Peter H Klopfer. American Scientist. Volume 84, Issue 4. Jul/Aug 1996.
Experimental studies of socialization in mammals date back to the 1950s. Much of the early work was modeled on the classic naturalistic studies of socialization and hen-chick relations in birds conducted by the European ethologists, principally by Konrad Lorenz and his associates. Beginning in the 1960s, my students and I also began to examine some of the mechanisms underlying the processes these investigators described. We were particularly interested in the process given the name Praegung, or imprinting. Imprinting is a metaphor: A blob of hot sealing wax can receive an imprint only for a short period before it cools and hardens. In the same way, it was thought, an infant might develop a stable and lasting preference for a parent model during an early “critical period.”
When we undertook these studies, we skipped happily from Lorenz’s chicks and geese to goats and elephant seals. Whatever our model, we had few inhibitions about extrapolating freely across taxa. Obviously, we believed, what we were studying had relevance to the human condition-why else would the National Institute of Mental Health have so generously supported us? My purpose here is to disabuse everyone, should that be necessary, of the appropriateness of our assumption. Was I ever heard extrapolating human maternal behavior from that of goats or lemurs? I was? Well, then, this is a mea culpa.
The work described below was performed at the Duke University Behavior Station and Primate Center over a period of three decades by more of my students and associates than I have time to list, except for the most significant contributors, who were Jack Hailman, David Gubernick, Cathy Dainis, Richard Hemmes and Leon Rosenson. They do not, however, share the blame for my casual extrapolations.
In Figure 1, an adult goat is nuzzling a newborn kid. Our reports incorporated many similar figures of other species. This adult is the infant’s own mother. Other adults, including those with infants of equal age, are generally unlikely to provide care for alien infants, at least in most of these species; indeed, a mother goat may even respond with lethal aggression toward a youngster not its own. In the case of my favorite subject, the domestic Toggenburg goat, both virgin and pregnant does behave hostilely toward kids, although their behavior can be very rapidly transformed.
Under normal circumstances a Toggenburg doe’s transformation from aggressor to nurturer happens with parturition: As the kid’s head emerges, its mother becomes transformed into a caring, solicitous animal. Even if the mother is given but a few minutes of contact with her infant before it is removed from her, the transformed state is maintained for some hours. The infant can be returned to her later, and she will accept it. What is more, she will accept its siblings, even when these have not had any contact with her-having, for instance, been enclosed in toweling and whisked away by the ever-present experimenter.
Alien infants will be rejected, however. And if a doe is altogether denied contact with any of her newborn during the first few minutes of their life, we found, her behavior is not transformed: She remains implacably hostile to the approach of those bothersome kids. Parallels with the process of imprinting in waterfowl are apparent, although here it appears that it is the mother that is fixated onto the kid, rather than the other way around: The kids are blithely indiscriminate in whom or what they approach to nurse. It also appears that two separate processes are involved, one that activates the female’s interest in the newborn, the other that allows for a discrimination between own and alien young (Klopfer, Adams and Klopfer 1964; Gubernick and Klopfer 1988).
Looking for Love
Our studies first involved determining the temporal limits to this transformation of the mother animal from careless to caring, and to the discrimination that she made between different youngsters. We then focused on the underlying mechanisms that control this system and quickly established that the critical event was the cervical dilation that took place in the final stages of parturition. By dilating the cervix manually Hemmes did this by inserting hydraulically inflatable balloons-we could even induce virgin females to adopt kids that, minutes before, they had attacked (Hemmes, unpublished thesis).
Earlier work by S. J. Folley and G. S. Knaggs (1965) had established that cervical dilation, at least in sheep, induced the release of the hormone oxytocin from the hypothalamus. Since the time course of oxytocin in the bloodstream is strikingly similar to the time for attachment of mother to infant-after 5 minutes oxytocin levels have dropped by almost 50 percent from their peak at parturition-we focused on the role of bloodborne factors. Hemmes exchanged blood between a virgin goat and one in labor and observed that with about one-quarter of the blood exchanged, the virgin animal began to display interest in kids. Injection of exogenous oxytocin had no discernible effect, but some years later Cort Pedersen transformed indifferent female and male rodents into maternally active animals by injecting minute quantities of oxytocin into the third ventricle of the brain, just above the hypothalamus. The response he elicited was specific to oxytocin as well as to that particular region of the brain (Pedersen and Prange 1979). We began to exult that we had found the hormone for mother love. I overconfidently titled a 1971 American Scientist article “Mother Love: What Turns It On.”
More recent work, on sheep, by K. M. Kendrick, E Levy and E. B. Keverne (1992), has provided evidence of specific changes in neural activity at parturition. By chronically implanting electrodes and dialysis tubes into the olfactory lobe, Kendrick, Levy and Keverne could track the changes in neurotransmitter release and electrical activity and correlate these to behavioral changes occurring at birth. Perhaps it is these changes that oxytocin induces, changes that presumably reflect alterations in sensory thresholds at the time of parturition (making olfactory inputs available that were previously not perceived) as well as changes in disposition.
