patterns are nativistically interpreted. "maturation" by Lorenz and Tinbergen as well as For example, what is the nature of the rat's by many other workers is not, as it at first appears, temperature-sensitivity which enables its nest- a reference to a process of development but rather building to vary with temperature? How does the to ignoring the process of development. To say of a animal develop its ability to handle food in specific behavior that it develops by maturation is tanta- ways? What are the physiological conditions which mount to saying that the obvious forms of learning promote licking of the genitalia, etc.? We want to do not influence it, and that we therefore do not ON LORENZ'S THEORY OF INSTINCTIVE BEHAVIOR 345 consider it necessary to investigate its ontogeny obscures the necessity of investigating develop- further. mental processes in order to gain insight into the actual mechanisms of behavior and their inter- Heredity-vs.-environment, or development? relations. The problem of development is the prob- Much the same kind of problem arises when we lem of the development of new structures and consider the question of what is "inherited." It is activity patterns from the resolution of the inter- characteristic of Lorenz, as of instinct theorists in action of existing structures and patterns, within general, that "instinctive acts" are regarded by the organism and its internal environment, and him as "inherited." Furthermore, inherited be- between the organism and its outer environment. havior is regarded as sharply distinct from behavior At any stage of development, the new features acquired through "experience." Lorenz (1937a) emerge from the interactions within the current refers to behavior which develops "entirely inde- stage and between the current stage and the en- pendent of all experience." vironment. The interaction out of which the organ- It has become customary, in recent discussions ism develops is not one, as is so often said, between of the "heredity-environment" problem, to state heredity and environment. It is between organism that the "hereditary" and "environmental" con- and environment! And the organism is different at tributions are both essential to the development of each different stage of its development. the organism; that the organism could not develop Modern physiological and biochemical genetics in the absence of either; and that the dichotomy is is fast destroying the conception of a straight-line more or less artificial. [This formulation, however, relationship between gene and somatic character- frequently serves as an introduction to elaborate istic. For example, certain strains of mice contain attempts to evaluate what part, or even what a mutant gene called "dwarf." Mice homozygous percentage, of behavior is genetically determined for "dwarf" are smaller than normal mice. It has and what part acquired (Howells, 1945; Beach, been shown (Smith and MacDowell, 1930; Reeler, 1947a; Carmichael, 1947; Stone, 1947).] Lorenz 1931) that the cause of this dwarfism is a deficiency does not make even this much of a concession to of pituitary growth hormone secretion. Now what the necessity of developmental analysis. He simply are we to regard as "inherited"? Shall we change states that some behavior patterns are "inherited," the name of the mutation from "dwarf" to others "acquired by individual experience." I do "pituitary dysfunction" and say that dwarfism is not know of any statement of either Lorenz or not inherited as such-that what is inherited is a Tinbergen which would allow the reader to con- hypoactive pituitary gland? This would merely clude that they have any doubts about the cor- push the problem back to an earlier stage of de- rectness of referring to behavior as simply "in- velopment. We now have a better understanding herited" or "genically controlled." of the origin of the dwarfism than we did when Now, what exactly is meant by the statement we could only say it is "genically determined." that a behavior pattern is "inherited" or "genically However, the pituitary function developed, in controlled"? Lorenz undoubtedly does not think turn, in the context of the mouse as it was when that the zygote contains the instinctive act in the gland was developing. The problem is: What miniature, or that the gene is the equivalent of an was that context and how did the gland develop entelechy which purposefully and continuously out of it? tries to push the organisms's development in a What, then, is inherited? From a somewhat particular direction. Yet one or both of these pre- similar argument, Jennings (1930) and Chein (1936) concluded that only the zygote is inherited, formistic assumptions, or their equivalents, must or that heredity is only a stage of development. underlie the notion that some behavior patterns are "inherited" as such. There is no point here in involving ourselves in tautological arguments over the definition of The "instinct" is obviously not present in the zygote. Just as obviously, it is present in the be- heredity. It is clear, however, that to say a be- havior of the animal after the appropriate age. The havior pattern is "inherited" throws no light on its problem for the investigator who wishes to make a development except for the purely negative implica- causal analysis of behavior is: How did this be- tion that certain types of learning are not directly havior come about? The use of "explanatoryv involved. Dwarfism in the mouse, nest-building in categories such as "innate" and "genically fixedJ' the rat, pecking in the chick, and the "zig-zag THE QUARTERLY REVIEW OF BIOLOGY

dance" of the stickleback's courtship (Tinbergen, that the species-characteristic nature of many be- 1942) are all "inherited" in the sense and by the havior patterns may result partly from the fact criteria used by Lorenz. But they are not by any that all members of the species grow in the same means phenomena of a common type, nor do they environment. Smith and Guthrie (1921) call such arise through the same kinds of developmental behavior elements "coenotropes." Further, it is not processes. To lump them together under the rubric at all necessary that these common features of the of "inherited" or "innate" characteristics serves to environment be those which seem a priori to be block the investigation of their origin just at the relevant to the behavior pattern under study. point where it should leap forward in meaningful- Lorenz's frequent assumption (e.g., 1935) that the ness. [Anastasi and Foley (1948), considering data effectiveness of a given stimulus on first presenta- from the field of human differential psychology, tion demonstrates an innate sensory mechanism have been led to somewhat the same formulation specific for that stimulus is not based on analysis of the "heredity-environment" problem as is of the origin of the stimulus-effectiveness, but presented here.] merely on the fact that Lorenz has eliminated the major alternative he sees to the nativistic Taxonomy and Ontogeny explanation. Lorenz (1939) has very ably pointed out the Thorpe and Jones (1937) have shown that the ap- potential importance of behavior elements as parently innate choice of the larvae of the flour moth taxonomic characteristics. He has stressed the fact by the ichneumon fly Nemerites as an object inwhich that evolutionary relationships are expressed just to deposit its eggs is actually a consequence of the as dearly (in many cases more clearly) by similari- fact that the fly larva wasfed on the larvae of the ties and differences in behavior as by the more flour moth while it was developing. By raising commonly used physical characteristics. Lorenz Nemerites larvae upon the larvae of other kinds of himself has made a taxonomic analysis of a family moth Thorpe and Jones caused them, when adult, of birds in these terms (Lorenz, 1941), and others to choose preponderantly these other moths on have been made by investigators influenced by which to lay their eggs. The choice of flour-moth him (Delacour and Mayr, 1945; Adriaanse, 1947; larvae for oviposition is quite characteristic of Baerends and Baerends-van Roon, 1950). This Nemerites in nature. In view of Thorpe and Jones' type of analysis derives from earlier work on the work, it would obviously be improper to conclude taxonomic relations of behavior patterns by from this fact that the choice is based on innately- Whitman (1898, 1919), Heinroth (1910, 1930), determined stimuli. Yet, before their paper was Petrunkevitsch (1926), and others. published, the species-specific character of the be- Lorenz's brilliant approach to the taxonomic havior would have been just as impressive evidence analysis of behavior characteristics has had wide for "innateness" as species-specificity ever is. influence since it provides a very stimulating frame- Taxonomic analysis, while very important, is not work in which to study species differences and the a substitute for concrete analysis of the ontogeny specific characteristics of behavior. However, it of the given behavior, as a source of information does not necessarily follow from the fact that be- about its origin and organization. havior patterns are species-specific that they are "innate" as patterns. We may emphasize again Levels of Organization that the systematic stability of a characteristic Levels of "Innateness" does not indicate anything about its mode of de- velopment. The fact that a characteristic is a good Animals at different evolutionary levels show taxonomic character does not mean that it de- characteristic differences in the extent and manner veloped autonomously. The shape of the skull of learning. In addition, within the same animal's bones in rodents, which is a good taxonomic charac- behavior different activities may be more or less ter (Romer, 1945), depends in part upon the pres- susceptible to the influence of learning, and may be ence of attached muscles (Washburn, 1947). We affected in different ways by learning (Schneirla, cannot conclude that because a behavior pattern is 1948, 1949a). taxonomically stable it must develop in a unitary, Lorenz explains these facts in terms of the rich- independent way. ness of the animal's instinctive equipment. As In addition it would be well to keep in mind described above, his conception is that instinctive ON LORENZ'S THEORY OF INSTINCTIVE BEHAVIOR 347

