Suomi learned his trade as a student and protégé of, and then a direct successor to, Harry Harlow, one of the 20th century’s most influential and problematic behavioral scientists. When Harlow started his work, in the 1930s, the study of childhood development was dominated by a ruthlessly mechanistic behavioralism. The movement’s leading figure in the United States, John Watson, considered mother love “a dangerous instrument.” He urged parents to leave crying babies alone; to never hold them to give pleasure or comfort; and to kiss them only occasionally, on the forehead. Mothers were important less for their affection than as conditioners of behavior.
With a series of ingenious but sometimes disturbingly cruel experiments on monkeys, Harlow broke with this cool behavioralism. His most famous experiment showed that baby rhesus monkeys, raised alone or with same-age peers, preferred a foodless but fuzzy terrycloth surrogate “mother” over a wire-mesh version that freely dispensed meals. He showed that these infants desperately wanted to bond, and that depriving them of physical, emotional, and social attachment could create a near-paralyzing dysfunction. In the 1950s this work provided critical evidence for the emerging theory of infant attachment: a theory that, with its emphasis on rich, warm parent-child bonds and happy early experiences, still dominates child-development theory (and parenting books) today.
In the years since Suomi took over Harlow’s Wisconsin lab as a 28-year-old wunderkind, he has both broadened and sharpened the inquiry Harlow started. New tools now let Suomi examine not just his monkeys’ temperaments but also the physiological and genetic underpinnings of their behavior. His lab’s naturalistic environment allows him to focus not just on mother-child interactions but also on the family and social environments that shape and respond to the monkeys’ behavior. “Life in a rhesus-monkey colony is very, very complicated,” Suomi says. The monkeys must learn to navigate a social system that is highly nuanced and hierarchical. “Those who can manage this, do well,” Suomi told me. “Those who don’t, don’t.”
Rhesus monkeys typically mature at about four or five years and live to about 20 in the wild. Their development parallels our own at a fairly neat 1-to-4 ratio: a 1-year-old monkey is much like a 4-year-old human being, a 4-year-old monkey is like a 16-year-old human being, and so on. A mother typically gives birth annually, starting at around age 4. Though the monkeys copulate all year, the females’ fertility seasons are only a couple of months long. Since they tend to occur together, a troop usually produces crops of babies that have same-age peers.
For the first month, the mother keeps the baby attached to her or within arm’s reach. At about two weeks, the baby starts to explore, at first within only a few feet of its mother. These forays grow in frequency, duration, and distance over the next six to seven months, but rarely do the babies pass out of the mother’s sight line or earshot. If the young monkey gets frightened, it scampers back to the mother. Often she’ll see trouble coming and pull the infant close.
When the monkey is about eight months old—a rhesus preschooler—its mother’s mating time arrives. Anticipating another child, the mother allows the youngster to spend more and more time with its cousins, with older siblings in the maternal line, and with occasional visitors from other families or troops. The youngster’s family group, friends, and allies still provide protection when necessary.
A maturing female will stay with this group all her life. A male, however, will leave—often under pressure from the females as he gets rowdier and rougher—when he’s 4 or 5, or roughly the equivalent of a 16-to-20-year-old person. At first he’ll join an all-male gang that lives more or less separately. After a few months to a year, he’ll leave the gang and try to charm, push, or sidle his way into a new family or troop. If he succeeds, he becomes one of several adult males to serve as mate, companion, and muscle for the several females. But only about half the males make it that far. Their transition period exposes them to attacks from other young males, attacks from rival gangs, attacks from new troop members if they play their cards wrong, and predation during any time they lack a gang’s or troop’s protection. Many die in the transition.
Very early in his work, Suomi identified two types of monkeys that had trouble managing these relations. One type, which Suomi calls a “depressed” or “neurotic” monkey, accounted for about 20 percent of each generation. These monkeys are slow to leave their mothers’ sides when young. As adults they remain tentative, withdrawn, and anxious. They form fewer bonds and alliances than other monkeys do.
The other type, generally male, is what Suomi calls a “bully”: an unusually and indiscriminately aggressive monkey. These monkeys accounted for 5 to 10 percent of each generation. “Rhesus monkeys are fairly aggressive in general, even when young,” Suomi says, “and their play involves a lot of rough-and-tumble. But usually no one gets hurt—except with these guys. They do stupid things most other monkeys know not to. They repeatedly confront dominant monkeys. They get between moms and their kids. They don’t know how to calibrate their aggression, and they don’t know how to read signs they should back off. Their conflicts tend to always escalate.” These bullies also score poorly in tests of monkey self-control. For instance, in a “cocktail hour” test that Suomi sometimes uses, monkeys get unrestricted access to a neutral-tasting alcoholic drink for an hour. Most monkeys have three or four drinks and then stop. The bullies, Suomi says, “drink until they drop.”
The neurotics and the bullies meet quite different fates. The neurotics mature late but do okay. The females become jumpy mothers, but how their children turn out depends on the environment in which the mothers raise them. If it’s secure, they become more or less normal; if it’s insecure, they become jumpy too. The males, meanwhile, stay within their mothers’ family circles an unusually long time—up to eight years. They’re allowed to do so because they don’t make trouble. And their longer stay lets them acquire enough social savvy and diplomatic deference so that when they leave, they usually work their way into new troops more successfully than do males who break away younger. They don’t get to mate as prolifically as more confident, more assertive males do; they seldom rise high in their new troops; and their low status can put them at risk in conflicts. But they’re less likely to die trying to get in the door. They usually survive and pass on their genes.
