AniTNlILeaming&: Behavior1977,5 (4), 325-335

The origin and functions of adjunctive behavior

JOHN L. FALKRutgers University, New Brunswick, New Jersey 08903

The major determinants of schedule-induced or adjunctive behavior are reviewed briefly.Adjunctive behavior and its ethological equivalent, displacement activity, has a stabilizingfunction on agonistic, mating, parental, and intermittent-feeding behavior when any of theseactivities are in unstable equilibrium with an escape vector. This buffering action of adjunctivebehavior is analogous to the diversity-stability rule of ecology in which an increase in thediversity of species stabilizes the populations of the component species, thereby preventingextinction. Opposing behavioral vectors in unstable equilibrium can function to exaggeratecertain behavioral adjuncts that preexist in a situation. The resulting increase in diversifica­tion (information] augments the overall stability of the opposing-vector circumstance, con­serving the context. This strengthening process is discussed in relation to ritualization andthe preadaptation of exaggerated behavior to new functions.

The nature and determinants of adjunctive be­havior were described in a previous article (Falk ,1971). Briefly, "adjunctive behavior" refers to be­havior that is maintained at a high probability bystimuli which derive their exaggerated reinforcingefficacy primarily as a function of scheduleparameters governing the availability of another dassof reinforcing events. The present account willattempt to answer the questions of how the ad­junctive determination of behavior may have evolvedand what adaptive functions it might serve. Con­siderable information has accumulated concerningthe variables which determine adjunctive behavior.A few of the major features will be described at theoutset, but the treatment is far from complete. Thisis neither a review nor even a synoptic summary.

Animals (rats, mice, pigeons, squirrel and rhesusmonkeys, chimpanzees) reduced in body weight andgiven small portions of food at certain fixed orirregular times in daily sessions quickly develop ex­cessive and persistent behavioral adjuncts. Forexarnple, rats earning 45-mg food pellets by lever­pressing on a variable-intervall-min schedule or on afixed-interval 90-sec schedule for a few hours perday drink extensive draughts of water after con­suming each pellet (Falk, 1961, 1969). This schedule­induced polydipsia continues daily as lang as thegenerator schedule conditions remain in effect.Approximately 10 times the amount of water isdrunk compared to a condition in which the animalis given the same amount of food daily as a singleportion and remains in the situation for the same

This manuscript was facilitated by support from Grant DA 1110from the National Institute on Drug Abuse, Grant AAOO253 fromthe National Institute of Alcohol Abuse and Alcoholism, and a Bio­logical Seiences support grant (NIH) to Rutgers University. Theauthor's address is: Psychology Building, Rutgers University, BuschCampus, New Brunswick, New Jersey 08903.

number of hours. The animal is never water deprived,yet the intermittent delivery of food increases thereinforcing value of water by a mechanism whichis not at all explieable by metabolie or water balanceconsiderations (Falk, 1969).

The generator conditions produced by intermittentfood schedules can induce exaggerated behaviorsother than polydipsia. Schedule-induced attackagainst a conspecifie, an inanimate object, a con­specific model, or a conspecific visual photo targethave been studied extensively (e.g., Azrin, Hutehinsan,& Hake, 1966; Cohen & Looney, 1973; Flory, 1969;Hutehinsan, Azrin, & Hunt, 1968). Scheduled fooddelivery also can induce escape behavior. Animalswill emit operant behavior which results in the ter­mination of the schedule (time out) or in the removalof stimuli associated with the schedule (escape fromthe discriminative stimuli) (e.g., Azrin, 1961; Brown& Flory, 1972; Thompson, 1964). Schedule-inducedadjunctive wheel running (Levitsky & Collier, 1968),pica (Villarreal, Note 1), pecking (Miller & Gollub,1974), and airstream licking (Mende1son & Chillag,1970) are ot her activities which occur excessivelyand in a fixed temporal relation relative to the induc­ing events (food-portion delivery) of the generatorschedule. However, generator conditions da notnecessarily involve food deprivation and the inter­mittent delivery of small food portions. Recentresearch with rats (King, 1974; Wayner, Singer,Cimino, Stein, & Dwoskin, 1975) and humans(Kachanoff, Leveille, McLelland, & Wayner, 1973;Wallace, Singer, Wayner, & Cook, 1975; Wallace &Singer, 1976) has demonstrated the induction ofvarious adjunctive behaviars in which the generatorconditions constituted the scheduling of water ,running-wheel access, monetary gain, game playing,or problem solving rather than food. Thus, adjunc­tive behavior reveals a considerable degree of

325

326 FALK

generality both in the kinds of commodities or be­havior sequences whose intermittency constitutes agenerator condition and in the variety of resultingadjunctive activities.

While a host of conditions are demonstrablemodulators of adjunctive behavior, two determinantsare of particular importance: the intermittence valueof the generator schedule and the deprivation statusof the scheduled consummatory behavior. When thefirst of these factors, intermittence value, was variedsystematically, a sornewhat counterintuitive func­tion was obtained. Many fixed-interval schedulevalues between 2 and 300 sec were explored for therat. Adjunctive drinking increased as a function ofthe fixed-interval food schedule value up to120-180 sec and decreased to a low level at the300-sec value (Falk, 1966). Subsequent experimentshave confirmed this bitonic function, not only forschedule-induced polydipsia in the rat (e.g., Flory,1971), but also for other adjunctive activities and forspecies other than the rat. For example., abitonicfunction has been reported for schedule-inducedpolydipsia in the rhesus monkey (Allen & Kenshalo,1976) and schedule-induced attack and escape in thepigeon (Brown & Flory, 1972; Flory, 1969).

The other major factor deterrnining the rate ofadjunctive behavior, the deprivation status of thescheduled consummatory behavior, has received lessexperimental attention than the bitonic function, butthe results have been quite consistent. The degree ofschedule-induced polydipsia (Falk, 1969), attack(Dove, 1976), and airstream licking (Chillag &Mendelson, 1971) produced by intermittent foodschedules were all inverse functions of the body­weight level at which animals were maintained asdetermined by the total food rations allowed.

Thus, adjunctive activities are systernatically re­lated to two main constraining conditions. Theactivities are bitonic functions of the generator inter­mittence value and increasing monotonic functionsof the degree of deprivation. They are excessive andpersistent. A behavioral phenomenon which encom­passes many kinds of activities and is widespreadover species and high in predictability ordinarily canbe presumed to be a basic mechanism contributingto adaptation and survival. The puzzle of adjunctivebehavior is that, while fulfilling the above criteria,its adaptive significance has escaped analysis, In­deed, adjunctive activities have appeared not onlycuriously exaggerated and persistent, but alsoenergetically quite costly. They do not have the rela­tion of unconditioned, conditioned or operant re­sponses, or fixed-action patterns to the conditionsgenerating them. Consequently, traditional accountsof behavior afforded by Pavlovian and operant con­ditioning and by ethology seem not to apply. How­ever, similarities between the conditions producingboth adjunctive and dis placement activities have

been suggested (Falk, 1971), and this analysis will beexpanded.