behavior is sharply different from all behavior much of the learning- which is antecedent to the leading up to the performance of the instinct. This emergence of nest-building in the rat occurs after "appetitive" behavior is conceived of as the sole birth. evolutionary source of all learned and intelligent Shall we call those behavior patterns "innate" behavior. Thus he says: which develop before birth and not those which develop after? This would be fruitless in view of ". . .appetitive behavior, as the sole root of all "variable" behavior, not only is physiologically some- the demonstrated existence of prenatal condition- thing fundamentally different from the automatism of ing (Ray, 1932; Gos, 1933; Spelt, 1948; Hunt, instinctive behavior, but . . .the two different processes 1949), and unsatisfactory in view of the problem appear as "substitutes" (vikariierend) for each other, in that the higher (ph~logenetic)development of the of the so-called postnatal "maturation" of various one makes the other superfluous and stops its develop- "innate" behavior patterns (Grohmann, 1938). ment. The reaching of a higher psychic performance goes hand-in-hand with a reduction of the automatisms But we must recognize that different behavior that take part in the action, leaving a behavior pattern patterns may involve learning at different on- with the same function as the one originally existing" togenetic stages to different extents, and in differ- (Lorenz, 1937a). ent ways. For example, much less of the behavior Again: of the rat is divectly a consequence of the specific "It is a peculiarity of many behavior patterns of characteristics of its structure than in the case of higher animals, that innate instinctive elements and in- the earthworm (Maier and Schneirla, 1935). The dividz~ally-acqzlired elements immediately jollow each other, within a functionally unitary chain of acts. . . I involvement of learning in the development of the have characterized this phenomenon as instinct-training rat's behavior is different from and occurs at interlacement. Similar interlacements occur between instinctive acts and intelligent or insightful behavior. different developmental stages from that of the ... The essence of such an interlacement is that, within chick. Further, some responses of the rat (such as a chain of innate instinctive acts there is a definite licking of a painful spot) are very much less sub- point, which point is innately determined, where a learned act is inserted. This learned act must be ac- ject to change by learning than others, such as quired by each individual in the course of its ontogenetic care of young (Sperry, 1945; Uyldert, 1946). These development. In such a case, the chain of innate acts are not differences in the number of behavioral has a gap, in which, instead of an instinctive act, there is a 'capacity to acquire' " (Lorenz, 1937a). [All em- elements which are "innate," but rather in the way phases are Lorenz's.] in which the structures are involved in the develop- It is apparent that Lorenz regards differences in ment of behavior at different evolutionary levels the extent to which learning occurs as representing and for different behavior patterns. differences in the size of the gaps in the chain of Lorenz does not fully utilize the idea of levels of innate behavior. He considers any given "com- organization of behavior, apparently because his ponent" of behavior as "innate" or not "innate." concept of "innateness" is not the result of analysis This is entirely consistent with his virtual identi- of the development of behavior; it is in part the fication of "innate" with "autonomously result of a preconception that "innate" and "not- developing." innate" are the two categories into which behavior However. we have already tried to make it clear logically falls. Consequently Lorenz and his school that behavior patterns classified as "innate" by any have classified behavior as "innate" and "not- criterion do not all fall into the same category innate" on the basis of criteria which when care- with respect to embryonic origin, developmental fully examined appear to be arbitrary. Their category of "innate" therefore includes very differ- history,.. or level of organization. Lorenz notes that more or fewer of the components of behavior may ent kinds of behavior, which involve learning in be "innate." But nowhere does he recognize that many different ways. Lorenz's concept of "innate" one component may be more or less "innate" or behavior represents a lumping-together of many "innate" in one or another manner. We may different kinds and levels of behavior on the basis call attention to an important difference be- of an essentially phenotypic classification, and the tween the pecking of the chick and the nest- imposition of preconcieved categories upon that building of the rat, both behavior patterns classification. which develop without specific practice of the Evolutionary Levels patterns: a major part of the learning which appears to be antecedent to the emergence of Since Lorenz does not discuss the existence of pecking in the chick occurs before hatching, while qualitative differences with respect to modes of 348 THB QUARTERLY REVIEW OF BIOLOGY development within his category of "innate" be- structures so that the underlying processes and havior it is not surprising that his conception of the mechanisms are not the same. evolution of behavior lacks any notion of qualita- Lorenz's application of the concept of evolution- tive change. Lorenz maintains at all levels a sharp ary change does not consist of analyzing the differ- distinction between "instinctive acts" and "ap- ent ways in which behavior patterns at different petitive behavior" (which includes alI oriented, evolutionary levels depend on the structure and goal-directed, and variable types of behavior at all life of the organism. It consists rather of abstract- levels). He says: ing aspects of behavior, reifying them as specific autonomous mechanisms, and then citing them as "If we consider the unbroken series of forms of corresponding modes of behavior, which extends in a demonstrations of "evolution" in a purely descrip- smooth progression from protozoa to man, we must tive taxonomic sense. Taxonomically, this pro- determine that we cannot distinguish between taxis, on the one hand, and, on the other, behavior guided cedure is often extremely valuable, but by its by the simplest intelligence (Einsicht). We cannot here implicit assumption that "elements" of behavior distinguish between taxis and, in the case of our frog, maintain their nature regardless of change in the an intelligence which might (anthropomorphically speaking) be limited to the knowledge: 'There sits the organization in which they are embedded (more fly' " (Lorenz, 1937a). properly, we should say from which they emerge),- .. it hinders rather than helps analysis of the behavior This is restated in a later paper (Lorenz, 1939): patterns themselves. "No sharp line can be drawn between the simplest orienting-reaction and the highest 'insightful' Levels of Neural Organization behavior." It might be pointed out that whether we can Lorenz characterizes each instinctive act as distinguish various levels of behavioral organiza- depending on a specific center in the central nervous tion depends in part on our assiduity in attempting system which continuously produces a type of to distinguish them. Preconceptions about the excitation specific to the act, and which is partly number and kind of categories into which behavior "used up" when the act is performed. He uses the ought to fall naturally has an important effect on concept "used up" quite literally, even suggesting the kind of examination we make of behavior the existence of act-specific substances (Lorenz and patterns and the kinds of distinctions we find our- Tinbergen, 1938). One of the principal types of selves able to make among them. evidence used by Lorenz to support this conception In the quotation above we have translated as is the lowering of the threshold for release of the 'lsrnooth" (progression) Lorenz's word "stufenlose," act as a function of lapse of time since performance which might be more literally translated "without of the act. That is, the longer the animal has gone steps" or "without levels." This is a gratuitous and without performing the act, the easier it is to elicit. very misleading oversimplification on Lorenz's This is taken by Lorenz as evidence of accumula- part. The transition from protozoa to man is not tion of the reaction-specific energy in the central "step-less." There are characteristic structural nervous system. Lowering of the intensity of differences between phyletic levels, and these performance upon repeated elicitation is taken as differences are responsible for characteristic differ- further proof, since it may indicate the using up of ences in the organization of behavior. A protozoan the excitation faster than it can be produced. is not like a simpler man. It is a different kind of Lorenz and Tinbergen offer observations along organism, with behavior which depends in differ- these lines on mammals, birds, fish, and insects. ent ways on its structure. The analysis of behavior The hunting behavior of the dog (Tinbergen, mechanisms at different levels (Schneirla, 1946) 1942), food-begging of a young bird (Tinbergen and shows that it is frequently misleading to speak of Kuenen, 1939), fighting in a fish (Tinbergen, 1942), behavim patterns or elements as homologous when courtship flights of a butterfly (Tinbergen, Meeuse, they seem to serve similar (or the "same") func- Boerema, and Varossieau, 1942) are all offered as tions and have superficially similar characteristics. examples of instinctive acts having this kind of Analysis of structural organizations out of which physiological basis. the specific behavior patterns emerge shows that Lorenz and Tinbergen adduce as physiological similar behaviors at different phyletic levels often evidence for this interpretation a series of studies are end-products of evolutionary selection leading by von Holst (1935-1937), on the mechanisms of to the similar behavior, but deriving from different locomotion in fishes. Von Holst observed that ON LORENZ'S THEORY OF INSTINCTIVE BEHAVIOR 349

almost completely deafferented fishes show some Gray and Lissman (1940,1946a, 1946b; Gray, 1939, of the coordinations of locomotion. He concluded 1950; Lissmann, 1946a, 1946b) have studied the ef- that the basic movement patterns of locomotion fect of deafferentation on locomotion in fish and are the result of the accumulation of locomotion- amphibians. Both regard their evidence as being specific energy in the central nervous system inde- against the probability of automatic-rhythmic pro- pendent of peripheral activity, and are the result duction of coordinations in the central nervous sys- of a central (non-reflex) coordination. tem, even at the fish level, and even for locomo- In his original consideration of von Holst's tion. Lissmann, in fact, designed his experiments work, Lorenz (1937b) stated that it would be pre- specifically in view of von Holst's observations, and mature to make positive assertions about the direct explicitly in view of the use made of the latter by relevance of that work to his instinct theory. Lorenz. He concluded, on the basis of a complete af- Lorenz and Tinbergen (1938), at about the same ferent isolation of the central nervous system, that time, stated that the conception of locomotion as there was no central automatic production of excita- an example of instinctive coordination in Lorenz's tion. Itshould benoted that the experiments of Weiss sense might be a very rough simplification of the (1936, 1937a-d, 1941, 1950) support the conclu- facts. sion that spinal centers in amphibians are so organ- However, over the years since then, the relevance ized that the coordination of locomotor patterns is of this kind of evidence in their writings seems to dependent upon characteristics of the centers. have increased, although some doubt has been Gray and Lissmann's experiments, however, show thrown on the validity of von Holst's conclusions. that proprioception actually plays a major role in One might question the direct relevance of the the normal ambulatory rhythms, even of these neural mechanisms of locomotion in fish and am- animals. phibians to the explanation of the origins of com- Tinbergen (1942) has expanded Lorenz's concept plex LLinstinctive"behavior in birds and mammals. of neural organization to include higher levels of Tinbergen is aware of the dangers inherent in the physiological and behavioral organization than the procedure of using physiological evidence from stereotyped "instinctive act" or consummatory lower evolutionary levels, lower levels of neural act. Tinbergen conceives of instinctive behavior in organizations, and simpler forms of behavior as general as being hierarchically organized in the analogies for the support of physiological theories individual. For example, in the reproductive be- of behavior mechanisms at higher and more com- havior of the stickleback Tinbergen sees three plex levels. For example, after a description of this main levels of organization, hierarchically ar- aspect of Lorenz's theory, Tinbergen (1948a) says: ranged. The highest level represents the reproduc- "These formulations are supported by the entirely tive drive in general. This corresponds to a center independent investigations which have been con- at a high level in the nervous system which when ducted during the last ten years on the central activated (by external conditions, hormones, or nervous mechanisms of locomotion. Here, to be sure autonomous cyclicity) sends impulses to a whole on a lower level of integration [my emphasis- group of intermediate centers, making the latter D. S. L.], we are brought to a fundamentally similar capable of activity. Each