How intelligence is wired: innate complex patterns or acquired coordinations?

One of the main features of “instinctive” behaviour is that fixed action patterns are common to all members of the species (species-typical) and therefore they serve characteristic of the species. We thus may expect that all sticklebacks will perform their zigzag courtship dance in similar ways; all squirrels will gnaw nuts applying similar techniques, and all wolves will stop aggression aimed at them by high ranked individuals turning the unprotected neck as a sign of strict obedience. There are, however, individual deviations, and we can start from three examples of animals behaviour listed above.

Mainardi (1976) observed how several wolves hunted one poor male wolf that definitely demonstrated signs of obedience turning its neck, bending down, wagging and low smiling. All these signals were of no help, aggressive mates tore him to shreds. Something was wrong with his signs of obedience, and we will never know exactly what. At least, we no more believe in absolute power of signs of obedience in animal behaviour, although they work in most cases.

In the previous issue I have cited results of Eible-Eibesfeldt’s (1961-a) experiment with squirrels gnawing nuts. Most of them acquired the most efficient technique. There were, however, displays of individual variance. Some squirrels learned to open nuts by gnawing a hole through a few closely spaced furrows instead of following the path of least resistance. One squirrel achieved almost instant success by gnawing a hole into the base of the nut, and continued to use this technique. This example shows how early experience influence the display of species-specific behaviour.

Sticklebacks have to seek a compromise between aspiration for calling to female by their zigzag dances and risk to attract predators’ attention. In open places they dance on the verge of dangerous, and intensity of zigzags depends on the level of risk (Candolin and Voigt, 2003). Environmental influences thus shift displays of species-specific courtship behaviour, although it is rather conservative.

Of course, it is logically impossible to test whether behaviour would develop the same way in all environments. At the same time, the “sameness” of behaviour between members of the same species does not exclude the possibility that all members of a particular species share common learning experiences. Eibl-Eibesfeldt (1970) calls learned behavioural sequences acquired coordinations, taking into account that they contain fixed action patterns as elements which are combined into the functional whole by learning. Variants of hunting behaviour in mammals and birds can serve as good examples of acquired coordinations.

It is tempting to include all complex species-specific stereotypes into this set. We however take chances to find a few individuals in populations that display very complex behavioural patterns “all and immediately”. Even a single individual of that sort displays clearly that we deal with an innate behavioural pattern. Insight thus can be obtained from studies at the population level. We have got it in our laboratory when first applied Haspar Hauser method to studying development of two complex stereotypes in ants, namely, aphid tending and hinting jumping victims.

First, we examined a question whether ants learn to milk aphids or this very complex stereotype is innate (Reznikova and Novgorodova, 1998). It is known that in many ant species aphids’ sweet excretions are one of the main sources of carbohydrates for adult family members. Until recently, nothing was known about the roles of innate and acquired behaviour in ant-aphid interactions. We conducted experiments on three control colonies of red wood ants and three experimental “Haspar Hauser” colonies of the same species composed of individuals raised from pupae in separate laboratory nests and deprived of the experience of communication with adult ants as well as with aphids. We timed behaviour of 230 individually labelled “Kaspar-Hauser” ants throughout their daily cycles of activities until the end of their seasonal activity.

Under natural conditions, the behaviour of an aphid milker during direct contact with an aphid is stereotypical and specific. An ant strokes an aphid’s abdomen with its antennae, which are folded so that their ends are close to the ants’ trophi. In this way the ant is begging for a drop of the sweet excretion; the ant immediately catches the drop and puts it into its crop. It is important to note that during trophallaxis, i.e. the exchange of liquid food belched from the crop with other ants; the antennae are folded in similar manner. Trophallaxis is the basic process common not only for ant species but also for all social hymenopterans. Kloft (1960) compared the aphid’s abdomen to the head of an ant offering liquid food. The behaviour of the aphids closely resembles that of ants during trophallaxis and apparently triggers the same behavioural stereotype in naï ve aphid milkers. However, the behaviour of an ant eating carbohydrate liquid food from an open trough or encountering various objects is distinctly different. In this case the ant feels the object with extended, almost straight antennae, with the frequency of tapping reflecting the degree of the ant’s interest in the object; however, the position of the antennae themselves does not change.