The reader is now requested to bear with me whilethe case for certain ecological and evolutionary anal­ogies is slowly, and I hope not too painfully, built.These will converge later in the paper to providea basis for ascertaining the origin and functionsof the seemingly profligate behaviors termed "ad­junctive. "

TUE ECOLOGICAL ANALOGY:STABILIZING FUNCTION OF NETWORKS

An ecosystem is defined as all the organisms withina given region, their interactions with the physicalenvironment, and, most importantly, their inter­actions among themselves. These latter interactionsconstitute the food web or predator-prey relations.The food web consists of food chains, with eachlink in the chain being a particular predator-preyrelation. The links are located at different trophiclevels. The green plants, dependent upon energyfrom the sun, define the first level, the herbivoresconsuming them define the second level, and thecarnivores living on them compose the third level.Carnivores consuming third-level carnivores wouldconstitute a fourth trophic level. Feeding may spanmore than one trophic level. For example, the omni­vores eat both plant and animal materials.

The food webs of the species populations consti­tuting different ecosystems can take many differentforms. Nevertheless, a roughly acceptable notion,at least as a point of departure, is that the stabilityof an ecosystem increases as the number of foodweb linkages increases. Thus, the greater the numberof plant and animal species participating in an eco­systern, the less each species will fluctuate innumbers. The argument that species communitystability is dependent upon trophic web complexitysterns from several lines of evidence. Both mathe­matical considerations of predator-prey interactions(MacArthur, 1955) and field observations indicatethat a community containing just a single preyspecies and its enemy oscillates widely in numbers(Elton, 1958, p. 146; Odum, 1971, chap. 7). AI­though there is little stability, extinction does notnecessarily occur. The community is subject toperiodic "outbreaks" of eaeh species, especially insimple environments (e.g., small islands, cultivatedmonocultures, arctic regions). The addition of one ormore prey species would diminish such fluctuations.The ease with whieh invading speeies are able tocolonize the relatively simple environments of islandscontaining rather few species further illustrates thevulnerability of simple ecosystems. The harsh condi­tions of the arctie apparently limit the number ofspecies, and stability conditions are correspondinglydelicate (MacArthur, 1955). Such regions are, as

would be expected, characterized by large, cyclicoscillations in their fauna which "may be due in partto the communities not being sufficiently complexto damp out oscillations" (Hutchinson, 1959). Bycontrast, the tropical rain forest with a high degreeof trophic complexity is unusually stable (Elton,1958, pp. 148-149).

As sparse islands are susceptible to species in­vasions and the arctic to population oscillations oflarge magnitude, the artificially simplified cropcommunities planted in large stands by man are alsoextremely vulnerable. Monoculture reduces thediversity of flora, and such special crops usuallylack the development of typical attached fauna.Generalized pesticides further reduce both herbivoreand carnivore species diversities, sometimes quitedrastically, thereby limiting pest enemies along withthe pests. Thus, pest outbreaks are frequent due tothe extreme simplicity of the monoculture. Also, pestenemies fail to proliferate as these parasites andpredators often require nourishment derived fromthe secretions of plants absent from monocultures.In contrast with this picture, pest outbreaks do notoccur in tropical forests where the great diversity ofplant and animal species constitutes a highly complexand stable comrnunity. Likewise, the intercroppingof a great diversity of plants practiced by the swiddenagriculturalists of Malaysia and Indonesia imitates thediversity and stability of the jungle which it tem­porarily replaces (Geertz, 1963, chap. 2).

Community stability, then, is a function of thenumber of links among species in the food web, orstated another way, "the amount of choice of theenergy in going through the web measures the sta­bility" (MacArthur, 1955). This can be viewed interms of what occurs when one species increases ordecreases. As it increases, competitors on the sametrophic level may decrease, and predators andparasites may switch to the more abundant prey.Conversely, a decrease in the species would producea switch in preference by predators to less rare,alternate prey. The more diverse and complex eco­systems may be more stable since the individualspecies populations would be protected from thepossibility of large fluctuations by the bufferingaction of other species groups. This relation is some­times referred to as the diversity-stability rule (Wilson& Bossert, 1971, p. 140). The rule was formalizedby MacArthur (1955), who applied the Shannon­Wiener formula for the measurement of informationin a system 10 the analysis of ecological communities.The diversity or stability of a community is

sH= -~PilogPi'

1

where s is the number of species considered and Piis the proportion which the population of the ith

ADJUNCTIVE BEHAVIOR 327

species is of the total number of individualscomposing the species population.

The rule and its most frequent mathematical ex­pression do not have the status of a mathematicaltheorem. Mathematical models have been proposedthat do substantiate the rule given certain assump­tions or conditions (Leigh, 1965; May, 1973, pp. 102­107; Webber, 1974). But, in general, "the greater thesize and connectance of a web, the larger the numberof characteristic modes of oscillation it possesses:since in general each mode is as likely to be un­stable as to be stable (unless the increased complexityis of a highly special kind), the addition of moreand more modes simply increases the chance for thetotal web to be unstable" (May, 1973, p. 75). How­ever, May (1973, pp. 75-76) indicates that "thebalance of evidence would seem to suggest that, inthe real world, increased complexity is usuallyassociated with greater stability." This occurs be­cause ecosystems are mathematically atypical."Nature represents a small and special part ofparameter space" (p. 76). A similar conclusion wasreached by Rogers and Hubbard (1974) in examininga parasite-host model in the light of actual data oninsect parasites and predators. The real speciespopulations did not show the instability predictedby the deductive model. On the other hand, thereare data which indicate that an increase in speciesdiversity does not always yield an increase in stability(Hairston et al., 1968; Watt, 1968, pp. 43-45). Thus,the diversity-stability rule must be treated withcaution. Its deductive underpinnings are suggestive,but not compelling. The relevant data likewise fallshort of being entirely persuasive. Nevertheless, therule is in accord with the analysis of several generalecological situations, as described at the beginning ofthis section. It is worth considering its applicabilityto analogaus relations in the realm of behavior.