of these intermediate position by the researches of von Holst, Weiss, centers corresponds to a behavior pattern involved Gray, Lissmann, and others." in reproductive activities: fighting, nest-building, In this case, Tinbergen's mention of the fact courting, parental behavior, etc. Each of these that the physiological evidence comes from a intermediate centers, in turn, activates (or con- lower level of integration is actually embedded in tributes to the disinhibition of) a group of lower an expanded use of these data to support theories centers each of which coordinates a particular act based on observations at higher levels. This is which is released by an innate releasing pattern. merely a formal bow to the concept of levels, which For example, the fighting center, when activated appears to strengthen the form of the argument by external stimulation (which can only occur while actually weakening its content. when its threshold is lowered by activity in the In point of fact, it is now doubtful whether even superordinated center for the reproductive drive so simple a behavior pattern as locomotion, in so in general) puts the animal in "fighting mood" simple a vertebrate as a fish, is really organized which makes possible the performance under in the way that Lorenz's instinct theory demands. proper stimulus-conditions of each of the acts 350 THE QUARTERLY REVIEW OF BIOLOGY involved in fighting: biting, chasing, threatening, produced in such centers. As suggested by Gray etc. The latter are the "consummatory acts" or and Lissmann, rhythmicity of behavior is much "instinctive acts" of Lorenz. [N. B. The slight more parsimoniously explained in terms of periodic differences in terminology sometimes occurring shifts in balance between central and peripheral between Tinbergen and Lorenz do not usually processes or interaction between different central represent theoretical or conceptual disagreements.] processes, than in terms of the production of Tinbergen cites the work of Hess (1943, 1949) as periodic impulses by a single "center" which, in demonstrating the reality of these autonomous Lorenz's treatment, has the character of a "thing centrally-coordinated centers of activity. By in itself." electrically stimulating various points in the hypo- Lorenz (1950) describes in some detail a hy- thalamus, Hess was able to cause cats to perform draulic model, or analogy, of the instinct mechan- sleeping, eating, and other behavior which was not ism, including a reservoir of excitation and de- related to specific external stimulation, and which vices for keeping it dammed up (innate releasing ceased upon cessation of the stimulation. Tin- mechanism) until appropriate keys unlock the bergen regards these observations, with those of sluices. Hydraulic analogies have reappeared so von Holst, as demonstrating the reality of Lorenz's regularly in Lorenz's papers since 1937 as to justify picture of centers of automatic-rhythmic produc- the impression that they are not really analogies- tion of action-specific excitation. they are actual representations of Lorenz's con- Hess found, however, that there was consider- ception of the origin and channelling of "in- able variation in the responses to repeated stimu- stinctive energy." [The basic assumptions-of a lation of a specific spot in the hypothalamus. special center producing a reservoir of energy Stimulation of the same point might elicit quite specific for each instinct, and of devices for dis- different responses, depending upon the conditions tributing the energy-are very similar to those of of afferent inflow. This led Hess to conclude that MacDougall (1923, 1930).] there was not strict localization of function in the There is no neurophysiological evidence for such hypothalamus as (he assumed) there is in the cor- hydraulics in the brain. Aside from the contro- tex. [It might be pointed out that recent discus- versial aspects of the idea of automatic-rhythmic sions of cortical function indicate considerable production of excitation, such hydraulic concep- doubt about the reality of localized functions tions simply do not conform with what we actually isomorphic with their behavioral expressions, even know about the complexities of brain function (cf. in the motor areas of the cortex (Lashley, 1923; Fulton, 1949). Hines, 1947; Clark and Ward, 1948; Clark, 1948).] The actual physiological relationships underly- Now, a strictly punctate localization of function ing behavior patterns must be analyzed for the is not necessary, either in cortex or in hypo- different behavior patterns concerned. The as- thalamus, in order that these organs be able to sumption which underlies Lorenz's approach to the serve organizing and coordinating functions. In neurophysiological basis of behavior is that the the light of Hess's work there is no doubt that the neural events underlying behavior patterns must lower-level details and components of many be- somehow be isomorphic with the behavior itself. havior patterns are coordinated and integrated in He is thereby led to assume that behavior patterns the hypothalamus. But it is difficult to see how the having similar functional characteristics must be shifting locus of this integration can be reconciled caused by identical neural mechanisms. Lashley with the conception of a celzter which produces an (1942) has pointed out the erroneous nature of excitation specific for the behavior patterns con- such reasoning. It is by thus abstracting phenotypic cerned. It is equally difficult to reconcile the fact resemblances in behavior at different levels, and by that the function of a "center" depends partly on gratuitously transferring physiological explanations the type of afferent inflow present with the notion from one level to another that Lorenz creates the of the center as a place where excitation is produced impression that "instinctive" acts are grounded in for a particular kind of act. a common type of mechanism which is the same at different evolutionary levels. Neither do the researches of Hess, nor those of Gray and Lissmann, support the idea that rhythm- Levels of Behavioral Function icity or cyclicity of behavior is a function of the As already pointed out, a serious question facing periodic reaching of a threshold level of energy all investigators of animal behavior is the extent to ON LORENZ'S THEORY OF INSTINCTIVE BEHAVIOR 35 1

which different mechanisms may be assumed to be tozoan response and the human. At each level, the identical because of the apparent similarities in the mechanisms underlying this characteristic and behavior patterns they underlie. By this I do not widely-distributed response (Maier and Schneirla, mean to imply that the similarities may be un- 1935) are derived from the specific structure of the important, but only that functionally similar be- organism in question. The behavior patterns are havior patterns may be effectuated through very not homologous, although they may in some cases dissimilar causal mechanisms. And if the causal be based to different extents and in different ways mechanisms hypothesized in the case of one of the on more or fewer homologous structures. The behavior patterns are conceptually reified and amlysis of the behavior at each level must be in applied to other patterns or other animals, because terms of its emergence from the structure of organ- of the fact that the similar behavior patterns are isms at that level, as indicated in the examples of subsumed under the same term or included in the the amoeba and the neonate child, and not on super- same category or concept, the analysis of the ficial comparisons of the behavior with similar mechanisms actually operating in the different behavior patterns at other levels. cases is seriously hampered. Rather than making a Lorenz's concept of "instinct" represents, I developmental analysis of the processes concretely think, precisely this kind of undesirable reification underlying each behavior pattern, the predominant of a hypothesized mechanism. Lorenz's use of the tendency is to carry out brief studies on a variety term "instinct" does not denote merely a group of of selected examples assumed to demonstrate the behavior patterns characterized by certain com- validity of the a priori "principle" or the reality mon functional characteristics; it denotes a definite of the hypothesized structure or "center." class of things-a specific group of homologous This practice may produce very fallacious structures underlying acts whose characteristics are results. For example, both the amoeba and the isomorphic with those of the structures. And the neonate infant will move toward weak stimulation nature of the structures is inferred from the be- and away from strong stimulation (the amoeba as havioral characteristics, supported by physiological a whole, the child locally). In both animals, this evidence the inadequacy of which has already serves the biological function of bringing the organ- been pointed out. ism into contact with food (and for the child, pro- This reification of the concept of "instinct" leads tection), and away from contact with harmful to a "comparative" psychology which consists of stimuli. This similar biological utility is a sufficient comparing levels in terms of resemblances between basis of explanation for the evolutionary develop- them, without that careful consideration of dijer- ment of the similar modes of behavior in the two ences in organization which is essential to an organisms. understanding of evolutionary change, and of the But the mechanisms underlying the response in historical emergence of new capacities. Thus the the two animals manifestly must be very different. lowering of intensity of response as a consequence In the amoeba, the differential response to weak of repeated elicitation, in the case of certain sexual and strong stimuli is caused by the differential activities of a butterfly (Tinbergen, Meeuse, effects of the weak and strong stimuli on the sol-gel Boerema, and Varossieau, 1942) and of a fish relationship in the protoplasm of the single cell (Tinbergen, 1942) is taken in both cases as verify- (Mast, 1926). In the neonate child, the basis is ing Lorenz's assumption of the nature of the more obscure. Schneirla (1939) has suggested that organizing center for an instinctive act. The fact initially it is the result of differences in arousal- that some behavior patterns of a butterfly may threshold between flexor and extensor muscles of exhibit functional similarities to some behavior the limbs, so that they respond optimally to differ- patterns of a fish isinteresting as an indication that ent impulse-frequencies in afferent volleys, corre- similar response characteristics may be species- sponding to different intensities of stimulation. preserving in both cases. But it is not very judi- These two behavior patterns may seem function- cious, and actually is rash, in view of the very ally quite analogous. Can we say that they are different types of organization involved in the homologous? This would obviously be absurd. structure and the behavior of the two animals They represent two totally different kinds of ad- concerned, to assume that the mechanisms under- justment, both selected (in the evolutionary sense) lying the two similar response characteristics are because they serve the same kind of function. Nor in any way identical, homologous, or even similar, is there a "smooth progression" between the pro- or that there is any historical (evolutionary) con- 352 THE QUARTERLY REVIEW OF BIOLOGY tinuity between them as such. Yet this is precisely Tinbergen (1942) further classifies taxes into the basis of Lorenz's whole treatment of "instinct" several categories, based partly on Kiihn's (1919) and evolution. analysis: (1) tropotaxis, equivalent to Loeb's In addition to distorting comparative study and tropisms, in which the animal turns until the the study of evolutionary change, this reification relevant stimulus is equally intense on both sides; of "instinct" has unfortunate effects on the study (2) telotaxis, a visual orientation based on fixation of ontogenetic development. The