In the deprivation (Kaspar Hauser) experiment, when ants encountered aphids for the first time, they perceived the aphids as any other unknown object; the ants felt the aphids with extended antennae and did not remain near them for long. An ant behaved in this manner until it accidentally touched a drop of an aphid’s excretion and had to test it when cleaning its antennae or legs. After this, the ant’s behaviour substantially changed: instead of tapping at the aphids the ant began to stroke them with folded antennae, thus asking for the sweet excretion. This change was gradual. At first, the ant only slightly folded the antennae, so that they tapped the aphid’s sides (in normal contact, ants pat aphid’s backs); the movements of the antennae were uncoordinated. We observed this stage in the behaviour of all “beginners”. After successful contact with the first aphid, an ant began to perceive other aphids. The ant stopped them and tried to milk them, lengthening the contact until a drop of excretion emerged. At this stage, the movements of antennae were more coordinated; however, the ant was usually unable to catch the drop in time, and hence, was compelled to clean itself continually (Fig. VII -8). The development of milking behaviour, including the stage of asking and waiting for the drop of excretion, was accomplish 60-90 min after the ants had faced aphids for the first time. The behaviour of naï ve ants during their subsequent contacts with aphids did not differ from the behaviour of control, “wild” ants.

We think that, in this case, innaterecognition of the objects of thespecies-specific instinctive behaviouroccurred. When the ants perceived the stimuli that came from the aphids, innate recognition was complemented with acquired reactions, and all behavioural elements form an integrate behaviour. Apparently, social facilitation (see Part VIII) was also involved in this process. Those ants that were the first to appear in the aphid colony took considerably more time to develop the behaviour. In sum, direct interaction with aphids in ants consists of short innate behavioural sequences which are likely to be the same with those involved into the process of trophallaxis, one of the basic process integrating a family of social hymenopterans into the whole. Whereas the process of aphid milking seems to be completely wired, the division of labour in groups of aphid milkers appeared to be determined by social experience.

It was difficult for us to believe that so complex behavioural pattern as aphid milking is innate in ants. Further, it was very likely that another complex behaviour, hunting for jumping springtails, is learned in those ant species which encounter springtails very often (a good chance to eat them) but lack special morphological structures (snap-on mandibles) intended to facilitate catching springtails in highly specialised tropical Dacetine ants. As it was briefly noted in Chapter 22, we carried out many experiments with “wild” and naive families of Myrmica rubra, abundant inhabitants of soil and litter in the forest and steppe-forest zones. To observe the interaction of the ants with active prey, we put live jumping springtails (Tomocerus sibiricus) into glass containers (6 cm in diameter and 12 cm in height, 30 animals per container). The containers had a gypsum bottom covered with a " splints" made out of plastic bottles (this substrate mimicked forest litter but was transparent). Ants were put into containers individually, transferring them with a small brush. Their responses to the prey were timed and video recorded; the records were then viewed in the slow-motion mode, fixing and drawing in detail individual frames (Reznikova and Panteleeva, 2001). The members of the control families caught jumping springtails quite effectively. For instance, in one series of experiments, 116 of 214 tests ended in catching the prey; in the remaining tests, ants also responded to the springtails aggressively. Successive “springtail hunters” demonstrated searching and sufficiently specific behaviour. These ants moved relatively fast and freely through the bulk artificial litter. Once the ant found itself in the immediate proximity to a springtail, it attacked the prey: bent the abdomen and head to the thorax, jumped to the springtail and fell on it in the similar manner as a fox falls on a mouse (Fig. VII-9). At a final stage of this behaviour sequence the ants differs sufficiently from a fox setting its sting in motion. However, naï ve individuals behaved rather differently. One “Haspar Hauser” ant after another, being at the age of full mature, behaved toward a potential victim in the same friendly manner as if it was its nestmate. We observed numerous antennal contacts of ants with springtails; i.e., ants responded to springtails as to conspecific animals rather than potential prey. If the springtail made occasional jerks, the ant jumped aside. As our laboratory and field observations have demonstrated, it took from several weeks to several months of practice, trials and errors, to build up the character of a successful springtail hunter. At this stage of our study we believed that individual and social experience strongly dominate the development of springtail-hunting in Myrmica ants moreover this species lack specialised morphological adaptations for catching jumping victims. This suggestion agrees with the fact that under natural condition the higher population density of potential victims was fixed in ants feeding territory, the more numbers of successful hunters were encountered within ants’ families.

We nevertheless did not stop our efforts to examine naï ve Myrmica ants in order to reveal more details of the scenario of hunting behaviour (Reznikova and Panteleeva, 2005). At last, in one of experimental families, we found 7 out of 123 naive ants that were able to catch the prey, and all of them exhibited an " at once and entirely" FAP during the final act of the hunt and had no noticeable differences from the adults (Fig VII-10). One of them caught the prey at an age of seven days and the others, at an age of 14 days. In contrast to the control ants, they, however, stayed with their prey on the laboratory arena instead of transporting it to the nest. If we transferred them, together with their prey, to the nest by means of a brush, they left the springtail near the brood. Having found the killed prey, the members of the naive family that stayed in the nest carried it to the remote part of the nest, far away from the brood, and did not use it to feed the larvae. Our observations showed that the broods of the naive families fed on fodder eggs laid by adult worker ants. Thus, the hunting in the young families " ran idle"; i.e., the prey was not used for its intended purpose.