Just as an uncomplicated ecosystem is subjectto instability or wide, cyclic oscillations which maybe remedied by an information increase in the speciesnetwork, an increase in the probability of preexistingbehaviors (viz, adjunctive behavior generation) maylimit unstable fluctuations in situations containingrelatively few behavioral vectors. Thus, the aug­mented adjunctive activity in a behaviorally un­stable situation may serve to strengthen a fluctuatingor deteriorating behavior rather than to occlude it.On the other hand, the vulnerability of a relativelysimplified island or monoculture ecosystem to in­vasion by a colonizing species has its counterpartin the excessive and intrusive aspects of adjunctivebehavior in the relatively barren situations that con­stitute both controlled experimental spaces and lifesituations lacking alternative opportunities or reper­toires. Thus, adjunctive behavior can either stabilizepreexisting, adaptive behavior in an otherwise un­stable situation or be so invasive in vulnerable situa-

328 FALK

tions that it eliminates other species of behavior.Whether this latter outeome is ereative or toxie forthe organism depends upon the overall adjustivevalue of the newly dominant behavior (Falk, 1971).If the adjunetive behavior is ehronie aleohol or otherdrug overindulgenee, then it would be eonsideredbehaviorally disadvantageous (Falk & Samson,1975). If it is a persistent artistie or scientificendeavor, it would be judged as creatively adaptive.The exaggerated and persistent aspect of adjunctivebehavior occurring in vulnerable contexts shouldnot be overemphasized. Rather, the conservative,stabilizing function of adjunctive behavior has prob­ably provided its major, selective, evolutionary ad­vantage.

ROMER'S RULE AND PREADAPTATION

Evolutionary change often is described in termsof special innovations which meet the specificchallenges and opportunities posed by altered en­vironments. While new circumstances may requirenew adaptive modes, there is also a conservative sideto evolutionary change. The variant selected may beone which initially facilitates the eontinuation of aprevious adaptive mode in altered cireumstances.Only later does the selective change develop a trulyinnovative relation to the altered ecological nicheof the species. For example, Romer (1959) describesthe evolution of the amphibians from 1ungfish 1ivingin stagnant fresh waters. With the seasonal droughtsof the Devonian period, the pools would have shrunkand become crowded as weIl as exhausted of food.Such periodic recessions of their pools produced aselection pressure favoring lungfish with fins strongenough to get thern to another pool. Strong finswhich could support weight out of the water weredeveloped initially not to invade the land, but ratherto continue an aquatic 1ife by moving to new pools."Land limbs were developed to reach the water, not10 leave it" (Romer, 1959, p.94). But once thisrather conservative step toward maintaining theiraquatic existence had been taken, it permitted theexploitation of a new ecological niche and a majorevolutionary opportunity. The innovation of ter­restrial life was actually a consequence of selectionpressure to maintain an aquatie one. In a similarvein, Romer (1959, p. 327) suggests that aneestralman may have been forced into ground locomotionfrom his arboreal existence because the expandinggrass lands and the consequent shrinking of forestareas required hirn to cross grassy prairies to reachpatches of forest. Just as walking developed so thatamphibians could reach pools, walking in man'sancestors may have developed so that he could main­tain an arboreallife.

These examples have been generalized by Hockett

and Ascher (1964) into what they call Romer's rule:"The initial survival value of a favorable innovationis conservative, in that it renders possible the main­tenance of a traditional way of life in the face ofchanged circumstances." Later, such innovationsmay have more radieal evolutionary implications,such as a changeover from aquatic or arboreal toterrestrial life. But, according to Romer's rule, theimmediate consequence of successful innovation is astabilization of the current adjustive context. Thekey notion here is that a contextual conservatism,operating through innovation, results in the stabiliza­tion of an unstable situation. In Romer's examples,the innovations were in mobility modes, but there areother ways in which innovations can occur. Ratherthan developing a new mobility mode, a species mayextend greatly a particular mode by gradually comingunder the guidance of new controlling stimuli.Skinner (1975) has described how, as the continentsdrifted slowly apart, distant migrations to breedinggrounds in turtles, fish, eels, and birds could havedeveloped gradually from the selection of genotypesand the transfer of stimulus control. Rather than newmodes cf mobility, various migrating species havedeployed increased responsiveness to new classes ofstimuli which have come to control the swimming orflying and permit the extensive trips neeessary toretain their traditional breeding grounds.

Romer's rule can be applied, thus, not only to in­novation in the modes of mobility, but to the extremeextensions of mobility classed as migrations. Sincethe rule deals with contextual conservatism, it iseven more probable that mechanisms would evolvewhich limit mobility in situations where this main­tains currently favorable adjustments. One waylimited mobility occurs is by ancillary behaviorcoming under the strong control of stimuli in theimmediate environment. It was indicated in theprevious section that such local diversifieation ofbehavior might promote the stabilization of anunstable situation. How this occurs will be amplifiedin later sections.

The contextual conservatism (Romer's rule) ofselective modifications which occur in response to achanged environment does not preclude the oppor­tunistie, adaptive significance these innovations maypossess with respect to other niehe possibilities. Theeonditions that allow some strueture or funetion tochange from control hy one kind of selective ad­vantage to another have been described by evolutiontheorists as preadaptation. Bock (1959) considersa structure "to be preadapted for a new function ifits present form which enables it to discharge itsoriginal funetion also enables it to assurne the newfunction whenever need for this function arises."The structure is molded into its preadapted form bythe selective advantages it affords the original func-

tion. For example, vertebrate cutaneous vasculariza­tion is welldeveloped. It has arespiratory function infish and amphibians, serves to take in or dischargeheat in reptiles, and maintains body temperaturewithin narrow limits in warrn-blooded animals(Cowles, 1958). The basic structure (cutaneousvascularization), while selected originally for itsrespiratory function, was preadapted for tempera­ture regulation. Another example is afforded by theremarkable degree to which the reptiles ancestral tobirds were preadapted for flight. These reptiles werearboreal, bipedal, with strong anterior extremities;even feathers were present (Mayr, 1960).

Before applying Romer's rule and the notion ofpreadaptation to the origin and functional implica­tions of adjunctive behavior, the nature of displace­ment activity will be discussed, as it forms a bridgebetween phenomena observed in natural settings andthe adjunctive behavior generated in operant condi­tioning situations.

DISPLACEMENT ACTIVITIES:STIMULUS CONTROL AND

FUNCTIONAL SIGNIFICANCE

The theory of natural selection implies that thebehavior patterns characterizing different animalspecies are interpretable in terms of their adaptivevalues. Displacement activities are difficult tointerpret within this framework, as the behaviorsequences are curiously intrusive. They often aredescribed as "out-of-context" behaviors, not onlyin relation to the behaviors occurring immediatelybefore and after the displacement activity, but alsoas adjustments to the current environmental situ­ation. These "incongruous" and "irrelevant"activities were related to Freudian symptomaticactions in early ethological theory and viewed as"outlets through which the thwarted drives can ex­press themselves in motion" (Tinbergen, 1952). Thenotion that displacement behavior functions to ridthe organism of surplus and potentially damagingimpulses was abandoned later by Tinbergen (1964)and other ethologists for various reasons. Chiefly,it became increasingly evident that the activities werenot so irrelevant to the current stimulating circum­stances as had been supposed.

Displacement aetivities occur in agonistic, sexual,or parental situations and specifically in conjunctionwith interruptions in the progress of behaviorscharacteristic of these situations. So ut he r ncormorants may show displacement brooding inintervals of fighting, or may displacement-bathewhen surprised by a human intruder (Armstrong,1965, p. 109). Many species of birds show displace­rnent feeding and preening activities during intervalsin intraspecific fighting or mating (Tinbergen, 1952).