development of movements so that either eye can serve as the sole an "instinctive" act inevitably appears to Lorenz to receptor; and (3) menotaxis, like telotaxis except be the self-differentiation of a preformed, autono- that the orientation, instead of being toward the mous thing. Thus Lorenz sees the developing be- stimulus, is at a constant angle from it. To these havior of the animal as progressing toward the full- categories of Kiihn, Tinbergen added a fourth: blown "instinct" rather than as developing out of pharotaxis, in which the animal is oriented to a interactions among processes present at that stage. part of the visual field defined in terms of its rela- This is a teleology which is inherent in Lorenz's tion to the rest of the field, irrespective of the approach, and which cannot be eliminated by his animal's orientation to the field. formal attempt to deny teleological and purposive This classification of "taxes" solely in terms of a procedures and to exclude the terminology. highly restricted definition of the receptor processes For example, Lorenz mentions the development inevitably lumps together many very different of fighting behavior in ducks. When fighting with processes. For example, our amoeba and newborn another drake, an adult drake will grasp its op- infant both show a "turning-to" reaction to mild ponent's neck in its bill and strike at him with a stimulation. What possible category, based on the wing. Lorenz noted that ducklings whose wings had characteristics of the turning, could properly not yet feathered would perform the same move- include both of these movements as examples of one ments even though the stubby, unfledged wing kind of process? To say that the movement of the was not yet long enough to strike the opponent. child and of the amoeba are both a taxis is to admit Lorenz's interpretation (1937a) is that the in- that the word "taxis" does not define a group stinctive act had matured before the full matura- of behavior patterns which have common tion of the structure which was used by it. This mechanisms. interpretation does not explain what the duckling Tinbergen (1942) makes this explicit when he is doing; rather, it prevents the investigator from says: ". . .in the concept of pharotaxis the part seeing the problem of what it is about the duckling played by mnemic processes is not taken as a and its situation (and its ontogenetic history) criterion, because in tropo-, meno-, and telotaxis which gives rise to this kind of behavior. This the criterion upon which the 'distinction' is based type of theory apparently causes the investigator also leaves this topic out of consideration . . . to look at the process of development in such a Menotaxis, for instance, can be innate or learned." way that the problem of the origin of this behavior, What then remains of Tinbergen's classification? and its cause and role in the duckling are not con- Tinbergen himself is aware that the members of sidered by him at all. In the light of our previous any of his taxis-categories probably differ widely discussion, it would appear that these are the in ontogenetic origin and central mechanism. How crucial problems, and that a theory which makes can the classification be justified? A preliminary them appear as relatively irrelevant to the explana- classification has heuristic value only if the mem- tion of the development of fighting behavior must bers of a given class are thought to be representa- be seriously lacking. tive of similar dynamic processes which can be This conceptual merging of very different levels investigated. That, in fact, is the purpose which on the basis of superficial similarities permeates Tinbergen assumes for his classifications. But in the system. For example, the concept of "taxis" as the case of taxes, the classifications are known to a meaningful class of behavior elements seems to be contain different levels of organization and differ- based on such a procedure. Lorenz defines a taxis ent processes. In this case the classification is as a movement which is continuously oriented based on the analogizing which appears to be with respect to the stimulus (thus distinguishing it basic to the Lorenz approach. sharply from an instinctive act which, once started, Lorenz (1939) and Tinbergen (1942) have both is centrally coordinated, independent of the pointed out that, under the influence of natural receptors). selection, widely divergent species may develop ON LORENZ'S THEORY OF INSTINCTIVE BEHAVIOR 353 similar characteristics which should not be assumed science and so many other activities certainly have to be homologous. Tinbergen (1948) g'ives an connections with internal factors of this kind." example of this "convergent evolution": "The most Here the implication seems to be that, because striking example of how far convergencies can go both are "spontaneous" and neither is mainly in these phenomena, is given by L. Tinbergen caused by proprioceptive stimuli or hormones (1939) in his study of the mating behavior of the (itself a gratuitous assumption), therefore the cuttlefish Sepia ofiicinalis. Parallel with the de- causes of hunting activity in the dog are the same velopment of eyes in cephalopods (convergent to as (or belong to the same class as) those of scientific those of fish) the courtship of the cuttlefish has activity in human beings! It is obvious that this evolved into a typically visual one closely resem- argument is based on the most casual and un- bling the courtship of certain sexually dimorphic analytical kind of comparison, and a lack of con- fish, lizards, and birds. . ..This state of affairs cern with the specific origins of the behavior pat- closely resembles that found by Noble and Bradley terns at issue. (1933) in [the lizard] Sceloporus. In both species Lorenz (1937b), speaking of the evolutionary the male's display is primarily a means of threaten- relation between instinctive and learned acts, ing other males. . . ." says: It will be noted that Tinbergen specifically notes "The presence of an instinctive act also seems to be that the resemblance is caused by convergence, detrimental to the development of an intelligent process rather than homology. However, his treatment of having the same function. At least, it is true of humans. the behavior patterns involves the implicit as- To be convinced of the correctness of this statement, one has only to consider the behavior of highly intelligent sumption that the convergence is one of meclza- men who have otherwise good critical faculties, when nisms. Actually, as far as we know, the convergence they 'fall in love' to carry out the undoubtedly instinc- is only of outcomes. The assumption that the tive reaction of mate-selection. The already-mentioned example of the ravens and jackdaws shows that higher mechanisms underlying these similar outcomes are psychological development may occur without any equally similar is both characteristic and reduction of the instinctive, innate members of a be- gratuitous. havior chain. ..." I include the last sentence to show how very The Human Level easily Lorenz switches from man to bird without This analogizing and confusion of levels becomes any apparent awareness that he is discussing phe- patently shallow when either Lorenz or Tinbergen nomena which may be very different. The point of discusses human behavior. Lorenz's statement seems to be that men fall in For example, Tinbergen (1942) says, "The love irrationally because "falling in love" is an activation of other drives, too, leads to searching instinctive reaction released by an innately-deter- behavior. Classical examples are the searching for mined situation. In this case the unreality of the a nesting site in birds, for a house in man, etc." concepts used is apparent to any student of human It is difficult to see what valid explanatory purpose behavior, although it may not be so to one of bird can be served by such an inappropriate juxta- behavior-a fact which itself indicates that the position, based on the mere fact that the outcomes source of the unreality lies partly in Lorenz's are similar from the human point of vim. merging of different levels on the basis of super- Tinbergen (1942), speaking of "instinctive"acts ficial similarities. which appear without external stimulation, as the Many other examples of Lorenz's interpretation result of extreme lowering of the threshold because of human behavior could be cited. For example of long non-elicitation, says, "Lorenz . . . dis- (Lorenz, 1940), he interprets the relative attractive- covered that various activities may occur in cases, ness to women of several breeds of dog in terms of where neither proprioceptive stimuli nor hormones the degree to which they fit the innate perceptual could possibly be the driving causes. The simplest pattern releasing instinctive maternal behavior in instance of this kind of vacuum activity is the the human individual! This, again, is entirely hunting behavior of the well-fed dog. As every derived from a too facile analogy with less complex dog-owner knows, a dog can by no means be pre- kinds of animals. Recent work with chimpanzees vented from making hunting excursions by supply- reared in darkness (Riesen, 1947) and with con- ing it with ample food. Other instances of a similar genitally blind human beings whose sight had just kind are familiar to us by introspection. Sports, been restored by surgery (Senden, 1932) indicates 354 THE QUARTERLY REVIEW OP BIOLOGY

that, at least at these phyletic levels, any response animal is set into activity. He says (1937b), "It is to or perception of visual shapes, proportions, sizes, one of the most important and most remarkable and relationships can only occur as the result of a features of the instinctive act, that the organism long and complex learning process. Under these does not wait passively for its release, but actively circumstances it is most difficult to assign any seeks these stimuli." This active seeking is called meaning whatever to Lorenz's assertion that these "appetitive behavior" (Craig, 1918). It may range responses are based innately on perceptual charac- from simple turning movements ("taxes") to the teristics of shape and proportion (Hebb, 1949). most complicated kinds of intelligent behavior. As [Lower mammals apparently require less learning has already been pointed out, Lorenz regards these for the establishment of some of their characteristic as being continuous with each other, both being modes of response to the visual field (Hebb, 1937a, (at different stages of evolutionary development) 1937b; Lashley and Russell, 1934).] means of bringing the animal into a situation con- The interpretation of human behavior in terms taining the stimuli which will release an instinctive of physiological theory based on lower levels is act (viz., eating, copulation, etc.). carried one step further when Lorenz (1940) Lorenz recognizes a few instances of motivation equates the effects of civilization in human beings the source of which is peripheral (viz., hunger, with the effects of domestication in animals. He defecation, etc.). But he adopts the characteristic states that a major effect is the involution or procedure of lumping together all "goal-directed" degeneration of species-specific behavior patterns (i.e., adaptive) behavior which does not have an and releaser mechanisms because of degenerative immediately obvious peripheral motivation under mutations, which under conditions of domestica- the rubric of "appetitive behavior." His conception tion or civilization are not eliminated by natural of "peripheral sources of motivation" is practically selection. He presents this as a scientific reason for limited to the examples just given. I do not recall societies to erect social prohibitions to take the any reference in any of the writings of either Lorenz place of the degenerated releaser mechanisms which or Tinbergen to the autonomic nervous system, or originally kept races from interbreeding. This is to the possibility of qualitatively different roles presented by Lorenz in the context of a discussion of the autonomic nervous system at