The Trouble with Testosterone

by Robert M. Sapolsky

  Intrigued, Wrangham sent samples of the plant for analysis to Eloy Rodriquez, a biochemist at the University of California in Irvine. Rodriquez discovered that the leaves contain thiarubrine-A, a reddish oil known to be a potent toxin against fungi, bacteria, and parasite nematodes. Indeed, he noted that the youngest Aspilia leaves, much the preferred kind among the chimpanzees, are far richer in the oil than the older leaves are. And he thought it significant that the apes kept the drug under their tongues before gulping it down: the tongue is dense with minute blood vessels, and so a substance held there can bypass the digestive juices and head straight into the circulation. (Physicians and victims of angina are well aware that nitroglycerin placed under the tongue can rapidly suppress chest pain.)

  On the basis of his laboratory studies, plus Wrangham’s observations, Rodriquez estimated that in each eccentric bout of Aspilia feeding, the chimpanzees consumed just enough thiarubrine-A to kill between 70 and 80 percent of the parasites in their digestive tracts, without harming other, useful intestinal bacteria. Wrangham and Rodriquez also noted that Tanzanian tribespeople have long chewed Aspilia leaves to heal stomachaches and a variety of other ills. Hence the two coworkers speculated that the chimpanzees were using the plant as medicine.

  In the decade since that work was published, reports of self-medication by chimpanzees, and other primates, have become all the rage. Such is the vogue in academic circles that these studies travel under the imposing name zoopharmacognosy; a few years ago the American Association for the Advancement of Science even convened a symposium on the subject. And certainly, some compelling data were presented, including a report by the anthropologist Karen B. Strier of the University of Wisconsin in Madison. Strier has spent some ten years studying the Brazilian muriqui, or wooly spider monkey, one of many endangered species of the rain forest. Just before the onset of their mating season, Strier notes, muriquis seem hellbent on reaching a rather inconvenient corner of their territory, where they consume large amounts of the legume Enterolibium contortisiliqiuim. As it happens, the fruit of the plant contains the steroid stigmasterol, a precursor of progesterone, the crucial female reproductive hormone. Thus, Strier speculates, consuming the legume helps bring on the mating season for the muriquis.

  Apes and monkeys are far from being the only animals that appear insightful about eating foods for curing what ails them. There seems little doubt that creatures of every stripe have a host of strategies for detoxifying their diets, and in some cases the maneuvers are fairly complex. For example, in the endless wars between plants and animals, many plants have evolved leaves laced with secondary plant compounds, among which substances are cardiac toxins, hallucinogens, antifertility drugs, and growth inhibitors. Such chemicals have no effect on the plants, but they can be deadly poisonous to any animal foolish enough to eat them. Not to be outdone, animals have evolved counteroffensive patterns of behavior that enable them to eat the plants: they may follow up a poisonous meal by eating something that can detoxify the poison. Rats, for instance, often consume clay after eating highly toxic plants; the clay absorbs the poisons.

  Animals also appear to correct their own dietary deficiencies. In his famous “cafeteria” studies a half century ago the psychobiologist Curt P. Richter broke up a balanced rat diet into its constituent parts, serving up eleven small trays of proteins, oils, fats, sugars, salt, yeast, water, and so forth. Remarkably, the rats picked out an efficient diet that, with fewer calories, made them grow at a rate faster than that of rats fed normal chow.

  Oh, those animals, and the wise and wonderful things they know. Wrangham, Rodriquez, and Strier, all leaders in the field, have been scrupulously careful in their reports and speculations. Nevertheless, the lay press has had a field day with these observations, nearly implying that somewhere out there, some chimp knows the cure for cancer. Despite the tremendous appeal of this new field, a surprising amount of work remains to be done before the most excited—and exciting—claims about zoopharmacognosy can be endorsed.

  To what extent do animals’ efforts to find pharmacologically active plants truly constitute self-medication? One step toward an answer is to ask another question: Does the purportedly medicinal food of animals really work? That issue is best illustrated if one imagines a reversal of roles between human and animal: A chimpanzee, for its science-fair projects, analyzes the remains in a potato chip bag discarded by a person in Topeka. Among the chemical constituents in the potato chips, the ape finds one compound that, when analyzed in a test tube, displays anticholera activity. Does that mean Topekans can prevent or cure cholera by eating potato chips? Analogously, is the amount of thiarubrine-A consumed by chimpanzees enough to kill the kinds of parasite that occur in the animals?

  That question might seem to have been settled by Rodriquez, when he estimated that the amount of thiarubrine-A ingested by chimpanzees is enough to kill a large percentage of nematodes. But there is a catch: Rodriquez was working in the test-tube world of purified thiarubrine-A. In real life some of the drug could theoretically be inactivated by other compounds in the biochemically complex Aspilia leaf. And not all of the drug absorbed from under a chimpanzee’s tongue is likely to reach its gut; some of the drug might be broken down first. In short, it must be shown experimentally that the quantity of Aspilia leaves consumed by the animal leads to fewer gut parasites. Such an experiment has not yet been reported.