ADJUNCTIVE BEHAVIOR 329

In experiments where one or more eggs of the black­headed gull were removed from the nest, vigorousdisplacemen: preening or nest building occurred asreactions to the interferenees with brooding(Moynihan, 1953). Displacement nest digging andfanning are observed in the male three-spined stickle­back during courtship behavior when the female failsto follow the male toward the nest (Tinbergen &VanIersel,1947).

In general, a wide range of displacement activitieshas been studied in various species, includingmammalian forms. Displacement feeding, drinking,sexual display, mounting, sleeping, and groominghave been observed, as well as those activities men­tioned previously. The situations producing displace­ment activities involve interferenee with appetitive orconsummatory behavior which is in progress. Thisean occur in several ways (Arrnstrong, 1950;Tinbergen, 1952). Another organism or physiealevent may intrude upon the situation. Incompatiblebehaviors ean alternate and inhibit each other. Thissecond type of event is described in ethologicalaccounts as a conflict situation either between sexualand agonistic behaviors in courtship, or betweenagonistic and incubation behaviors in a broodingsituation. The conditions under which conflictingbehaviors occur also give rise to various displacementactivities. Finally, a behavior can be terminatedrapidly by the sudden eessation of stimuli elicitingor sustaining the behavior. For example, "the red­necked phalarope, where the female takes theinitiative in courtship and eoition ... often attacksthe male immediately after eoition. Also, she goes onland and indulges in a bout of nest-building"(Tinbergen, 1952). In a similar vein, Armstrong(1950) notes that displacement aetivities oeeur "aftera goal has been attained or the threshold of expressionof a behavior-pattern has been raised." It is im­portant to note that, in all of the above situations,the ongoing behavior was interfered with or sum­marily terminated and that displacement aetivitiesresulted. These curious outcomes are no longerlooked upon as an epiphenomenon of eentralnervous design ("sparking-over"). It is reeognizednow that whatever may be the ultimate explanationof their eausation, the dynamics of displacementactivities are under the control of current environ­mental stimuli.

Stimulus control factors are important in deter­mining the particular displacement aetivitiesprodueed. For example, Morris (1954) noted thatthe displacements of the zebra fineh depended uponwhat stimuli were nearby. Feeding was the displace­ment behavior observed when food was in thevieinity, mounting occurred when a female was near,and when neither were easily available, sleep orcomfort responses resulted. Formerly, it had been

330 FALK

assumed that the stimulus situation appropriate tothe observed displacement activities was absent, andthat the behaviors were thus entirely incongruous.But increasing evidence pointed to the conclusionthat when the behavior dominating a situation be­comes, for whatever reason, impeded, behaviorappropriate to other concurrent stimuli becomesprimary (Vanlersel & Bol, 1958). Events whichimpede the behavior dominating the situation,whether these events are barriers, intrusions by acompeting conspecific or potential predator, con­flicting behavior, or the absence of appropriate sus­taining stimuli, all produce displacement activities bya permissive action. Detailed analyses of displace­ment fanning in the three-spined stickleback(Sevenster, 1961) and displacement preening in thechaffinch (Rowell, 1961) showed that external stimuliwhich facilitated the normal performance of theseactivities also increased their probability when theyoccurred as displacement behaviors. Experimentsby McFarland (1965) on displacement feeding in thewater-deprived Barbary dove induced in a variety ofcarefully controlled situations demonstrated theimportance of both behavioral history and currentsituational stimuli.

As indicated previously (Falk , 1971), there is con­tinuity in the processes determining displacement andadjunctive behaviors. Both occur when powerfullydetermined, ongoing behavior is impeded; in both,the particular behavior next in evidence is a functionof available facilitating stimuli. Further , thebehaviors in both situations have a certain aspect ofincongruity: they are excessive and appear to con­tribute little of adaptive significance to the be­havioral situation impeded.

It is worth considering this last point-the in­trusive and incongruous aspect of displacement andadjunctive behaviors. As noted, displacement be­haviors are under the control of faci1itating stimuliand thus not entirely incongruous with respect tosituationa1 vectors. Likewise, adjunctive behaviorsare not newly engendered activities. They are in­creases in the magnitude and probability of behaviorsalready present as moderate, base-rate responses tothe general situation (Falk, 1971). Both displacementand adjunctive activities probably are produced bymuch the same stimuli that produce these behaviors(drinking, aggression, nest-building, etc.) in other,more usual situations. The unusual features involvethe temporallocus of such behavior and its apparent­ly nonfunctional exaggeration.

Persistent, nonfunctiona1 behaviors with widespecies representation would be something of apuzzle to exp1ain in terms of natural selection. Butrecent work on dis placement indicates that theprocess may possess certain adaptive functions.Wilz (1970) studied nest-re la ted displacement

activities (fanning, creeping through) which occurduring courtship in the male three-spined stickleback.These displacement activities are quite frequentfollowing "dorsal pricking," in which, instead ofleading the female towards the nest, the male pushesthe female away with the dorsal side towards thewater surface. Such pricking occurs if the male hasjust been stimulated by the intrusion of anothermale. Experimental analysis revealed that the dis­placement activities facilitated the transitionfrom aggression to courtship. When nest creeping­through responses were prevented after pricking,biting of the female was frequent and zig-zag leadingwas infrequent. The displacement activities were acause rather than an effect of a transition fromaggressive to sexual responses. The ritualization ofdisplacement activities into communicative displayswhich function to mitigate aggression and promotesexual responses in partners has been well described(Tinbergen, 1952). It now appears that displacementactivities may be effective in promoting this transi­tion in the performer prior to their selective ritual­ization into communicative displays.

There are numerous descriptions which suggest.that displacement activities function directly asdistractions luring predators away from the bird'snesting area. Armstrong (1949a) notes that displace­ment brooding in a spot where there are neither eggsnorchicks leads predators away from the actualnest site. Displacement feeding in the lapwing occurswhen it does not approach its nest in the presenceof a predator. The feeding behavior may functionto deceive predators in that they are less likely tostalk a bird apparently feeding than one broodingor tending chicks (Armstrong, 1949a). More complexdiversionary displays, such as injury simulation,arise through the ritualization of agonistic andmating display components (Armstrong, 1949b;Williamson, 1950).

These studies indicate that, even prior to theirritualization into display, dis placement behaviorsafford the organism definite adaptive advantages,Armstrong (1950) suggests how various maintenanceactivities such as nest building and nest lining haveevolved directly from displacement movements.Species which are prone to displacement activities,according to his analysis, exhibit these modifiedbehaviors when faced with altered environmentalcircumstances. This enables them 10 incorporatethese modifications into new adaptive patterns ratherthan requiring the evolutionary construction of ad­justive modifications more slowly from smaller be­havioral units. Alcock (1972), in a similar vein, hasshown how various instances of the use of tools byfeeding animals could have arisen directly from dis­placement and redirection activities. For example,Darwin's woodpecker finch of the Galapagos carries

about a cactus thorn or twig with which it pokesinto tree cracks, dislodging insects and larvae whichit then seizes (Lack, 1947, pp. 58-59). This finch hasa beak and tongue length not as suitable as a truewoodpecker's for retrieving its prey from crevices.Alcock (1972) suggests that seeing nearby prey ob­tainable only intermittently may have producednest-material-gathering displacement activity withnearby twigs. If reattracted by an insect while stillholding a twig, a jab of the beak could now beeffective in dislodging the prey. Selection pressurewould then favor displacement-prone members ofthe species.