The " at once and entirely" FAP of ants hunting for a jumping springtail, which is so difficult to catch, indicates that the specific stereotype of hunting behaviour may be expressed as an integrated set of behavioural sequences. However, the expression of this stereotype is variable within a family. Only in a small proportion of ants (less than 10%) was hunting behaviour expressed at an early imaginal stage. For comparison, recall that the formation of such a complex behavioural pattern as begging for honeydew from symbiotic aphids was observed in all naive ants within 60-90 min after the first contact with a drop of honeydew. In contrast, the scenario of the formation of interaction with the difficult-to-handle prey requires is based on the multistage maturation and completion of the species-specific FAP, which probably includes elements of social learning (see Part VIII). This example enables us to examine complex behavioural patterns in many members of species investigated because apparent learning by trial and errors can be hardly distinguishable from maturation, and a single exemplar possessing the " at once and entirely" spontaneous behaviour sequence can argue for “instinctive” behaviour.

Very similar results with ours on hunting Myrmica were obtained in the study on tool use in New Caledonian Crows, indicating that one individual can change experimenter’s mind from recognising individual experience to maturation as the main factor guiding complex and apparently intelligent behaviour. We have already described New Caledonian Crows in Part VI as the most prolific avian tool users (see: Hunt, 1996; Kenward et al., 2005). Students of tool using in this species have elaborated a hypothesis about cumulative cultural evolution (Hunt and Gray, 2003; Kacelnik et al., 2004). However, recent development study on a group of four hand raised juvenile individuals (Kenward et al., 2005) showed that one individual spontaneously manufactured and used tools in sophisticated manner, without any contact with adults of its species.

Experimenters raised chicks in artificial nests and subsequently transferred them to enriched aviaries that contained twigs of assorted shapes and sizes, and food items hidden in holes and cervices. None of the subjects was ever allowed to observe an adult crow. Two of them were housed together and were given regular demonstrations by their human foster parents of how to use twig tools to retrieve food. The other two were housed individually and never witnessed tool use. All four crows developed the ability to use twig tools. Although the tutored crows paid close attention to demonstrations, there was no qualitative difference between them and the untutored birds in their tool-oriented behaviour. Experimenters first observed the successive food retrieval from a hole by the tutored birds when they were 68 and 72 days old and by the untutored birds at 63 and 79 days old.

The authors also tested the juvenile’s response to leaves from trees from genus Pandanus, similar to those from which wild crows make tools. The leaves were mounted on wooden frames so that the birds could access them roughly as they would in the wild. On the first day that he was presented with Pandanus, the untutored individually housed male Corbeau (then aged 99 days) produced a straight tool from one side of the leaf by using a swift “cut-tear-cut” action. Immediately after producing the tool, Corbeau carried it to a crevice where food was often hidden and used it as a probe, a sequence that he has since repeated several times to successfully retrieve food.

This finding demonstrates that New Caledonian Crows have inherited characteristics that support tool making and use and possibly guides fast learning processes enabling birds to solve their living problems applying tools. The fact that inherited predisposition can account for a complex behaviour such as tool manufacture highlights the need for controlled developmental investigations in those species that seemingly show intelligible and especially culturally transmitted behaviour. Social learning however may play a significant role in transmitting specific techniques within populations even if it only facilitates maturation of innate behavioural patterns. We will consider this aspect of learning in Part VIII.

24. IMPRINTING

Imprinting, in general sense, is the phenomenon by which many animals form special attachments to objects to which they are exposed during specific, sensitive (and “critical”) periods of their life.

Within the whole spectrum of combinations between innate and learned behaviour imprinting provides an amazing and very curious example of genetic and environmental influence on animal behaviour. Lorenz (1935) used the noun “Prä gung”, or “Imprinting”, to refer to the process of rapid bound-formation early in the life of precocial birds (ducks, geese, and the like). Lorenz himself “imprinted” this idea from his teacher Oskar Heinroth, a German zoologist and ethologist. In his paper (1911) Heinroth described the behaviour of incubator-hatched graylag goslings. These newly hatched birds showed no fear, and attached themselves readily to humans. Such man-attached goslings do not show any inclination to approach and stay with parent-geese; they behave as if people were their parents. This kind of attachment-behaviour was described by Heinroth by the verb “einzuprä gen”, corresponding to the English “to stamp in” or “to imprint”; and the word “stamped in” had earlier been used by Spalding (1873), and by Thorndike (1898) in relation to firmly acquired modes of behaviour. Further imprinting has been found in many species, from ants to humans, and it appeared to include reactions not only for visual but also for acoustic and olfactory stimuli.