The bullies fare much worse. Even as babies and youths, they seldom make friends. And by the time they’re 2 or 3, their extreme aggression leads the troop’s females to simply run them out, by group force if necessary. Then the male gangs reject them, as do other troops. Isolated, most of them die before reaching adulthood. Few mate.
Suomi saw early on that each of these monkey types tended to come from a particular type of mother. Bullies came from harsh, censorious mothers who restrained their children from socializing. Anxious monkeys came from anxious, withdrawn, distracted mothers. The heritages were pretty clear-cut. But how much of these different personality types passed through genes, and how much derived from the manner in which the monkeys were raised?
To find out, Suomi split the variables. He took nervous infants of nervous mothers—babies who in standardized newborn testing were already jumpy themselves—and gave them to especially nurturing “supermoms.” These babies turned out very close to normal. Meanwhile, Dario Maestripieri of the University of Chicago took secure, high-scoring infants from secure, nurturing mothers and had them raised by abusive mothers. This setting produced nervous monkeys.
The lesson seemed clear. Genes played a role—but environment played an equally important one.
When tools for the study of genes first became available, in the late 1990s, Suomi was quick to use them to more directly examine the balance between genes and environment in shaping his monkeys’ development. He almost immediately struck gold, with a project he started in 1997 with Klaus-Peter Lesch, a psychiatrist from the University of Würzburg. The year before, Lesch had published data revealing, for the first time, that the human serotonin-transporter gene had three variants (the previously mentioned short/short, short/long, and long/long alleles) and that the two shorter versions magnified risk for depression, anxiety, and other problems. Asked to genotype Suomi’s monkeys, Lesch did so. He found that they had the same three variants, though the short/short form was rare.
Suomi, Lesch, and NIH colleague J. Dee Higley set about doing a type of study now recognized as a classic “gene-by-environment” study. First they took cerebral spinal fluid from 132 juvenile rhesus monkeys and analyzed it for a serotonin metabolite, called 5-HIAA, that’s considered a reliable indicator of how much serotonin the nervous system is processing. Lesch’s studies had already shown that depressed people with the short/long serotonin-transporter allele had lower 5-HIAA levels, reflecting less-efficient serotonin processing. He and Suomi wanted to see if the finding would hold true in monkeys. If it did, it would provide more evidence for the genetic dynamic shown in Lesch’s studies. And finding such a dynamic in rhesus monkeys would confirm their value as genetic and behavioral models for studying human behavior.
After Suomi, Lesch, and Higley had grouped the monkeys’ 5-HIAA levels according to their serotonin genotype (short/long or long/long, but not short/short, which was too rare to be of use), they also sorted the results by whether the monkeys had been raised by their mothers or as orphans with only same-aged peers. When their colleague Allison Bennett charted the results on a bar graph showing 5-HIAA levels, all of the mother-reared monkeys, no matter which allele they had, showed serotonin processing in the normal range. The metabolite levels of the peer-raised monkeys, however, diverged sharply by genotype: the short/long monkeys in that group processed serotonin highly inefficiently (a risk factor for depression and anxiety), whereas the long/long monkeys processed it robustly. When Suomi saw the results, he realized that he finally had proof of a behaviorally relevant gene-by-environment interaction in his monkeys. “I took one look at that graph,” he told me, “and said, ‘Let’s go pop some champagne.’”
Suomi and Lesch published their results in 2002 in Molecular Psychiatry, a relatively new journal about behavioral genetics. The paper formed part of a surge of gene-by-environment studies of mood and behavioral disorders. That same year, two psychologists at King’s College, London, Avshalom Caspi and Terrie Moffitt, published the first of two large longitudinal studies (both drawing on life histories of hundreds of New Zealanders) that would prove particularly influential. The first, published in Science, showed that the short allele of another major neurotransmitter-processing gene (known as the MAOA gene) sharply increased the chance of antisocial behavior in human adults who’d been abused as children. The second, in 2003 and also in Science, showed that people with short/short or short/long serotonin-transporter alleles, if exposed to stress, faced a higher-than-normal risk of depression.
These and dozens of similar studies were critical to establishing the vulnerability hypothesis over the last few years. Yet many of these studies also contained data that supported the orchid hypothesis—but went unnoticed or unremarked at the time. (Jay Belsky, the child-development psychologist, has recently documented more than two dozen such studies.) Both of Caspi and Moffitt’s seminal papers in Science, for example, contain raw data and graphs showing that for people who did not face severe or repeated stress, the risk alleles in question heightened resistance to aggression or depression. And the data in Suomi and Lesch’s 2002 Molecular Psychiatry paper, in which peer-reared monkeys with the risky serotonin-transporter allele appeared to process serotonin inefficiently, also showed that mother-reared infants with that same allele processed serotonin 10 percent more efficiently than even mother-raised infants who had the supposedly protective allele.