RITUALIZATION, ROMER'S RULE,AND TUE ADAPTIVE VALUEOF ADJUNCTIVE BEUAVIOR

Ritualization, a concept introduced into ethologyby Huxley, is defined as the process whereby main­tenance and displacement activities have becomespecialized for communicative functions throughselective pressures. The resulting displays are thoughtto provide unambiguous signals which minimizedamage in intraspecies agonistic encounters andpromote efficient releasing stimuli in sexual andsocial bonding situations (Huxley, 1966). However,the agonistic encounters involve not only intraspeciessignals but also interspecies predator-prey relations.The latter include ritualized activities resulting ininjury- and chick-simulation distraction displays(Armstrong, 1949a, 1949b; Williamson, 1950). In­timidation and cryptic displays occur in moths. Wing"eyespots" are revealed in the intimidation displaywhich mimic predator (owl) eye patterns, and un­responsiveness to strong stimulation (e.g., handling)by potential predators characterizes the crypticdisplay (Biest, 1957). Distraction, intimidation, andcryptic displays having interspecies predator-defensevalue appear to be products of ritualization processessimilar to those yielding intraspecies communicativefunctions.

As a behavior becomes ritualized, it is thought tobe released by stimulus situations different fromthose that produced the ancestral behavior. Forexample, the preening involved in courtship displayis described as being emancipated not only fromits original maintenance and comfort context butalso from a displacement role. It is evoked by a stim­ulus complex relevant to courtship and mating. Asritualization proceeds, presumably the behavior losesits responsiveness to the usual stimuli that facilitateor inhibit it in other contexts, and in this sense itprogresses beyond displacement behavior, whichretains at least a portion of its ancestral responsive­ness. The emancipation of displacement activitiesfrom their original controlling stimuli and their

ADJUNCTlVE BEHAVIOR 331

attachment to other types of stimulus control as theybecome ritualized in to displays has been describcdalmost entirely in terms of the eommunicative func­tions served. However, there are similarities in re­sponse topography dynamics in ritualized and ad­junctive behaviors whieh suggest that the character­istic intensification may have sources other thansocial display selective pressures. Both behaviorsare exaggerated either in movement amplitude orintensity. Both are protracted and acquire rhythmicqualities. Both reveal a "typical intensity" (Morris,1957): when they occur, their form and intensityresist alterations by stimulus variables that usuallyproduce graded changes in the behavior. There isproof that at least one adjunctive behavior, schedule­indueed polydipsia, is emaneipated from its ancestralcontrolling variables. For example, presession waterloading failed to attenuate the polydipsie behaviorto a significant degree (Falk, 1969). In contrast, suchloading considerably reduces drinking producedby conditions affecting body water status.

Ritualization, the ceremonious exaggeration andpersistence of what would appear to be ancillarybehaviors under certain environmental conditions,oceurs, then, in some situations unrelated to socialcommunicative display. Ritualized behavior mayacquire a communicative function in a succeedingphase; but neither its inception as a dis placementactivity nor its exaggeration into a strong and per­sistent behavior are necessarily in the service of com­munication. It is more likely that ritualized or ad­junctive behavior is preadapted for elaboration intocommunicative display functions (Spurway &Haldane, 1953). Ritualized displays are somewhatmaladaptive in that they often make the organismeonspicuous to predators as weil as to potentialmates. Adjunctive behavior too is a considerableinvestment in time and energy, the utility of whieh isnot readily evident. But undoubtedly selective ad­vantages acerue to species prone to the ritualizationof dis placement aetivities under the appropriateenvironmental conditions. It is probable that thisoceurs in aecordance with Romer's rule: the ritual­ized or adjunctive behavior maintains adjustment tothe current eontext. By emphasizing an ancillarybehavior already possessing some adaptive signifi­cance in that situation, the entire context is stabilizedand conserved (cf. the diversity-stability rule).Romer 's rule, as formulated by Hockett and Ascher(1964) and quoted previously, may now be restated asit applies to the functional significance of adjunetivebehavior: The adaptive value of adjunctive behaviorconsists in its maintenance of the organism 's engage­ment with a problematic, but overall favorable,situation.

Wh at eharacterizes the "problematic" situationsreferred to above, and how does adjunctive behavior

332 FALK

stabilize such situations? Conditions yielding dis­placement activities-agonistic, mating, incuba­tion, and parental care contexts-were describedpreviously. Their problematic aspect often occurswhen the appropriate behaviors are impeded: theantagonist makes the outcome of an agonisticcontext uncertain, mating encounters run unevencourses fraught with agonistic components, whileincubation and parental situations are interrupted byreturning mates, group members, or by the approachof potential predators. These conditions involveescape components. Hut if a territory is to be de­fended, mating accomplished, or offspring con­served, these situations must be preserved. If escapebehavior were to dominate these situations, evenmomentarily, there would be an appreciable chancethat territory, mates, and progeny would be lost.There is danger of such an eventuality in just thesetypes of situations because they are composed of twoopposed components (fight vs. flight: etc.) in un­stable equilibrium. Adjunctive behaviors or displace­ment activities have a protective or buffering actionin that they effect the continued engagernent of theorganism with the problematic situation. Time­sharing occurs between the problem situation andone or more adjunctive behaviors until resolution ofthe opposing vectors of the problem situationdevelops. Given the unstable equilibrium of thegenerator vectors, the resolution is less likely to beprecipitous under circumstances where adjunctivebehavior inclusion diversifies the situation. Ad­junctive behavior acts as a stabilizing, protectiveagent, preserving the opposing-vector situation whileit develops. This stabilization occurs when ancillary,adaptive behaviors which receive facilitatory stimulipreexistent in the situation become exaggerated andritualized. Mechanisms apparently have evolvedby genetic selection that strengthen adjunctive be­havior under the opposing-vector conditions definedby generator situations.

The adaptive significance of adjunctive behaviorresides in the several consequences that flow from itsincorporation and exaggeration in generator circum­stances. (a) Such situations are stabilized by thediversification of the network of behaviors broughtabout by adjunctive behavior inclusion and per­sistence. (b) Spatially, adjunctive behavior rnain­tains the organism in the situation by occludingescape behavior, thereby postponing aresolutionof the circumstances until the interplay of generatorvectors evolves further. This serves to pace thegenerator-situation behavior, which might other­wise resolve precipitously given the current dangersand constraints. An unstable equilibrium is con­verted into a stable one, allowing the generatorvectors to develop at a measured pace. (c) Finally,adjunctive behaviors are complex responses of con-

siderable strength and duration which are pre­adapted for other potential functions, such as com­munication, abode construction, and tool use. AsSkinner remarked in a passage following his dis­cussion of preadaptation, "behavior may have ad­vantages which have played no role in its selection"(Skinner, 1966).