Indeed, the first scientific description of the phenomenon of imprinting was done by Spalding (1873). He discovered that newly hatched chicks followed almost any moving figure. Spalding regarded such behaviour as “unacquired, ” that is, instinctive rather than learned. He then concluded that early learning takes over at some stage from instinct; chicks which at first follow instinctively then they learn who their mother or mother-substitute is, and are eventually able to discriminate between her and other figures. Spalding drew attention to a remarkable feature of such early behaviour, namely that a chick would follow its mother, only if it had the opportunity to do so early enough in life. If faced with its mother for the very first time after the opportune or sensitive period has
passed, the chick would fail to follow it and would show no affinity whatever to it; furthermore, it would not subsequently be able to develop any attachment to its mother. We now say that a chick becomes imprinted to the mother-figure when it learns its characteristics, and forms a tie to it. Spalding reported that this development was confined to a short period soon after hatching; and this has since become known, with some degree of justification, as the critical period for imprinting.

Lorenz went further than Spalding in that he specified the characteristics of imprinting that differed fundamentally from
what he called “ordinary learning”, and thereby generated much
interest in it, which, in turn, has resulted in further systematic observations and experimentation in this area of animal behaviour. Precocial chickens turned out to be ideal subject for studying impinting. Lorenz recalled in his Nobel Lecture:

“Selma Lagerlö f's Nils Holgersson was read to me - I could not yet read at that time. From then on, I yearned to become a wild goose and, on realising that this was impossible, I desperately wanted to have one and, when this also proved impossible, I settled for having domestic ducks. In the process of getting some, I discovered imprinting and was imprinted myself. From a neighbour, I got a one day old duckling and found, to my intense joy, that it transferred its following response to my person. At the same time my interest became irreversibly fixated on water fowl, and I became an expert on their behaviour even as a child “.

Lorenz started his systematic study on imprinting in the middle of

1930-s. In 1950-70s many experiments were conducted on imprinting (Guiton, 1959; Gottlieb, 1961, 1971; Bateson, 1966, 1979; Landsberg and Weiss, 1976). Hess (1959) suggested an experimental set up that allowed observing and recording behavioural patterns of ducklings and goslings following different things that imitated their mothers such as balls, boxes, dummy birds of different sizes and colours.
The main and most widely cited Lorenz’s experiment with gosling was the following. Lorenz divided a nest of goose eggs, leaving half with the mother and putting the rest in an incubator. The geese raised by their biological mother showed normal behaviour, following her around during their youth and growing up to interact with other geese. When the incubated eggs hatched, the goslings spent their first few hours with Lorenz instead of their mother. From then on, these goslings followed Lorenz around, showing no recognition of their own mother or even adults of their own species. As adults, these geese continued to prefer humans over members of their own species. Lorenz's experiment showed that geese have no instinct telling them who their mother is, or who is a member of their species. Instead, they respond to and identify with the first moving object they encounter.

Lorenz and his followers marked out several specific features of imprinting. First, imprinting could take place only during a brief critical period in the individual's life; and second, once it had taken place, it could not be reversed. This feature dramatically differs from associative learning that is highly volatile, usually fading with time. Furthermore, imprinting was reported to occur very rapidly, without any trial and error; and, above all, imprinting would show itself, at maturity, in a courtship directed towards the original mother-figure or figures similar to her. One more amazing characteristic of imprinting is that stressful stimuli fortify “stamping in”. If there is an increased level of stress at the time of the original imprinting, the learning paradoxically becomes faster. Hess (1964) demonstrated that if in the laboratory set up obstacles are placed in the runway between the duckling and the followed object then the response the duckling subsequently exhibits is more determined and energetic. It may be that this enhances an individual duck family's level of imprinting at times of greatest need, for instance when the threat of predators or the distraction of other broods is a particular problem.

It was later questioned whether these features would separate imprinting sharply from other forms of learning. Indeed, Lorenz himself, some twenty years after the appearance of his early papers, expressed the view that imprinting might be a type of conditioning. Whether it is, or is not, would depend partly on how narrowly or broadly conditioning is defined.

Several forms of imprinting are distinguished.

Filial/maternal imprinting is a kind of imprinting in which the offspring will follow its mother, or respond to the mother’s calls. The same term is also used with respect to mothers that give birth to offspring: maternal imprinting occurs during a brief period after parturition in which mothers learn to recognise the voice and/or odour of their young.