different of the scientific justification for the then existing phyletic levels, or to the possibility of complex (1940) German legal restrictions against marriage peripheral changes caused by hormones, as sources between Germans and non-Germans. of motivation (cf. Beach, 1948). One result is that The directness of Lorenz's application of the the referring of motivation to the action-specific concept of innate releasers to human social rela- "centers" in the central nervous system is often tions may be gaged by the following quotation like the concept of "innate behavior" itself, simply (1940): "The face of an Asiatic is enigmatic to us a substitute for actual analysis of the biology of the because the physiognomic characteristics to which specific case. For example, Tinbergen (1951) says our innate perceptual patterns respond are not that injection of prolactin into a dove has two connected with the same behavioral characteristics effects: (1) it causes development of the crop as in our race. . . . In all likelihood, this function gland, and (2) it causes brooding behavior. It is (of recognizing facial characteristics) cannot be thus more or less taken for granted that the be- substituted for by experience, as has been deter- havioral effect of the hormone is somehow a mined by many people who are acquainted with specific one; developmental analysis of relation- foreign races." ships between broodiness and crop-gland or brood- Social psychologists will all agree that the vari- patch stimulation is excluded by the nature of the ous degrees of difficulty which different people instinct theory, and of the consequent theory of have in learning to recognize and respond to facial motivation. In the case of "dominance" behavior expressions in a culture different from their own is in the chimpanzee Birch and Clark (1950) have at least partly dependent upon the attitude with shown that behavioral effects of hormones may which they approach the strange culture to be- actually be mediated by peripheral structures in gin with. situations where it is not at all apparent a priori that "proprioceptive stimuli" can play a role. The Sources oj Motivatim Lorenz regards all purposive (adaptive) behavior Lorenz states that as the level of action-specific as being directed toward the performance of the energy in the central nervous system rises the instinctive acts. For Lorenz, it is the perjornacance of ON LORENZ'S THEORY OF INSTINCTIVE BEHAVIOR 355 the instinctive act itself which serves as the animal's normally lies with its eyes closed, and opens them goal. Thus he says (1937a): only when it is gaping. Tinbergen and Kuenen state that the bird will gape in response to a moving ". . .in a man working with the motive of getting food, the behavior directed toward this goal includes visual stimulus as soon as the eyes open, and that many of the higher psychic performances of which he the innate releasing pattern for gaping therefore is capable; the 'motive' (goal)-the instinctive act of 'biting and chewing1-has become drawn back to the includes visual stimulation. For the first day or so end of a long series of acts, without, however, thereby after the eyes open gaping is not directed toward a in any way denying its fundamentally instinctive visual stimulus; even though the stimulus will nature." elicit gaping, the gaping is still directed vertically Thus to Lorenz, the statement "man works in upward. However, after about one day the gaping order to be able to have released [by food] the in- begins to be directed toward certain defined parts stinctive act of biting and chewing" is the same of the visual stimulus (highest, nearest, break in kind of statement as "the frog turns to the right outline, etc.). in order to be able to have released [by the sight Tinbergen and Kuenen's conclusions are that the of a fly in front of him] theinstinctiveact of flipping "centrally-coordinated" instinctive act and the out his tongue." He regards these two goal-directed (continuously-directed) taxis mature at different behavior patterns as being (in the evolutionary rates, the taxis not maturing until 10 days or so sense) continuous with each other, and both as of age while the instinct is fully mature at hatching. having the same kind of relationship to the in- In addition, they conclude that the adequate stinctive act which is the end-member of the stimulus-situations for releasing and for direction behavior chain. However, such a formulation is of the act are different, and are both innate. misleading and of little heuristic value. The actual First, a word about the "maturation of the complexity and variety, and situational relevance, taxis." It is not clear why the animal's experience of the sources of human motivation make such during the first day after its eyes open is not an statements meaningless, not merely because adequate reason for its development of orientation human motivation is more complicated than that toward the visual stimulus. Tinbergen and Kuenen of the frog, but because it is qualitatively different maintain in their discussion that some of the in organization and development. specific features of the stimulus toward which Tinbergen's equation of the causes of sports orientation occurs are not learned by direct ex- and scientific activity with those of hunting in the perience. However, it is not clear that the orienta- dog, because both appear to be internal and tio~toward the visual stimulus is not a result of "self-exciting," is perhaps an extreme example of experience. Even their limited discussion of pos- the result of analogical methods of approach, and sible learning is based on inferences from incidental of the belief that every behavior must have some observations, indicating that Tinbergen and center isomorphically corresponding to it in the Kuenen's orientation toward Lorenz's theory led nervous system. them to discount the serious possibility of learning being involved. HYPOTHESIS AND OBSERVATION The "innateness" of response to moving visual stimuli is quite ambiguous. It will be recalled that The "Innate Releasing Mechanism" the birds lie with their eyes closed for much of the It may be instructive to examine some ways time just after their eyes have first opened. I in which Lorenz's theoretical approach is expressed quote from Tinbergen and Kuenen's protocols: in an investigation of behavior. "5/26/36. Black Thrush, 9-10 days. Lifting and Tinbergen and Kuenen (1939) studied the moving back and forth of the wooden covers evokes no stimulus situations eliciting and directing the reaction. They gape immediately to a tap on the nest. gaping (food-begging) movements of young When the gaping subsides, we move our hand back and forth over the nest, and the birds instantly beg strongly. thrushes. The gaping movement consists at first "Later on the same day: Tap on the closed box of vertically directed stretching of the neck, and releasing gaping. Subsequent lifting of the cover does opening of the mouth. The birds are blind at not; the eyes are closed. Tap causes gaping. After cessation of the reaction, the animals remain lying hatching; their eyes do not open until about 9-10 with open eyes. We hold over the nest, one after the days of age. During this first blind phase gaping other: a black disc, a white wooden rod. . .and a black wand. . . . All the objects are reacted to by violent can be most easily elicited by tapping or jarring gaping. the substrate. When the eyes first open the bird "5/10/36. Song Thrush, 10 days. Preliminary lifting 356 THE QUARTERLY REVIEW OF BIOLOGY of the cover causes no reaction; light tap does. After a priori convinced that the developmental process the birds begin to calm down, a finger shown over the nest immediately releases gaping." is a maturational one, and that they therefore do not have to analyze its conditions. The Lorenz It is apparent that normally the first times that theoretical approach tends to restrict the recogni- young thrushes see visual stimuli they are already in tion of significant details and to obscure possibly a state of gaping excitement, since at first their eyes relevant features of developmental processes. are open only when they gape. Many cases of "innate releasing mechanisms" I have verified these findings on young red- seem to suffer from a similar approach. It will be winged blackbirds (Agelaius phoeniceus), on which recalled that the innate releasing mechanism is I could repeat all of Tinbergen and Kuenen's ob- regarded as a "preformed neural mechanism" servations. However, I was able by watching the (Lorenz and Tinbergen, 1938) for the release of the birds for several consecutive hours to note several instinctive act. Tinbergen refers to the releasing occasions on which one or another of the birds was stimuli as "sign stimuli" because they "represent" lying quietly with its eyes open, when it had not the biologically appropriate object of the instinc- recently gaped. Such birds would not gape in re- tive act. One might ask "Sign of what? Sign to sponse to a moving finger above the head, al- whom?" There is a subtle anthropomorphism though they might move their heads to fixate about the concept of innate releasing mechanisms the finger. If I tapped the nest, thus causing gaping, which is not at first apparent. For example, Lorenz and then moved my finger over the birds when and Tinbergen (1938), in discussing the egg-rolling gaping was subsiding, or shortly thereafter, the of the gray goose, speak of an innate releasing bird would gape instantly and vigorously. What pattern corresponding to the situation "egg out- is meant then by the statement that the birds side the nest." Now, "egg outside the nest" is not gape "innately" to visual stimulation? It would be the perceptual situation to the bird-it is the per- easy to produce ad hoc assumptions about tem- ceptual situation to the human observer. When porary changes in threshold of the innate releasing Lorenz and Tinbergen investigate the effective mechanism as a result of tactual stimulation. But features of the situation, they are looking for a these must be recognized as ad hoc. The possibility pattern of stimuli corresponding to a "patternJ' should be recognized that any stimulation to which which they presume to exist in the central nervous the bird is sensitive will increase the activity of system. Consequently, there is never any analysis the bird when it is already gaping, and may be- of any possible specific relationships between come associated with gaping, so that the later- effective stimuli and the structure or physiology apparent specificity of response to visual stimuli of the organism concerned. Thus the described may be a consequence, not of innate connections, stimulus-situations become structured in human but of the conditions under which visual sensi- terms (bird of prey, vegetation, the parent's head, tivity normally first becomes possible. In addition, etc.) instead of in terms indicative of the problems these birds must be fed almost every hour, and the of specific relationships between the structure and possible relevance of association of visual stimula- function of the animal being investigated. This tion with food reinforcement should not be over- approach, again, derives from Lorenz's identifica- looked. In this connection we may note the experi- tion of every behavior pattern with a specific ments of Padilla (1930), who found that chicks hypothetical "center," rather than with the that were kept in the dark and force-fed for the coming into play of specific structural-functional first twelve days of life, so that they had no op- relationships, which may be very different in portunity to associate pecking behavior with visual different kinds of organisms and for different be- stimuli or with food, would when placed in a normal havior patterns. feeding situation starve to death without ever giv- For example, Tinbergen (1948b) says: "The ing any sign of the allegedly "innate" pecking escape reactions of many birds from passing birds behavior. of prey are a response to a type of movement and It should be noted that the conditional nature to a special characteristic of shape, namely, 'short of the effectiveness of visual stimuli is indicated by neck.' " Now, it is certainly true that many birds Tinbergen and Kuenen's own protocols, but that perform "escape" movements at the sight of a evidently these authors have not really considered "short-necked" bird flying overhead (Kratzig, these facts. This, I think, is because they are 1940). But Tinbergen says 'Lshort-necked" rather ON LORENZ'S THEORY OF INSTINCTIVE BEHAVIOR 357