OF SPECIES AND PATcn FUGACITY

Some species are highly mobile hunters (manycarnivores and herbivorous seed and fruit eaters),exploiting their environmental food sources bymoving from place to place. These movements arecontrolled in a complex manner by the "patchy"availability of food sources, so that as one patchbecomes sparce the animal moves to a more favor­able one (e.g., Charnov, 1976; MacArthur & Pianka,1966). The herbivorous grazers and browsers areless mobile, exploiting relatively small, denselysupplied horne ranges (McNab, 1963). Still otherspecies move to a favorable place where, with im­mobile waiting, prey may be ambushed (e.g., bob­cats, leopards, and many spiders and snakes). Finally,some animals stay within one locus and capture preythrough lure display (e.g., angler fish: Wickler, 1968,p. 124ft).

The mobility of an organism in response to currentfood availability is a function of the particularspecies' foraging mode and the food-resource statuswithin the patch of present residence. Species withlow mobility as a dominant feature of their foragingmode might be expected to show less escape be­havior from a domain of attenuated food acquisi­tion rate. If this is the case, then they should evidenceless adjunctive behavior as a result of food schedulemanipulation. Adjunctive behavior has been demon­strated in several species; nevertheless, some speciesmay not yield schedule-induced polydipsia (e.g., thegolden hamster: Wilson & Spencer, 1975) or otheradjuncts when intermittent-food generator condi­tions are imposed. It is possible that if adjunctivebehavior is, as suggested, a stabilizing activity main­taining the organism's engagement with a situationcontaining escape components, a minimallyfugacious species may not require the diversificationof adjunctive behavior to protect a nonoptimal, butfeasible, feeding situation. The present evidence onspecies differences with regard to adjunctive be­havior is too fragmentary to enable the extractionof general principles. Hut it may be of interest tospeculate concerning species of intermediate mobility.These include species which are fairly mobile betweenthe long bouts of immobility required for ambush.Finer, less extravagant, stabilizing displacementmovements would be expected in such animals for atleast two reasons. First, with the rather static ambush

foraging mode, the escape component generatedby feeding intermittence should be greatly diminished.Second, the requirement for stealth would precludeexcessive or exaggerated movement. Possibly thesmall-amplitude tail twitches of felines as they lie inwait for prey to pounce on is a behavioral diversifica­tion facilitating the required stasis. Likewise, manysnakes vibrate the tail in agonistic situations, abehavior which may occlude escape. Further, insome snakes, tail-tip twitching and coloration havedeveloped to resernble an insect larva. This luredisplay probably derived from the preadapted an­cestral displacement activity. Additional lure displaysare described by Alcock (1975, p. 343ff) and Wickler(1968).

It is probable that the contextual conservatismeffected by displacement activities which limit thefugacity of ambush feeders evolves into lure displaysunder the selective pressure of more efficient preda­tion. This may be compared with the enforced limitson fugacity required by human tasks and social inter­actions which must be done for long periods in oneplace (e.g., writing an articJe, assembly line work,crowded parties). These circumstances often inducea common set of adjunctive behaviors: smoking,fluid and food overindulgence, and ritualized verbaldisplays, all of which function to sustain situationaltasks and interactions containing escape components(e.g., Roy, 1960). The adjunctive behaviors blockfugacity and keep the organism engaged with thesituation-on the job, This, too, can develop intoluring in a variety of situations. Consider thatexaggerated smoking or gum-chewing movements, oragitated vocal inflections (or fan manipulation insome cultures), function as lures, since they com­municate an unstable equilibrium status, and hencepossible vulnerability of the displayer.

Ta return to the initial theme of this section, thevery mobile hunters with large prey-search timesrelative to prey-pursuit times (MacArthur, 1972,p. 61ff), with rat her low patch-fugacity thresholds,should yield Iood-schedule-induced adjunctive be­havior most readily. The creatures with less mobilemodes of foraging may reveal less adjunctive be­havior, but it may not be inconsequential. Forexarnple, the ambush mode may remain open to low­amplitude adjunctive behavior inclusion as a stabilizerof enforced waiting. Also, they may depend uponhigh prey abundance, leaving them sensitive to smallcapture rate decrements.

ADJUNCTIVE BEHA VIOR LAWS

By way of a summary, four provisional laws arepresented with the hope that they can serve to unifythe direction of additional research efforts in thisarea.

ADJUNCTIVE BEHAVIOR 333

(I) Law 01 adjunct exaggeration. The more con­strained a situation (a) spatially, (b) in terrns ofgenerator-schedule commodity deficit, or (c) withrespect to the competition of scheduled behaviorwith an escape vector, the greater will be the intensityof the adjunctive behavior.

This law essentially defines the set of conditionsgoverning adjunctive response strengthening. Itattempts to state the circumstances increasing thereinforcing properties of specific response patternsand environmental commodities which are involvedin the increased probability of behavioral engage­ment termed "adjunctive behavior." In (a) above,the reference is literally to constrained spatial area.In the usual operant experimental space arrange­ment, it implies that larger areas should result insomewhat attenuated adjunctive behavior. The im­mobility contingencies imposed on ambush-rnodefeeders implies a similar constraint, but there are amore complexly determined set of relevant factors(discussed in the previous section) which limit theiradjunctive behavior intensity. In (b) above, most ofwhat is known concerning the relation betweencommodity deficit and adjunctive behavior has beengathered by food-ration manipulation of bodyweight, and the inverse relation was discussed brieflyin the introduction. In (c), the implication is that asthe schedule-determined and escape vectors approachequality, the situation increases in constraint. "Con­straint" here should be understood as the necessity ofengaging in behavior relevant to two opposingvectors. If one of the vectors departs from the pointof equality, it would gain ascendency and the situa­tional constraint would dissipate. Figure 1 showsthis constraint relation as it applies to adjunctivebehavior generation by schedule induction. The dia­gram assurnes that the conditions described in (a)and (b) have been met, at least to some minimallyeffective extent. The linear aspect of the functionshould not be regarded as a necessary feature of thelaw. The consummatory behavior function (schedule­determined vector) can be controlled rat her preciselyby the experiment er and simply reflects themanipulated condition indicated on the x-axis. Thereis little information available on the shape of theescape function. Existing information suggests that itis bitonic (Brown & Flory, 1972), as one might expectif it were similar to other schedule-induced behaviors(cf. introductory section). However, schedule-inducedescape has been explored by allowing animals toescape from the discriminative stimuli associatedwith schedules, rather than permitting spatial escape.These two kinds of escape may not be equivalent,but only experiment ation will tell. The ethologicaland ecological relations discussed involve mainlyspatial escape vectors (encounters containing agonisticcomponents, patch fugacity). The view implied by

334 FALK

SCHEDULED RATE OF CONSUMMATORY EVENTS

them. lf adjunctive behavior is invasive enough,schedule contingency requirements may not be met,increasing reinforcement intermittency. This couldresult in a further increase in the adjunct if the inter­mittence change lies on the rising limb of the bitonicfunction. A major interference by adjunctive be­havior with its generator conditions would, ofcourse, be self-limiting. More often the intrusive ordisadvantageous aspect of adjunctive behavior (e.g.,excessive smoking, drug taking, loquaciousness)interferes with adaptive behaviors other than thosetied to the generating conditions. But excessive andintrusive aspects of adjunctive behavior can be over­emphasized. lts major adaptive significance, as thelaw of protective inclusion states, lies in its stabilizingfunction .