Filial and maternal imprinting is typical for precocial birds and mammals (that is, those born in a relatively mature state). It also may well be -but this is still somewhat controversial-that imprinting also occurs in altricial species (that is, those that are rather immature when born). Altricial neonates are unlikely or unable to stray from their home base in the first few days of life and therefore do not need the same response as precocial youngsters. They learn similar lessons rather later in life during what are called " socialisation periods". These apply when the animal's sensory, motor and thermoregulatory systems are fully functional and they learn to move away from their mother and to interact with others of the same and other species. The window of opportunity for learning varies in different species. In dogs it is from 3 to10 weeks and in cats from 2 to 7 weeks, while in primates it is usually 6-12 months. Stimuli that the youngsters of each species are exposed to during these window periods will be accepted as “normal” (Immelman, 1972).

Most of the work on imprinting and in particular on filial imprinting has been done with birds because they are convenient research animals. They will imprint on practically anything-balloons, boxes, or even the cover under which an experimenter is hiding so as not to influence the behaviour of his charges (see: Wallace, 1979). Colour and shape do not seem to matter, but the imprinting takes place faster if the object contrasts with the background and if it makes some sort of noise. Klopfer (1959) revealed that young wood ducks actually imprint on sounds rather than on visual stimuli. In an experimental situation, a young wood duck will approach any sound, but later it will follow only familiar sounds. In nature these birds nest in dark holes, so that during the imprinting period they may never even see their mother.

In filial imprinting, once the young individual has formed an attachment to a particular object it avoids novel objects. There are conflicting pressures on the young to readily recognise and to follow its parent but also to recognise and to avoid other adults, as well as heterospecifics that are potential predators (Hinde, 1970). Such conflicting pressures in filial imprinting resemble those involved in sexual imprinting and mate selection, when it is advantageous to distinguish conspecifics from heterospecifics (Irwin and Price, 1999). We will see further that although filial imprinting is separable from sexual imprinting, the processes are similar in many ways.

It is possible that the human infants are tied to their mother by imprinting-like connections. Lorenz (1935) drew attention to certain analogies in human behaviour to the occurrence of imprinting to, or with, inappropriate objects; he had in mind human ways of acting which, as he put it, “appear in the form of pathological fixations on the object of an instinct”. Basing on the ethological concept of imprinting British psychiatrist Bowlby (1969) and American developmental psychologist Ainsworth (1989) elaborated attachment theory of parent-child and other close relationships. According to this theory, an attachment is a strong affectional tie that binds a person to an intimate companion. Human researchers have found many analogies between behaviours in human infants and infants of other primate species at early stages of establishing intimate connection with their mothers.

The dramatic effect of early experience and consequences of disturbance of imprinting-like behaviour in primates were first revealed in the investigations of H.F. and M.K. Harlow (1962). They raised rhesus monkeys without their mothers. The animals had only surrogate mothers, which were either covered with terrycloth or consisted of bare wire mesh. Attached to these models were bottles with nipples, from which the baby monkeys could suck their milk. The monkeys that were raised in this way later proved to be poor mothers. They did not nurse their young or only did so after some time, and even mistreated them being rejective or aggressive. Here an early childhood experience led to substantial disruptions of later social behaviour. In human beings this phenomenon is called hospitalism (Spitz, 1945). Emphasising this dramatic role of early experience, Harlow (1971) called his book about these experiments “Learning to Love”. It turned out a little bit later that there was no need to separate infant monkeys from their mothers for so long in order to observe displays of hospitalism in them. Hinde (1974) carried out much less traumatic experiments with the same species. He allowed mothers and their infants to develop normal bounds and only removed a mother from a group for several days. The young was not left alone, instead, it could enjoy treatment from other females in the group. Infants, however, hardly survive, and the harder the less harmonious were their previous relationships with their mothers, as if they possessed less “margin of safety”. Hind was able to distinguish between normal monkeys and those which had been separated from their mothers in their childhood even after several years.

Recent studies on wild elephants in Africa revealed displays of hospitalism in adolescents caused by seemingly weak changes in their social bonds. Being isolated from their supportive older caretakers (“allomothers’) because of fragmentation of their social environment by humans, young elephants display symptoms associated with human post-traumatic syndrome: abnormal startle response, depression, unpredictable asocial behaviour and hyperagression (Bradshaw et al., 2005).