than, for example, LLhavinga short and long pro- when her eggs are dyed . .. but is. .. disturbed if their. . . contour is altered by stickirg on a bit of clay jection at opposite ends, and moving so that the or putty. . . smoothness of outline is the essential short projection is anterior." His usage is, of course, character of the egg. This is the sort of property that more convenient, as he makes clear, B~~ in addi- can be most easily interpreted in terms of the inherent tendencies to functional organization in the nervous tion it derives from, and in turn reinforces, the network, Lorenzian notion that the "short-neckedness" is a "I do not mean to imply by this that the geometry of the web of the spider is exactly represented in the perceptual LLsign,,or stimulus77(Tinbergen, spider's brain. . .. The angle of radii may be deter- 1939) which corresponds innately to a preformed mined by the angle at which the legs are held (Peters, neural 1lreleasing" mechanism. ~ ~ of leading~ t 1937);~ the completeness~ d or incompleteness of the orb may depend upon the readiness with which certain to an analysis of the specific patterns of excita- postures are assumed in relation to gravity. . . . The tion of sensory elements in the bird's eye which simple nest of the rat is piled and pushed about until it are required for the elicitation of the response, satisfies certain sensory requirements of reduced heat loss. The orb of the spider is perhaps a composite of and a further consideration of the effect of such such sensory requirements combined with some patterning on central nervous activity in the specialized geometrical perceptions such as are illus- nervous systems of these birds, Tinbergen's trated by the rat's more ready recognition of a.. . circle than of irregular ink blots." terminology requires the identification of the bird's readiness to perform "escape movements" 1, contrast to this Lorenz (193j) has with a preformed "conception" of the short- used the analogy of a key unlocking a lock, to necked character of hawks. Thus, 'Ithe partridge describe the function of the releasing pattern in runs for 'Over from an Overhead Object with a releasing an instinctive act. To pursue this analogy, short neck," and "the goose back to the nest Lashley would regard it as the task of the lock- an object lying near the nest which 's smooth- smith-investigator to investigate all the char- contoured and hard-surfaced," become not de$ni- acteristics of lockopening devices, including keys, tions of the problem of how the structure of the pic~ocks, and any other means of opening the various birds makes it possible for them to react locks; and to consider these characteristics together to their environment, but rather solutions to the with what he knows of the structure of locks, the question: "What are the characteristics of these conditions of their use, their history, etc., in order two maIbers of the class of innate releasing to gain an understanding of how the functions of mechanisms?" The assumption Of the various kinds of locks are related to their strut- Lorenz's approach 's that these two Of be- ture. To Lorenz on the other hand, all the locks havior are related to environmental stimuli by are basically alike, so that investigations of the means of mechanisms that are basically identical characteristics of the keys required to open them except for the perceptual details When reveals nothing about internal differences among extended (as it is) to the whole animal kingdom, lock mechanisms, but only about the specific ar- this approach becomes profoundly anti-evolution- rangement of tumblers in lock. ary, in spite of Lorenz's concern with "cornpara- ~11of this should not be taken to mean that tive" studies. we do not recognize that relatively simple stimuli Lashley (1949) has noted with some may sometimes lead to the appearance of quite Lorenz's studies Of patterns. For ex- complex behavior. As a matter of fact, some of the ample he has said: best studies of stimulus-conditions eliciting various "A study of complex instincts requires a detailed types of animal behavior have been carried out by analysis of the exact stimulus or combination of stimuli Lorenz and Tinbergen and their associates (e,g. which call forth the behavior, combined with descrip- tions of the behavior elicited. This has been attempted Tinbergen and Perdick, 19j0). The point is not under controlled conditions only for some instinctive to deny the existence of simple stimuli which under behavior of birds (Lorenz, 1935)." some conditions lead to complex behavior. Rather, However, a closer examination of Lashley's it is that the assigning of the locus of activity to concepts and those of Lorenz will show that the a hypothetical center in the brain, with characteris- subsequent development of Lorenz's approach was tics predeterminedly and isomorphically correspond- not at all in the direction anticipated by Lashle~. ing to those of the stimulus situation, represents an Lashley (1949) says: unphysiological way of thinking disguised in "The nesting tern seems to notice no difference physiological terms. 358 THE QUARTERLY REVIEW OF BIOLOGY

"Vacuzsm Activities" is no "reaction-specific energy" being built up. The The so-calked "vacuum activities" or Leer- periodicity is a result of the periodicity of the laufreaktionen are regarded by Lorenz and Tin- queen's egg-laying, which is not a "center" having bergen as evidence of the accumulation of any characteristics corresponding to the behavior. And even this is not a direct relationship. If the reaction-specific energy in the instinctive center until it "forces" its way through the inhibiting number of larvae in a colony is experimentally innate releasing mechanism and "goes off" without reduced by 50 per cent, thus reducing their total any detectable external stimulus. stimulating effect, a normal nomadic phase cannot Lees (1949) has cited the example of the cyclical occur. Recent findings (Schneirla and Brown, colony activities of the ant Eciton hamaturn 1950) have in fact confirmed the hypothesis that each of the regular large-scale egg-delivering (Schneirla, 1938) as an example of "something akin to 'vacuum activity.' " Colonies of this army episodes in the queen's function basic to the cycle ant pass regularly through statary and nomadic is a specific outcome of her over-feeding, due to a phases, each lasting about 20 days. As Lees points maximal stimulation of the colony by the brood. out [based on Schneirla's (1944) description]: This event, occurring inevitably at the end of each nomadic phase, is a "feed-back" type of function, "During the statary phase the bivouac, to which the not at 211 related to the implications of "vacuum single queen is confined, remains in situ and raiding activities are minimal. During the nomadic phase the activity." position of the bivouac is changed each nightfall and The restrictive nature of such categorical strong raiding parties emerge from the colony. This activity is in no way related to the abundance or theories as that of Lorenz is very well illustrated scarcity of food in the neighborhood.. .." by Lees' remarks on Eciton. The actual develop- ment process leading to the periodic performances This cyclic behavior thus appears to Lees to of this ant are well understood, and are known to have the character of a "vacuum activity," in have no essential relationship to any "reaction- that it occurs periodically without any noticeable specific energy" in any nervous system; further change in the external stimulus-conditions. This is they are known not to be "innate" as such very misleading, for Schneirla's (1938, 1944) (Schneirla, 1938). The processes leading to this analysis of this behavior has shown that the behavior surely have nothing to do with the change from statary to nomadic behavior is a con- processes leading to "vacuum activities" in a fish. sequence of the growth of a great new brood of Yet the superficial similarity is sufficient to cause ants. When the callow workers emerge from their Lees to cite the ant's behavior as an example of a cocoons, their movements stimulate the adult type of behavior described for vertebrates. This is workers to great activity. As the callows mature a good example of the tendency encouraged by and cease to be dependent on the adults, their such theories to look for cases fitting the theoretical energizing effect is lessened. At this point, the categories in many types of behavior, rather than emergence of wriggling larvae from the eggs supple- analysis of the processes involved in the develop- ments the diminishing activating effect of the ment of any one behavior pattern. callows on the adults. When the larvae pupate, and become inactive, the adults are no longer CONCLUSION subject to trophallactic (Wheeler, 1928) stimula- tion, and the colony changes to its statary period. We have summarized the main points of Lorenz's The point that is relevant to our discussion is instinct theory, and have subjected it to a critical that Schneirla's analysis leads to a conception that examination. We find the following serious flaws: is the opposite of that implied by the notion of 1. It is rigidly canalized by the merging of "vacuum activity." The periodic recurrences are widely different kinds of organization under in- not the result of the building up of energy in any appropriate and gratuitous categories. animal's nervous system. They are the result of 2. It involves preconceived and rigid ideas of the periodic recurrences of inter-individual stimu- innateness and the nature of maturation. lating effects. The behavior is not represented 3. It habitually depends on the transference of "in advance" in any of the animals in the colony; concepts from one level to another, solely on the it emerges in the course of the ants' relationships basis of analogical reasoning. with one another and with the environment. There 4. It is limited by preconceptions of isomorphic ON LORENZ'S THEORY OF INSTINCTIVE BEHAVIOR 359 resemblances between neural and behavioral underlie the development of "instinctive" phenomena. behavior. 5. It depends on finalistic, preformationist con- ceptions of the development of behavior itself. ACKNOWLEDGMENTS 6. As indicated by its applications to human I am greatly indebted to Dr. T. C. Schneirla (who psychology and sociology, it leads to, or depends originally suggested the writing of this paper) and to Dr. J. Rosenblatt for many stimulating and helpful on, (or both), a rigid, preformationist, categorical discussions of the problems discussed here. Dr. Schneirla conception of development and organization. in particular has devoted much attention to criticism Any instinct theory which regards "instinct" of the paper at various stages. The following people also have read the paper, in as immanent, preformed, inherited, or based on part and at various stages, and have made many helpful specific neural structures is bound to divert the suggestions and comments: Drs. H. G. Birch, K. S. Lashley, D. Hebb, H. Kliiver, L. Aronson, J. E. investigation of behavior development from Barmack, L. H. Hyman, L. H. Lanier, and G. Murphy. fundamental analysis and the study of develop- Since these scientists differ widely in the extent of their mental problems. Any such theory of "instinct" agreement or disagreement with various points of my discussion, I must emphasize that none of them is in inevitably tends to short-circuit the scientist's any way responsible for any errors of omission or com- investigation of intraorganic and organism-en- mission that may appear. vironment developmental relationships which Present address: Rutgers University, Newark 2, N. J.