\. Villarreal, 1. Schedule-induced pica. Paper presented at themeeting of the Eastern Psychological Association, Boston, April1%7.

REfERENCE NOTE

HIGH

CONSUMMATORYRIHAVIOR

lOW

...111

~Go111...•

;...iiiCl:

§Go

Figure 1. Adjunctive behavior probability as the inverse of thedifference between two functions: (1) the rate of generarer­schedule consummatory behavior, and (2) the probability ofescape behavior,

Figure 1 is that schedule-induced escape has a specialstatus. Its interaction with the consummatory be­havior function is responsible for the induction of theother adjunctive behaviors.

(2) Law 0/ adjunct selection. The adjunctive be­havior which occurs, other factors remaining equal, isthat which receives maximal stimulus facilitation.

Present information on this point is derived mainlyfrom studies on the stimulus facilitation deter­minants of displacement activities and was discussedin a previous section.

(3) Law 0/ protective inclusion. A situation evok­ing two competing behavioral vectors of equivalentstrength which result in an unstable equilibrium,such that one of the vectors could become eliminatedor undergo oscillations of large magnitude, will in­crease in stability if one or more adjunctive behaviorsshare the behavioral domain.

(4) Law 0/ minimal intrusion. Other factorsremaining equal, the adjunctive behaviors thatemerge are those which maintain the organism inclosest proximity to the generating situation andoccur at times minimally competitive with the be­havior primarily determined by the generatorschedule.

In accordance with this law, adjunctive behaviortypically occurs at postreinforcement times and atpoints where reinforcement probability is low (Falk,1969, 1971). But, as the phrase "other factors re­maining equal" intimates, adjunctive behavior canbe so intrusive as to interfere with its generatingconditions. Often, however, it only exacerbates

REFERENCES

ALCOCK, J. The evolution of the use of tools by feeding animals.Evolution, 1972, 26, 464-473.

ALCOCK. 1. Animal behavior: An evolutionary approach.Sunderland, Mass: Sinauer Associates, 1975.

ALLEN. 1. D .• & KENSRALO, D. R., JR. Schedule-induced drinkingas a function of interreinforcement interval in the rhesus monkey.Journal of the Experimental Analysis of Behavior, 1976, 26,257-267.

ARMSTRONG. E. A. Diversionary display. Part \. Connotation andterminology. Ibis, 1949, 91, 88-97. (a)

ARMSTRONG, E. A. Diversionary display. Part 2. The nature andorigin of distraction display. Ibis, 1949, 91, 179-188. (b)

ARMSTRONG. E. A. The nature and function of displacement activi­lies. Symposia ofthe Society for Experimental Biology, 1950, 4,361-384.

ARMSTRONG, E. A. Bird display and behavior (Rev. 00.). NewYork: Dover, 1%5.

AZRJN, N. H. Time-out from positive reinforcement. Science, 1%1,133, 382-383.

AZRJN. N. H., Hurcmssos. R. R., & HAKE, D. F. Extinction­induced aggression. Journal of the Experimental Analysis 0/ofBehavior, 1966. 9. 191-204.

BLEsT, A. D. The evolution of protective displays in theSaturnoidea and Sphingidae (Lepidoptera), Behaviour, 1957, 11,257-309.

BOCK, W. J. Preadaptation and multiple evolutionary pathways.Evolution, 1959, 13, 194-211.

BRowN, T. G., & FLORY, R. K. Schedule-induced escape fromfixed-interval reinforcement. Journal 0/ the Experimental Analy­sis of Behavior, 1972, 17, 395-403.

CRARNOV, E. L. Optimal foraging, the marginal value theorem.Theoretical Population Biology, 1976, 9. 129-136.

CHILLAG. D., & MENDELSON. 1. Schedule-induced airlicking as afunction of body-weight deficit in rats, Physiology and Behavior,1971. 6.603-605.

COHEN. P. S.. & LOONEY, T. A. Schedule-induced mirror respond­ing in the pigeon. Journal ofthe Experimental Analysis 0/Behav­ior, 1973, 19,395-408.

COWLES, R. B. Possible origin of dermal temperature regulation.Evolution, 1958, 12, 347-357.

DOVE, L. D. Relation between level offood deprivation and rate ofschedule-induced attack. Journal pIthe Experimental Analysis 0/Behavior, 1976, 25, 63-68.

ELTON. C. S. The ecology of invasions by animals and plants .London: Methuen, 1958.

FALK. J. L. Produetion of polydipsia in normal rats by an intermit­tent food schedule. Science, 1961, 133. 195-196.

FALK. J. L. Schedule-induced polydipsia as a function of fixed in­tervallength. Journal of the Experimental Analysis of Behavior,1966. 9. 37-39.

FALK. 1. L. Conditions producing psychogenic polydipsia in ani­mals. Annals ojthe New York Academy of Seiences. 1969. 157,569-593.

FALK. J. L. The nature and determinants of adjunetive behavior.Physiology and Behavior, 1971, 6. 577-588.

FALK. J. L., & SAMSON. H. H. Schedule-induced physical depen­dence on ethanol. Pharmacological Reviews. 1975, 27.449-464.

FLORY. R. K. Attack behavior as a funetion of minimum inter-foodinterval. Journal ofthe Experimental Analysis ofBehavior, 1969.12, 825-828.

FLORY, R. K. The control of schedule-induced polydipsia: Fre­quency and magnitude of reinforcement. Leaming and Motiva­tion. 1971, 2, 215-227.

GEERTZ. C. Agricultural involution. Berkeley: University ofCalifornia Press, 1963.

HAIRSTON. N. G., ALLAN. J. D.. COLWELL. R. K., FUTUYAMA,D. J., HOWELL. J., LUBIN. M. D.. MATHIAS. J.. &VANDERMEER. J. H. The relationship between species diversityand stability: An experimental approach with protozoa andbacteria. Ecology, 1968, 49, 1091-1101.

