Attempts to trace particular patterns of behaviour back to the early action of certain genes, or to particular kinds of early experience, are often misconceived because of profound changes that occur at certain stages in development. Early influences may not necessarily exert detectable long-term effects on behaviour because of major changes in the organisation of behaviour that have occurred in between. Such a possibility is, of course, in stark contrast to traditional views of development, which tended to emphasise the important and far-reaching consequences of all events that occurred early in life.

The control of behaviour patterns and their biological functions are likely to change as development proceeds. While caution is needed when interpreting changes with age in terms of reorganisation of behaviour, activities that look the same at different ages may be controlled in different ways and may have different functions. The time a kitten spends in contact with its mother, for example, is influenced primarily by its need for milk early in life and by its need for comfort later. Some activities, such as suckling, are special adaptations to an early phase and drop out of the repertoire as the individual becomes nutritionally independent of its mother. Similarly, certain motor patterns and reflex responses that are present at birth have disappeared from the behavioural repertoire by the time the cat is a few weeks old.

Weaning
At around the time of weaning, towards the end of the second month, play changes markedly in character. The frequency with which kittens play with inanimate objects increases sharply at around 7-8 weeks of age, and many measures of play before this age do not predict the same measures in the same individuals at 8-12 weeks, after weaning is over.

Correlations between different measures of social play also break down at the end of weaning, as do correlations between some measures of predatory behaviour. Certain measures of social play become increasingly associated with some measures of predatory behaviour during the third month. This might indicate that motor patterns come under the control of new motivational systems as the kitten develops, some becoming controlled by the same factors that control predatory behaviour, and others by the factors controlling agonistic behaviour. Some playful motor patterns become increasingly associated with patterns of predatory behaviour, and some become associated with agonistic social behaviour.

In passing, it is worth pointing out that the different developmental time courses and general lack of intercorrelations between measures of social play and measures of object play indicate that these two forms of play are separately organised and separately controlled. Even in terms of the motor patterns used, object and social play differ distinctly in a number of respects; for example, repetition of certain motor patterns occurs frequently during object play but seldom during social play.

Tara Montgomery

Zoe, at 10 weeks, developing play behaviour.

Playful predators
Cats are, of course, formidable hunters and many of the motor patterns that appear in play resemble those used in catching and killing prey. Not surprisingly, many hypotheses about the function of play in cats have invoked links between play and later predatory behaviour, with play seen as a form of practice for adult predatory skills. However, little hard evidence has yet been produced to support this view. Play experience is most certainly not necessary for at least the basic elements of predatory behaviour to develop. For example, Thomas & Schaller ('Das Spiel der optisch isolierten Kaspar-Hauser-Katze,' in Naturwissenschaften, vol. 41 pp. 557-8, 1954) reported that 'Kaspar Hauser' cats which were reared in social isolation and without opportunities for visual experience, let alone play behaviour, nonetheless showed 'normal' predatory responses when presented with a prey-like moving dummy at 11 weeks of age.

However, the possibility remains that play may have subtle beneficial effects on predatory skills. The one experimental test of this hypothesis so far carried out failed to find any relations between early object play experience and later predatory skills in domestic cats. Cats which had no opportunities for playing with small, inanimate objects when growing up did not subsequently differ from kittens which had regularly played with objects, when their predatory skills were measured at six months of age (Caro, T. M. 'Effects of the mother, object play and adult experience on predation in cats,' Behavioral and Neural Biology, vol. 29 pp. 29-51, 1980). This failure to find an effect might have been due to insufficient differences in the experience of the normal and the deprived groups of cats, or to measures of predatory behaviour that were insufficiently fine-grained to pick up genuine differences in skill. Furthermore, the benefits of play may be missed, because a single experience of catching and eating a mouse can be enough to make a kitten a skilled mouse-killer thereafter. For all these reasons, the role of play in behavioural development continues to generate much discussion.

Later developments
Despite these indications that not all aspects of development are continuous, it is clear that many types of early experience can be related to what happens later in ontogeny. For instance, many measures of predatory behaviour at 1-3 months of age are positively correlated with the same measures taken at six months. Individual differences in behaviour early in development can, to some extent, predict individual differences later in life. Laboratory studies suggest that cats' choice of prey and their adult food preferences are strongly influenced by experience with their mothers when young. For example, cats are more likely to kill prey species with which they are familiar from experience as kittens. Similarly, cats which have had experience with a particular type of prey when young are more skilful at catching and killing the same type of prey when adult. This effect of early experience appears to be specific, in that early experience with one type of prey does not produce a general improvement in predatory skills when other prey species are considered.