Sexual/social imprinting is a kind of imprinting in which a young offspring imprints on members of its own sex, and when the offspring mature, they prefer their own species. Sexual imprinting arises as a consequence of learning about individuals and can create mate preferences within species (Bateson, 1966). As we just have seen, sexual and filial forms of imprinting are closely connected and determine in many respects the future of adult’s life. For instance, in duck species in which the sexes have different appearances only the male ducklings imprinting sexually on the mother. In these species, ducklings exposed to adult males during their imprinting period will later form homosexual bonds, even in the presence of females (Schutz, 1965).

Lorenz (1937) related a story about a male bittern which was raised by a zoo-keeper. Although the bittern was maintained with a female of its own species and eventually paired with it, the misimprinted male would drive the female away whenever the zoo-keeper approached and try to get the keeper to come into the nest to incubate the eggs. Thorpe (1956) reported the case of a gander’s alleged seven-year fixation to an oil-drum. Subsequent controlled experiments have confirmed the power of sexual imprinting. Cross-fostering experiments with sexually-imprinting species have succeeded in creating preferences of one species for another. For instance, pigeons raised by doves will later prefer to court other doves, rather than members of their own species (Sluckin, 1965).

It is worth to note that some traits are affected by sexual imprinting more than others. Lorenz noted with some amusement that jackdaws that had imprinted on him would court his favour by presenting him with juicy fresh earthworms and would even attempt to introduce these into his ear-holes. However, when not sexually aroused, these birds would happily join other jackdaws in flight. In sexually dimorphic species (in which the external appearance of males and females differ), sexual imprinting varies depending on whether the youngster is a male or a female. While a male mallard duckling will identify his future mate by relating it to the appearance of his mother (or attachment figure), the same does not apply to a female. For falcons imprinted on humans a combination of human and avian stimuli are required in order to elicit sexual responses.

Sexual imprinting is one of the several known non-genetic, yet social factors which influence mate preference and can facilitate the formation of new species (Bateson, 1978, 1983; Todd and Miller, 1991; ten Cate and Vos, 1999). Whether sexual imprinting can support the evolution of novel traits is still under debate (Hö rster et al., 2000). Laland (1994) designed a model which shows that when there is an asymmetrical mate preference sexual imprinting can support the evolution of novel traits. There is experimental evidence that sexual imprinting on both natural and artificial novel traits takes place. Ten Cate and Bateson (1989) showed that in Japanese quail Coturnix coturnix japonica offspring prefer mates which are slightly differ in plumage character from their parents. Experiments with artificially changing traits demonstrated that Javanese mannikin (Lonchura leucogastroides) males and females became sexually imprinted on a read feather on the forehead as a novel trait in parent birds. Those females imprinted on the red feather showed a strong preference for another novel red trait, red stripes at the tail, similar to mail with the red feather. Females transferred the preference for a red feather to the other novel red trait (Plenge et al., 2000; Witte et al., 2000). Further studies demonstrated that not all kinds of red details of ornament are suitable for birds to be imprinted on, and that red bill is too much. Males and females raised by a red bill father showed even a strong rejection to conspecifics of the opposite sex with a red bill (Hö rster et al., 2000). Seemingly, in other experiments females showed no preference for males with novel blue traits (Plenge et al., 2000). Thus, not all kinds of novel traits birds can be sexually imprinted on.

The formation of social attachment, that is, social imprinting, usually requires the presence of conspecifics at early age. For instance, dogs pass through a critical period for the development of social relationships during weeks 4 and 6 (Scott, 1992). During this time they form a close social bond to conspecifics or to man as a substitute, regardless of whether they are punished, fed, or treated indifferently. Scott and Fuller (1965) have emphasised that the internal processes on which this readiness for contact is based seems to be more important than the external factors.

There is a growing body of evidence that domestic animals include humans into their social environment and thus treat them as conspecifics (Broom, 1981). I have been impressed by one slide from Altbä ker (2004) presentation at 2^nd European Conference on Behavioural Biology that contained a bill “Beware of rabbits” together with a photo of a furious male rabbit trying to drive a human away from its territory. Rabbits are highly territorial animals. In Altbä ker’s experiments fear of humans has been eliminated in rabbits by exposing infant rabbits to the smell of humans at early critical periods. Early experience interfering with conspecific recognition resulted in rabbits fearless to humans. They gamely defended boundaries of their territories from “giant rabbits” as they considered human intruders.