HOCKETT. C. F., - & ASCHER, R. The human revolution.Al1Ierican Scientist . 1964. 52, 70-92.

HUTCHINSON, G. E. Homage to Santa Rosalia or why are thereso many kinds of animals? American Naturalist, 1959, 93.145-159.

HUTCHINSON. R. R., AZRIN. N. H., & HUNT. G. M. Attackproduced by intermittent reinforcernent of a concurrentoperant response, Journal ot' the Experimental Analysis ofBehuvior, 1968, 11,489-495.

HUXLEY. J. lntroduction, A discussion on ritualization of behaviourin animals and man. Philosophical Transactions of the RoyalSocietv ofLondon, 1966. 251. 249-271,

KACHA~oiF. R., LEVEILLE. R.. McLELLAND. J. P., & WAYNER,M. J. Schedule induced behavior in hurnans. Physiology andBehuvior, 1973. 11. 395-398.

KING. G. D. Wheel running in the rat induced by a fixed-timepresentation of water. Animal Leaming & Behavior, 1974.2. 325-328.

LACK. D. Darwin's finches. Cambridge: Cambridge UniversityPress. 1947.

LEIGH. E. G., JR. On the relation between the productivity,biomass, diversity, and stability of a community. Ecology ;1965,53.777-783.

LEVITSKY, D.. & COLLIER. G. Schedule-induced wheel running.Physiology and Behavior, 1968. 3,571-573.

MACARTHUR. R. H. Fluctuations of animal populations, and ameasure of community stability. Ecology, 1955. 36. 533-536.

MACARTHUR, R. H. Geographical ecology: Patterns in thedistribution ofspecies. New York: Harper & Row, 1972.

MACARTHUR, R. H., & PIANKA. E. R. On optimal use of apatchy environrnent. American Naturalist, 1966. 100. 603-609.

MAY, R. M. Stability and complexity in model ecosystems.Princeton: Princeton University Press. 1973.

MAYR. E. The emergence ofevolutionary novelties. In S. Tax (Ed.),Evolution ajter Darwin (Vol, 1). Chicago: University of ChicagoPress. 1960. Pp. 349-380.

McFARLAND, D. J. Hunger. thirst and displacement pecking inthe Barbary dove. Animul Behaviour, 1965, 13, 293-300.

McNAB. B. K. Bioenergetics and the determination of horne rangesize. American Naturalist. 1963. 97. 133-140.

MENDELSON. J.. & CHILLAG. D. Schedule-induced air lickingin rats. Physiology and Behavior, 1970. 5. 535-537.

MILLER. J. S., & GOLLUB. L. R. Adjunctive and operant boltpecking in the pigeon. Psychological Record. 1974. 24. 203-208.

MORRIS. D. The reproductive behaviour of the zebra finch

ADJUNCTIVE BEHAVIOR 335

iPoephila guttata), with special reference to pseudofemalebehaviour and displacement activities. Behaviour, 1954. 6.271-322.

MORRIS. D. "Typical intensity" and its relation to the problemof ritualization. Behaviour, 1957. 11, 1-12.

MOYNIHAN. M. Some displacement activities of the black-headedgull. Behaviour, 1953. 5. 58-80.

ODUM. E. P. Fundamentals of ecology (3rd ed.). Philadelphia:Saunders, 1971.

ROGERS. D.. & HUBBARD. S. How the behavior of parasit es andpredators promotes population stability. In M. B. Usher& M. H. Williamson (Eds.), Ecological stability. New York:Wiley. 1974. Pp. 99-119.

ROMER. A. S. The vertebrate story, Chicago: University ofChicago Press. 1959.

ROWELL. C. H. F. Displacement grooming in the chaffinch.Anima! Behaviour, 1%1, 9. 38-63.

Rov, D. F. Banana time: Job satisfaction and informalinteraetion. Human Organization, 1960, 18, 158·168.

SEVENSTER. P. A causal analysis of a displacement activity(fanning in Gasteroteus aculeatus L.). Behaviour Supplement,1961. 9.1-170.

SKINNER. B. F. The phylogenyand ontogeny of behavior. Science,1966. 153. 1205-1213.

SKINNER. B. F. The shaping of phylogenic behavior. Journal ofthe Experimental Analysis ofBehavior, 1975. 24. 117-120.

SPURWAY. H.. & HALDANE. J. B. S. The comparative ethologyof vertebrate breathing. I. Breathing in newts, with a generalsurvey. Behaviour, 1953. 6, 8-34.

THOMPSON. D. M. Escape from SD associated with fixed-ratioreinforcement. Journal ofthe Experimental Analysis ofBehavior,1964. 7. 1-8.

TINBERGEN, N. "Derived" activities; their causation, biologicalsigniticance, origin, and emancipation during evolution.Quarter/y Review ofBiology, 1952. 27, 1-32.

TINBERGEN . N. The evolution of signaling devices. In W. Etkin(Ed.), Social behavior and organization among vertebrates.Chicago: University of Chicago Press. 1964. Pp. 206-230.

TINBERGEN. N., & VANIERSEL, J. J. A. "Displacement reactions"in the three-spined stickleback. Behaviour, 1947, 1,56·63.

VANIERsEL. J. J. A., & BOL, A. C. A. Preening oftwo tern species.A study on displacement aetivities. Behaviour, 1958, 13, 1-88.

WALLACE. M.. SINGER. G.. WAYNER, M. 1., & COOK. P.Adjunctive behavior in humans during game playing. Physiologyund Behavior, 1975. 14. 651-654.

WALLACE. M.. & SINGER. G. Adjunctive behavior and smokinginduced by a maze solving schedule in humans. Physiology andBehavior, 1976. 17. 849-852.

WATT, K. E. Ecology and resource management. New York:McGraw-HilI. 1968.

WAYNER. M. J., SINGER. G.. CIMINO. K.. STEIN, J., & DWOSKIN.L. Adjunctive behavior induced by different conditions of wheelrunning. Physiology and Behavior, 1975, 14. 507-510.

WEBBER. M. I. Food web linkage complexity and stability in amodel ecosystem. In M. B. Usher & M. H. Williamson (Eds.),Ecological stability. New York: Wiley. 1974. Pp. 165-176.

WICKLER. W. Mimicry in plants and animals. New York:McGraw-Hill. 1968.

WILLIAMSON, K. The distraction behaviour of the Faeroe snipe./ bis. 1950. 92. 66-74.

WILSON. E. 0., & BOSSERT. W. H. A primer of populationbiology. Sunderland , Mass: Sinauer Associates, 1971.

WILSON. S.. & SPENCER. W. B. Schedule-induced polydipsia:Species limitations. Psychological Reports, 1975, 36. 863-866.

WILZ. K. 1. Causal and functional analysis of dorsal prickingand nest activity in the courtship of the three-spinedstickleback Gasterosteus aculeatus . Animal Behaviour, 1970,18.115-124.

(Received for publication May 6. 1977;revisionaccepted October 6. 1977.)