LIST OF LITERATURE ADRIAANSE,M. S. C. 1947. Antmophila campestris -. 1951b. Body chemistry and perception. In Latr. und Ammophila adriaansei Wilcke. Ein Perception: an Approach to Persoltality (R. R. Beitrag zur vergleichenden Verhaltensforschung. Blake and G. V. Ramsey, eds.), pp. 56-94. Ron- Behaviour, 1: 1-34. ald Press, New York. ANASTASI,A., and J. P. FOLEY,JR. 1948. A pro- BIRCH, H. G., and G. CLARK. 1946. Hormonal posed reorientation in the heredity-environment modification of social behavior. 11. The effects of controversy. Psychol. Rev., 55: 239-249. sex-hormone administration on the social domi- ARMSTRONG,E. A. 1947. Bird Display and Behaviour. nance status of the female-castrate chimpanzee. Lindsay Drummond & Co., London. Psychosom. Med., 8: 320-331. -. 1950. The nature and function of displace- -, and -. 1950. Hormonal modification of ment activities. Symp. Soc. exp. Biol., 4: 361-384. social behavior. IV. The mechanism of estrogen- BAERENDS,G.P. 1941. Fortpflanzungsverhalten und induced dominance in chimpanzees. J. comp. Orientierung der Grabwespe Ammophila campestris physiol. Psychol., 43: 181-193. Jur. Tijdschr. Ent., 84: 68-275. BIRD, C. 1925. The relative importance of matura- -. 1950. Specializations in organs and move- tion and habit in the development of an instinct. ments with a releasing function. Symp. Soc. Pedagog. Semin., 32: 68-91. cxp. Biol., 4: 337-360. -. 1933. Maturation and practise; their effects -, and J. M. BAERENDS-VANROON. 1950. An upon the feeding reaction of chicks. J. comp. introduction to the study of the ethology of cichlid Psychol., 16: 343-366. fishes. Behaviour, suppl. 1: 1-242. BREED, F. 1911. The development of certain in- BEACH, F. A. 1942. Analysis of factors involved stincts and habits in chicks. Behav. Monogr., 1: in the arousal, maintenance and manifestation 1-78. of sexual excitement in male animals. Psychosom. BROWMAN,L. G. 1943. The effect of controlled ,Wed., 4: 173-198. temperatures upon the spontaneous activity -. 1947a. Evolutionary changes in the physio- rhythms of the albino rat. J. exp. Zool., 94: 477- logical control of mating behavior in mammals. 489. Psychol. Ra., 54: 297-315. CARMICHAEL,L. 1936. A re-evaluation of the con- -. 1947b. A review of physiological and psycho- cepts of maturation and learning as applied to logical studies of sexual behavior in mammals. the early development of behavior. Psychol. Physiol. Ra., 27: 240-307. Rev., 43: 450-470. -. 1948. Hormones and Behavior. Hoeber, New -. 1941. The experimental embryology of mind. York. Psychol. Bdl., 38: 1-28. -. 1951a. Instinctive behavior: reproductive ac- -. 1947. The growth of the sensory control of tivities. In Handbook of Experimental Psychology behavior before birth. Psychol. Rep.,54: 316324. (S. S. Stevens, ed.) pp. 387-434. John Wiley & CHEIN, I. 1936. The problems of heredity and en- Sons, New York. vironment. J. Psychol., 2: 229-244. 360 THE QUARTERLY REVIEW OF BIOLOGY

CLARK,G. 1948. The mode of representation in Wirbeltieren. S. Ges. naturf. Fr., Berl., 1929: the motor cortex. Brain, 71: 320-331. 333-342. -, and J. W. WARD. 1948. Responses elicited HEPPEL,L. A., and C. L. A. SCHMIDT. 1938. Studies from the cortex of monkeys by electrical stimula- on the potassium metabolism of the rat during tion through fixed electrodes. Brain, 71: 332- pregnancy, lactation and growth. Univ. Cdq. 342. Publ. Physiol., 8: 189-205. COGHILL,G. E. 1929. Anatomy and the Problem of HESS, W. R. 1949. Das Zwischenhirn. Syndrome Behavior. Cambridge University Press, London. Lokdisationen, Funktionen. Benno Schwabe & CRAIG, W. 1918. Appetites and aversions as con- Co., Basel. stituents of instincts. Biol. Bull., Woods Hole, -, and M. BRWGGER.1943. Das subkortikale 34: 91-107. Zentmm der affektiven Abwehrreaktion. Helv. CRUZE, W. W. 1935. Maturation and learning in physiol. acta, 1: 33-52. chicks. J. comp. Psychol., 19: 371-409. HINES, M. 1947. The motor areas. Fed. Proc., 6: DELACOUR,J., and E. MAYR. 1945. The family 441-447. Anatidae. Wilson Bull., 57: 3-55. HOLST, E. VON. 1935. Alles oder Nichts, Block, FULTON,J. F. 1949. Physiology of the Nervous Alternans, Bigemini und verwandte Phlnomene System, 3d ed. Oxford University Press, New als Eigenschaften des Ruckenmarks. PfEiig. York. Arch. ges. Physiol., 236: 149-159. Gos, E. 1933. Les reflexes conditionnels chez l'em- -. 1936a. Versuchen zur relativen Koordination. bryon d'oiseau. Bull. Soc. Sci. Liege, No. 45: Pjliig. Arch. ges. Physiol., 237: 93-122. 194-199; NO. 67: 246-250. -. 1936b. Vom Dualismus der motorischen und GRAY, J. 1939. Aspects of animal locomotion. der automatischrhythmischer Funktion im Rucken- Proc. roy. Soc., B, 128: 28-62. mark und vom Wesen des automatischen Rhyth- -. 1950. The role of peripheral sense organs mus. PjQ. Arch. ges. Physiol., 237: 356-378. during locomotion in the vertebrates. Symp. -. 1936c. Ueber dem "Magnet-Effekt" als Soc. exp. Biol., 4: 112-126. koordinierende Prinzip 'im Riickenmark. PRiig. -, and H. W. LISSMANN. 1940. The effect of Arch. ges. Physiol., 237: 655482. deafferentation upon the locomotory activity of --. 1937. Regulationsfahigkeit im Zentralnerven- amphibian limbs. J. exp. Biol., 17: 227-236. system. Nalurwissenschaften, 25: 625631, 641- -, and ---. 1946a. Further observations on the 647. effect of deafferentation on the locomotory ac- HOWELLS,T. H. 1945. The obsolete dogmas of tivity of amphibian limbs. J. exp. Biol., 23: heredity. Psychol. Rev., 52 : 23-34. 121-132. -, and D. 0.VINE. 1940. The innate differential -, and -. 1946b. The co-ordination of limb in social learning. J. abnorm. (soc.) Psychol., 35: movements in the amphibia. J. exp. Biol., 23: 537-548. 133-142. HUNT, E. L. 1949. Establishment of conditioned GROHMANN,J. 1938. Modifikation oder Funktions- responses in chick embryos. J. comp. physiol. reifung? Ein Beitrag zur Klarung der wechsel- Psychol., 42: 107-117. seitigen Beziehungen zwischen Instinkhandlung JENNINGS,H. S. 1930. The Biological Basis om- und Erfahrung. Z. Tierpsychol., 2: 132-144. man Nature. Norton & Co., New York. HARTLEY,P. H. T. 1950. An experimental analysis KEELER,C. 1931. The Laboratory Mouse. Harvard of interspecific recognition. Symp. Soc. exp. Biol., University $Press, Cambridge. 4: 313-336. KINDER,E. F. 1927. A study of the nest-building HEBB, D. 0. 1937a. The innate organization of activity of the albino rat. J. exp. Zool., 47: 117- visual activity: I. Perception of figures by rats 161. reared in total darkness. J. genet. Psychol., 51: KOEHLER,0. 1950. Die Analyse der Taxisanteile 101-126. instinktartigen Verhaltens. Symp. Soc. exp. -. 1937b. The innate organization of visual Biol., 4: 269-303. activity: 11. Transfer of response in the discrimina- KRAMER,G. 1937. Beobachtungen uber Paamngs- tion of brightness and size by rats reared in total biologie und soziales Verhalten von Maueredechsen. darkness. J. comp. Psychol., 24: 277-299. 2.Morph. Okol. Tiere, 32: 752-784. -. 1949. The Organization of Behavior. J. S. KRXTZIG, H. 1940. Untersuchungen zur Lebens- Wiley & Sons, New York. weise des Moorschneehuhns, Lagopu~I. lagopus, HEINROTH,0. 1910. Beitrage zur Biologie, nament- wahrend der Jugendentwicklung. J. Orn., Lpz., lich Ethologie und Psychologie der Anatiden. 88: 139-166. Znt. orn. Congr., 5 (Berlin): 589-702. KWHN, A. 1919. Die Orientierung der Tiere in -. 1930. Ueber bestimmte Bewegungsweisen bei Raum. G. Fischer, Jena. ON LORENZ'S THEORY OF INSTINCTIVE BEHAVIOR 361

Kuo, Z. Y. 1932a. Ontogeny of embryonic behavior Anatinen. J. Orn., Lpz., 89 (Sonderheft): 194- in Aves. I. The chronology and general nature 294. of the behavior of the chick embryo. J. exp. -. 1950. The comparative method in studying zool., 61: 395-430. innate behavior patterns. Symp. Soc. exp. Biol., -. 1932b. Ontogeny of embryonic behavior in 4: 221-268. Aves. 11. The mechanical factors in the various -, and N. TINBERGEN.1938. Taxis und In- stages leading to hatching. J. exp. Zool., 62: stinkthandlung in der Eirollbewegung der Grau- 453489. gans. I. Z. Tierpsychol., 2: 1-29. -. 1932c. Ontogeny of embryonic behavior in MACDOUGALL,WILLIAM. 1923. An Outline of Psy- Aves. 111. The structure and environmental fac- chology. Scribner's, New York. tors in embryonic behavior. J. comp. Psychol., -. 1930. The Hormic Psychology. In Psy- 13: 245-272. chologies of 1930 (Carl Murchison, ed.), pp. 3-36. -. 1932d. Ontogeny of embryonic behavior in Clark Univ. Press, Wooster, Mass. Aves. IV. The influence of embryonic movements MAIER, N. R. F., and T. C. SCHNELRLA.1935. upon the behavior after hatching. J. comp. Principles of Animal P~ychology. McGraw-Hill Psychol., 14: 109-122. Co., New York. LASHLEY,K. S. 1923. Temporal variation in the MAST, S. 0. 1926. Structure, movement, locomo- function of the gyrus precentralis in primates. tion and stimulation in amoeba. J. Morph., 41: Amer. J. Physiol., 65: 585402. 347425. -. 1938. Experimental analysis of instinctive MILLER, N. E. 1951. Learnable drives and rewards. behavior. Psychol. Rev., 45 : 445471. In Handbook of Experimental Psychology (S. S. -. 