The knowledge about imprinting has practical applications in farming, veterinary and breeding practice. Chinese peasants have for centuries capitalised on the tendency to imprint in making ducks more effective in controlling snails that otherwise damage rice crops. By imprinting ducklings onto a special stick, the peasants can not only take their brood out to the paddy fields as required but, by planting the stick sequentially in different parts of the plantation, they can ensure that molluscs in all areas are imprinted for predation. Hawks are also subjects of sexual imprinting: falconers take advantage of this by wearing special hats to collect the sperm of their amorous human-imprinted chargers for use in breeding programs. Some farms are practicing imprinting of foals at birth to acquaint them early with such things as halters, clippers, and trailers. Exposing a foal to these objects and practices within the first 45 minutes of its life demonstrates to the foal that certain sights and sounds offer no danger. Here farmers are guided by ethological principles of “imprint training” suggested by Miller (1991). However, experimental investigations showed that foals hardly change their behaviour against their current behavioural stereotypes. Early handling that began 2 hours (Williams et al., 2002) and 5-7 (Lansade et al., 2005) hours after birth appeared to have only short-term effect on foal’s reactivity and manageability. It appeared to be easier to shape foal’s behaviour toward humans through positive human’s interactions with dams. It is likely that observations of mother by foals during early stages of life facilitate human-foals relationship (Henry et al., 2005).

Filial and sexual imprinting forms a basis for bond formation as well as for species and sex recognition. It may be easier to appreciate the role of imprinting in animals’ self-determination if we consider species that do not necessarily meet conspecifics in their development. Interspecific brood parasites (such as some cuckoos and brown-headed cowbirds), and megapodes are good examples. For brood parasites, several mechanisms of sexual and social self-determination have been hypothesized. It was suggested that the timing of imprinting is delayed until the young leave the nest and interact with conspecifics, that innate recognition of a conspecific’s features is more important than imprinting, and that both genetic factors and learning are involved in the development of social attachments (review in: Gö th and Jones, 2003). As it is known, brood parasites are mostly altricial and are initially confined to the nest where they cannot actively seek contact with conspecifics. Conversely, a megapode chick, being a typical representative of a precocial bird, might possess imprinting. However, it seems to be too much advanced. As Gö th and co-authors (Gö th and Jones, 2003; Gö th and Evans, 2004) have revealed by observations of incubated Australian brush-turkeys Alectura lathami, these “packing” chicks show “social behaviour without social experience” and this requires no postnatal learning. As we have seen above in this Chapter, megapode chicks hatch underground in mounds of leaf litter, where external heat sources incubate the eggs, and then they dig their way to the surface to lay themselves the road to completely independent life. Chicks hatch asynchronously and receive no parental care, so imprinting cannot occur (Gö th and Evans, 2004). Wong (1999) found that megapode chicks do not imprint on moving objects after hatching like chicks of other galliforms do. Nor do they have a tendency to aggregate with other chicks. They are capable of finding adequate food alone and of detecting predators innately, and these features should enable them to survive without assistance from others (Gö th, 2001).

Whereas young megapodes display independence from an early age, young ants aspire to tight social bonding with their nestmates. As many other social animals, ants possess social imprinting and this enables them to occupy an appropriate place in their society. The most striking example of imprinting in ants is the learning of colony odour (see: Hö lldobler and Wilson, 1990, for a review). The period of greatest sensitivity is usually within first few days after eclosion of the young from the pupae, although in some species the learning can begin during the larvae stage. There is some evidence that when no contact with nestmates is permitted during the sensitive period, later social behaviour can be seriously impaired.

All these examples show imprinting as a very specific innate form of learning which is likely to be under hard press of natural selection in many species. To put studying of imprinting in the context of animal intelligence, it should be noted that in some cases imprinting appears to be an initial stage of a long lasting specific learning rather than a shortly restricted event in animal’s life.

For example, Charrier et al.’s (2003) experimental field study of mutual recognition in mothers and pups of fur seals Arctocephalus tropicalis has revealed surprisingly long-term recognition. In pinnipeds, and especially in otariids, mothers and pups stay within close range during about a year and they develop a strong capacity to recognize each other’s voices. Pups become able to discriminate their mother’s voice a few days after birth. For females this discrimination seems to occur earlier, probably during the few hours after parturition. However, during lactation mothers are confronted with a major problem: the change of characteristics of their pup’s call caused by maturation of their brains and vocal tracts. During lactation period mother seals have to alternate suckling periods with foraging trips at sea and periodically leave their pups alone in the colony for 2-3 weeks. Mother-pups vocal recognition system is particularly effective, since, in spite of the high risk of confusion due to the great density of individuals in the colony, mother-pup pairs have been shown to meet in less than 11 min when the mothers return from foraging trip in sea. Researchers carried out playback experiments at the end of the rearing periods to test whether mothers can recognise the changing voices of pups. They presented females with recorded signals of their own pups at different age, as well of other pups. The tests were carried out when mothers and pups were separated in the colony. Playback experiments demonstrated that females remember and still recognise all the successive immature calls of their own pups and easily distinguish them from other pup’s calls. Thus, mothers must be capable of permanently learning their pup’s vocal characteristics. In the northern fur seal Callorhinus ursinus, another otariid species with a similar pattern of maternal attendance, mothers and pups are able to mutually recognise their calls for at least four years (Insley, 2000). This may be redundant regarding adaptive behaviour, and initial impulse of long lasting memorisation is likely to be originated by imprinting.