Evolution, Behavior, and Day Care
It will not have escaped you that this picturethe release of oxytocin brought about by dilation of the cervix, followed by changes in the sensitivity and activity of neurons of the olfactory bulb and, in turn, the recognition of neonatal animals and discriminations between own and alien ones-was built up from studies of sheep, goats and rats. We simply selected the most convenient animal for a particular experiment or hypothesis and assumed results from one applied to the other. How risky is this?
Since Darwin’s time, it has been legitimate to expect continuities in behavior as much as in structures. Indeed, in his notebooks, Darwin even seems to give primacy to behavioral continuities. In the 1900s and beyond, Charles O. Whitman, Oscar Heinroth, Alexander Petrunkevitch-illustrious names in the history of biology-used behavioral attributes to establish systematic relations, arguing that the nervous system was maximally conserved in evolution, and thus its output, behavior, would more conservatively reflect relationships than would feeding appendages.
A nice example is Ted C. Schneirla’s description of courtship in empeid flies. The male is prey for the female, so when the male wishes to copulate it must place itself at risk. However, a distracting courtship ritual has evidently evolved that protects the male. In one species a prey favored by the female is captured, immobilized and presented to be eaten while the male mates. In another, the prey is wrapped in silk. The parcel is offered to the female, who is then available for a more leisurely mating while engaged in unwrapping and devouring the contents of the parcel. A third species saves itself the bother of finding prey and merely envelops a piece of detritus in silk, fooling the female while mating takes place. In a fourth, merely the empty silk wrappings serve to distract the female. This behavioral sequence suggests the evolutionary history and relations of these species (after Schneirla, in Klopfer and Hailman 1967).
The problem such reasoning poses is perhaps best illustrated by a tale the eminent psychologist Dan Lehrman once told me. At Congressional hearings that were to decide the fate of federal childcare centers, opponents of such facilities produced witnesses from Harry Harlow’s lab, where the early work on the effect of rearing motherless infants with surrogates had been performed, using rhesus macaques. The result of such rearing was severely pathological behavior on the part of the infants, which persisted for many years. Ergo, surrogate parenting was to be avoided.
Other considerations aside, Lehrman pointed out, had the evidence been drawn from work with the closely related pigtail macaque rather than rhesus macaques, quite a different argument might have been made: Pigtails adapt to substitute parents. This difference is presumably related to the difference in the conditions under which the two species have evolved. But how can one distinguish between behavior that represents evolutionary conservatism from that which responds rapidly to changes in environmental conditions?
Congruity and Context
Present-day studies of systematics and evolution generally adopt a more sophisticated set of assumptions than those underlying the Darwinian texts to which I’ve been alluding. An example is the analysis of courtship behavior in a group of Australian parrots, which has been the subject of study by one of my students, John Rowden. What John is doing is to derive a tree of relationship that is independent of behavior and is instead based on DNA similarities. Onto this he maps the elements of the courtship behavior he has described, as well as differences in the habitat requirements of each species. Congruities in the mappings suggest which behavioral elements are adaptations to ecological conditions and which represent evolutionary continuities.
This method is obviously applicable only to groups of related species or genera, but there is sufficient interspecific variation in maternal behavior that we could expect to learn a great deal about what constrains it and how it evolves. Some years ago, I described subtle differences in the time course of maternal care and other details on mother-infant relationships in three species of lemur (Klopfer and Klopfer 1970); now these three are regarded as separate genera, making a mockery of our conclusions. T. T. McCabe and B. D. Blanchard in 1950 and more recently David Gubernick, a former postdoctoral student in my lab, provided similar pictures of three species of Peromyscus. One builds an elaborate nest, whereas the others create much simpler structures, among other differences. They also relied merely on suppositions when they interpreted these differences. Had they and we applied Rowden’s methods, we could have told a more convincing tale.
Even within our own species, Homo sapiens, seemingly inevitable outcomes of early experiences may prove to be inevitable only in certain contexts. Imagine physically restraining a newborn infant, preventing it from as much as wiggling its toes, and releasing it from such restraints for only a few hours each day, and imagine continuing this treatment for many months. Would you not expect a variety of pathologies to develop? Yet in the context of Navaho culture, where this practice is believed to protect infants from the frustration of being unable to control fine motor behavior, no pathologies appear (Chisholm, as discussed in Klopfer 1988).
Certainly, as a biologist of the Darwinian stamp, I believe there are commonalities across cultures, across species, and across all higher taxa, including commonalities in the physical and physiological substrates that underlie maternal care. However, the neurons and hormones that are implicated may express their outputs very differently. The scene recorded by a camera depends not merely on its optics, but also on the film and the photographer.