1942. The problem of cerebral organization in Stevens, ed.), pp. 435472. John Wiley & Sons, vision. Biol. Symp., 7: 301-322. New York. -. 1949. Persistent problems in the evolution of MORGAN, C. T. 1947. The hoarding instinct. mind. Quart Rev. Biol., 24: 2842. Psycl~ol.Rev.,54: 335-341. -, and J. T. RUSSELL. 1934. The Mechanism NISSEN,H. W. 1951. Phylogenetic comparison. In of Vision: XI. A preliminary test of innate or- Handbook of Experimental Psychology (S. S. ganization. J. genet. Psychol., 45: 136144. Stevens, ed.), pp. 347-386. John Wiley & Sons, LEES, A. D. 1949. Modern concepts of instinctive New York. behaviour (in section on "Entomology"). Sci. NOBLE, G. K., and H. T. BRADLEY.1933. The Prog. Twent. Cent., 37: 318-321. mating behavior of lizards. Ann. N. Y. Acad. LISSMANN,H. W. 1946a. The neurological basis of Sci., 35: 25-100. the locomotory rhythm in the spinal dogfish (Scyl- PADILLA,S. G. 1930. Further studies on the de- lium canicula, Acanthias vulgaris). 1. Reflex be- layed pecking of chicks. J. camp. Psychol., 20: haviour. J. exp. Biol., 23: 143-161. 413-443. -. 1946b. The neurological basis of the locomo- PATRICK,J. R., and R. M. LAUGHLIN. 1934. Is the tory rhythm in the spinal dogfish (Scyllium wall-seeking tendency in the white rat an instinct? canicula, Acanthias adgaris). 11. The effect of J. genet. Psychol., 44:378-389. de-afferentation. J. exp. Biol., 23: 162-176. PETERS,H. 1937. Studien an der Netz der Kreuz- LORENZ,K. 1931. Beitrage zur Ethologie sozialer spinne (Aranea diadema L.). Z. ~Morph. Okol. Corviden. J. Orn., Lpz., 79: 67-127. Tiere, 1: 126150. -. 1932. Betrachtungen iiber das Erkennen der PETRUNKEVITSCH,A. 1926. The value of instinct as arteigenen Triebhandlungen der Vogel. J. Om. a taxonomic character in spiders. Biol. Bull., Lpz., 80: 50-98. Wood's Hole, 50: 427432. -. 1935. Der Kumpan in der Umwelt des Vogels. RAY, W. S. 1932. A preliminary study of fetal J. Orn., Lpz., 83: 137-213, 289413. conditioning. Child Develpm., 3: 173-177. -. 1937a. Ueber den Begriff der Instinkthand- RIESEN, A. H. 1947. The Development of Visual lung. Folia biotheor., Leiden, 2: 17-50. Perception in Man and Chimpanzee. Science, -. 1937b. Ueber die Bildung des Instinkt- 106: 107-108. begriffes. Nalurwissenschaften, 25: 289-300, 307- RIESS, B. F. 1949a. A new approach to instinct. 318, 324-331. Sci. & Soc., 13: 150-154. -. 1939. Vergleichende Verhaltensforschung. -. 1949b. The isolation of factors of learning and Zool. Anz., 12 (Suppl. band): 69-102. -. 1940. Durch Domestikation verursachte native behavior in field and laboratory studies. Storungen arteigenen Verhaltens. 2. angew. Ann. N. Y. Acad. Sci., 51: 1093-1102. Psychol. Charakterkunde 59: 2-81. ROMER, A. S. 1945. Vertebrate Paleontology. Uni- -. 1941. Vergleichende Bewegungsstudien an versity of Chicago Press, Chicago. 362 THE QUARTERLY 8!,?%VIEWOF BIOLOGY

SCHNEIRLA,T. C. 1938. A theory of army-ant be- SPERRY,R. W. 1945. The problem of central nervous havior based upon the analysis of activities in a reorganization after nerve regeneration and muscle representative species. J. comp. Psychol., 25: transposition. Quurt. Rev. Biol., 20: 311-369. 51-90. STONE,C. P. 1947. Methodological resources for the -. 1939. A theoretical consideration of the basis experimental study of innate behavior as related for approach-withdrawal adjustments in behavior. to environmental factors. Psychol. Rev., 54: 342- Psychol. BuU., 36: 501-502. 347. -. 1941. Social organization in insects, as re- THOWE, W. H. 1948. The modern concept of in- lated to individual function. Psychol. Rev., 48: stinctive behaviour. Bull. Anim. Behav., 7: 1-12. 465-486. -, AND F. G. W. JONES. 1937. Olfactory condi- -. 1944. The reproductive functions of the army- tioning in a parasitic insect and its relation to the ant queen as pacemakers of the group behavior problem of host selection. Proc. roy. Soc., B, pattern. J.N.Y. ent. Soc., 52: 153-192. 124: 5681. -. 1945. Contemporary American animal psy- TINBERGEN,L. 1939. Zur Fortpflanzungsethologie chology in perspective. In Twentieth Century von Sepia ojicinalis L. Arch. nterl. Zool., 3: Psychology. (P. Harriman, ed.) pp. 306-316. 305-335. Philosophical Library, New York. TINBEXGEN,N. 1939. On the analysis of social or- -. 1946. Problems in the biopsychology of social ganization among vertebrates, with special ref- organization. J. abnm. (soc.) Psychol., 41: erence to birds. Amer. Midl. Nat., 21: 210-234. 385-402. -. 1942. An objectivistic study of the innate -. 1948. Psychology, Comparative. Article in behaviour of animals. Bibl. biotheor., Leiden, D, Encyclopedia Brittanica. 1: 39-98. -. 1949a. Levels in the psychological capacities -. 1948a. Physiologische Instinktforschung. Ex- of animals. In Philosophy for the Future (R. perientia, 4: 121-133. Sellars and V. J. McGiil, eds.). Macmillan & --. 1948b. Social releasers and the experimental Co., New York. method required for their study. Wilson Bull., -. 1949b. Army-ant life and behavior under dry- 60: 6-51. season conditions. 3. The course of reproduction -. 1950. The hierarchical organization- of nervous and colony behavior. Bull. Amer. Mus. nal. mechanisms underlying instinctive behaviour. Hkt. 94: 1-81. Symp. Soc. exp. Biol., 4: 305-312. -. 1950. The relationship between observation -. 1951. The Study of Instinct. Oxford Uni- and experiment in the field study of behavior. versity Press, Oxford. Ann. N. Y.Acad. Sci., 51: 1022-1044. -, and D. J. KUENEN. 1939. Ueber die auslosen- -, and R. Z. BROWN. 1950. Army-ant life and den und die richtunggebenden Reizsituationen dcr behavior under dry-season conditions. 4. Further Sperrbewegung von jungen Drosseln (Turdus m. investigation of cyclic processes in behavioral and merula L. und T. e, ericetorum Turton). Z. reproductive functions. Bull. Amer. Mus. nut. Tierpsychol., 3: 37-60. Hisl., 95: 263-354. -, B. J. D. MEEUSE,L. K. BOEREMA,and W. W. SCHOOLLAND,J. B. 1942. Are there any innate be- VAROSSIEAU.1942. Die Balz des Samtfalters, havior tendencies? Genet. Psychol. Monogr., 25: Eumenis ( =Satyrus) semele (L.). Z. Tierpsychol. 219-287. 5: 182-226. SENDEN,M. VON. 1934. Raum-und Gestaltauffas- -, and A. C. PERDICK. 1950. On the stimulus sung bei operierten Blindgeborenen vor und nach situation releasing the begging response in the der Operation. Barth, Leipzig. newly hatched Herring Gull chick (Lancs argenta- SHEPARD,J. F., and F. S. BREED. 1913. Matura- tus argentatus Pont.). Behaviour, 3: 1-39. tion and use in the development of an instinct. UYLDERT,I. E. 1946. A conditioned reflex as a J. Anim. Bchav., 3: 274-285. factor influencing the lactation of rats. Acta SINNOTT,E. W., L. C. DUNN,and TH. DOBZHANSKY. brev. neerl. Physiol., 14: 86-89. 1950. Principles of Genetics. McGraw-Hill Co., WARDEN,C. J., T. N. JENKINS,and L. H. WARNER. New York. 1936. Comparative Psychology, Vol. 111. Vertc- SMITH, P. E., and E. C. MACDOWELL.1930. An brutes. Ronald Press, New York. hereditary anterior-pituitary deficiency in the WASHBURN,S. L. 1947. The relation of the temporal mouse. Anat. Rec., 46: 249-257. SMITH, S., and E. R. GUTHRIE. 1921. General muscle to the form of the skull. Anat. Rec., 99. Psychology in Terms of Behavior. Appleton, New 239-248. York. WEISS, P. 1936. Selectivity controlling the central- SPELT,D. K. 1948. The conditioning of the human peripheral relations in the nervous system. Biol. fetus in utero. J. exp. Psychol., 38: 338-346. Rev., 11: 494531. ON LORENZ'S THEORY OF INSTINCTIVE BEHAVIOR 363

-. 1937a. Further experimental investigations on locomotion after the interchange of right and left the phenomenon of homologous response in trans- limbs. J. comp. Newol., 67: 269-315. planted amphibian limbs. I. Functional observa- -. 1941. Self-differentiation of the basic patterns tions. J. comp. Neurol., 66: 181-206. of coordination. Comp. Psychol. Monogr. 17: -. 1937b. Further experimental investigations 1-96. of the phenomenon of homologous response in -, 1950. Experimental analysis of coordination transplanted amphibian limbs. 11. Nerve re- by the disarrangement of central-peripheral rela- generation and the innervation of transplanted tions. Symp. Soc. exp. Biol., 4: 92-111. limbs. J.comp. Neurol., 66: 481-535. WHEELER,W. M. 1928. The Social Insects. Har- -. 1937c. Further experimental investigations court, Brace & Co., New York. of the phenomenon of homologous response in WHITMAN, C. 0. 1899. Animal behavior. Biol. transplanted amphibian limbs. 111. Homologous Lect. mar. biol. Lab. Wood's Holl, 1898: 285-338. response in the absence of sensory innervation. -. 1919. The behavior of pigeons. Pz~bl.Carneg. J. comp. Newol., 66: 537-548. Inst., 257: 1-161. -. 1937d. Further experimental investigations WIESNER,B. P., AND N. M. SIIEARD. 1933. Ma- of the phenomenon of homologous response in ternal Behaviour in the Rat. Oliver & Boyd, Lon- transplanted amphibian limbs. IV. Reverse don.