Recent studies have brought many new facts and ideas concerning imprinting, and at the same time they make boundaries of the concept of imprinting fuzzier and raise new questions and problems. There are compelling evidences that the social bond that develops through imprinting entails an addictive process that is mediated by the release of endorphins, the brains’ own opiates (Hoffman, 1996).

Bateson (2000; se also: Bateson and Martin, 2000) formulates a model in which imprinting is not an instantaneous and irreversible process but a much more flexible and less peculiar phenomenon. Imprinting does not necessarily occur immediately after birth but has a more flexible sensitive period affected by both experience and species-specific features. Imprinting is not a monolithic capability but is composed by several linked processes: (1) detection of a relevant stimulus guided by predisposition to what the animal will find attractive; (2) recognition of what is familiar and what is novel in that stimulus, which involves a comparison between what has already been experienced and the current input; (3) control of the motor patterns involved in imprinting behaviour.

Although imprinting can be functionally distinguished from learning involving external reward, both types of learning are deeply connected, as suggested by the possibility of transfer of training after imprinting.

In both filial and sexual imprinting animals can make associations between multiple traits that distinguish individuals or species. It is still an opened question about what traits are really learned and why some traits are not learned during imprinting (Witte et al., 2000; Hö rster et al., 2000). It is possible that characters of species-specific predisposition to establish strong attachment links basing on attractive traits works like “filter wheels” so that one traits loose attractiveness and others take on special significance during certain periods of animals’ life. Not all “filters” are compatible with evolutionary established tendencies so some traits have more power than other to be imprinted on, and some traits can not be imprinted at all.

CONCLUDING COMMENTS

Ecological and evolutionary aspects of animal intelligence are promising and fascinating topics of research. It is a challenging problem to understand specificity and limits of wild minds. Classic ethologists considered “mature versus nurture” actually a false dilemma: in almost all animals behaviours there is a mixture of both. Indeed, simple " nature/nurture" or " instinctive/learned" dichotomies have now been abandoned, and attention is now focussed on experimental investigation of what does and what does not influence behavioural development. Modern researchers are interested in how genetic and environmental factors interact.

There are, however, new problems in this field aroused by recent experimental studies. We have already met a great deal of examples that members of many species develop extraordinary abilities to solve their problems on the basis of very specific stereotyped behaviour that was shown to be innate. Does this mean that cognition in animals can be imagined as a thin layer of gloss on inherited species-specific stereotypes? Or do animals enjoy flexible and creative cognitive skills? Or should we seek a compromise between these alternatives?

Summarising the data concerning innate and learned rules that establish strict boundaries for flexibility in animal behaviour, we can conclude that even those signs of behaviour that are considered to be unique characteristics of instincts in classic ethology, may be adhere to learning on closer examination. Instinctive acts are often triggered by simple key stimuli but we know now that animals can learn to pick out some simple stimuli and ignore others in changeable environment. Wired stereotypic behavioural patterns similar in all members of species can actually be learned stereotypes just because animals pick out the same simple stimuli and learn the same. Indeed, this is not all the truth about the relative importance of genetics versus environment in animal behaviour

Here I am not trying to analyse the balance between nature and nurture in animal behaviour in general. Instead, I only concentrate on the problem of development of complex behavioural patterns. Analysing different aspects of the findings described above we can conclude that complex behaviour can develop on the basis of specific behavioural domains. Within these frames animals demonstrate sophisticated mental abilities and high flexibility of memory. In some cases a compromise between advanced intelligence and innateness can be found if we consider behaviour within the frame of the paradigm of “adaptation learning” (Reznikova and Dorosheva, 2005). That is, in order to solve vital problems animals are likely to choose more and more quickly and effectively a relevant strategy from a set of strategies that already exists rather than to contrive something completely new as an answer for a challenge from their environment.

Being an advocate of animal intelligence, I nevertheless would like to emphasise that inherited predisposition can account for different complex behaviours such as food hoarding, hunting, tool manufacture, communication, and even Machiavellian interaction with conspecifics. It seems that there are only narrow roads for displays of flexibility in animal behaviour. To clarify the question of interactions between inherited traits and individual and social learning it is necessary to consider social inter-relations and “language” behaviour. The question of flexibility of communication systems seems to be crucial in consideration of animal intelligence. We will discuss these problems further in this book.