  Another issue is the role of Aspilia in the chimpanzee’s overall diet. Is there really anything special about the plant, at least in its natural state? Is Aspilia the only plant near the Gombe reserve that contains thiarubrine-A? To put the issue into relief, if all the plants turn out to contain the compound, and if chimpanzees are thus eating thiarubrine-A all the time, why is a sick chimpanzee’s selective diet any more exciting than, say, the fact that people eat more chicken than pigeon? Wrangham has not reported whether thiarubrine-A is unique to Aspilia. But Strier, in her work on muriquis, notes that most legumes in the rain forest inhabited by the monkey contain ample amounts of stigmasterol, making the animals’ preference for E. contortisiliqiuim less impressive in that regard.

  Now to a tougher question: Do animals go out of their way to consume certain foods when they are sick? Does a case of cholera in Topeka cause Topekans either to start eating potato chips or, at least, to step up their daily chip consumption? When a chimpanzee winds up with a gut full of parasites, does it suddenly get a craving for some Aspilia leaves?

  The issue is critical. There is, as yet, nothing other than anecdotal evidence to suggest that the chimpanzees are showing intentionality. Intentionality, however, has been shown in one case, namely, the clay eating by rats I mentioned earlier. For example, after being fed the poison lithium chloride, rats eat clay to induce vomiting. If rats are conditioned to associate the taste of lithium chloride with saccharin, they will eat clay after eating saccharin without the lithium chloride. In other words, if a rat even thinks it has eaten a poison, it wants to eat clay. And that raises the most complex question of all: In such a valid case, how can the animals possibly know what cures their ailments?

  The easiest response would be to fall back on the kind of explanation I mocked above: animals just know, instinctively, what to eat. Richter thought his rats balanced their diets by responding to instinctual cravings. Cravings do exist, and they are understood physiologically. When animals or people are hungry, for instance, they crave food, and as part of the craving the brain makes the olfactory receptors in the nose highly sensitive to the odors of food. The food becomes easier to find. Even more specific cravings have been demonstrated: When a rat is deprived of salt, it instinctively seeks out salt and finds the taste more pleasurable than usual. The neuroscientist Thomas R. Scott and his colleagues at the University of Delaware have shown that salt deprivation in rats causes neurons in the brain that normally respond to the taste of sugar to become responsive to salt. In other words, when a rat craves salt, the salt tastes as good to the rat’s brain as sugar normally do
es.

  The trouble with invoking innate cravings to explain cases of self-medication is that only two specific innate cravings have been demonstrated: one for salt and one for water. And so the plot thickens. If an animal does, in fact, pick out the correct medicinal food at the right time, it probably learned to do it. How could that happen?

  Work by the psychologist Paul Rozin of the University of Pennsylvania throws some light on that question. Rozin wanted to study the building blocks of the nutritional expertise shown by Richter’s rats. As a test, he chose to determine whether a rat compensates when it is suddenly deprived of a single important nutrient. He discovered that a rat deprived of, say, thiamine prefers a new thiamine-rich diet to its old thiamine-free one.

  But what the rat learns is not that the new diet is good. Rather, the animal learns that the old one is bad. Animals are usually tentative about eating a new food; by only nibbling at it they are less likely to be seriously poisoned. But Rozin noted that animals with a thiamine-poor diet do quite the opposite. Given the choice, they enthusiastically dig into a new diet rich in thiamine—not because they sense the thiamine or its benefits but because anything is better than the thiamine-poor diet they have been getting. As proof, Rozin and his coworkers gave thiamine-starved rats a choice between their old thiamine-free diet and two new diets, one with thiamine and the other without it. In roughly equal proportions the rats all gravitated toward the two new menus, thereby getting, as a group, only half the thiamine replacement.

  Those findings seem to explain what is taking place when a sick animal tries a new food or, more precisely, rejects its old one. But how does the animal figure out that the new food causes it to feel better? Timing is a problem. As much as a week can pass between the consumption of a new “medicinal” food and the onset of its beneficial effects. But according to classical learning theory, associations are best made when they are closely yoked in time. How could a chimp form an association between feeling better and the Aspilia leaves it first sampled a week before?

  Work by the psychologist Martin E. P. Seligman, also of the University of Pennsylvania, may help explain the conundrum. According to Seligman, not all learned associations must be closely linked in time. He has named such delayed learning the “sauce béarnaise” syndrome, after an unpleasant experience with that rich food. Once, following a sumptuous dinner featuring a sauce béarnaise, Seligman spent an evening listening to Wagner at the opera. Back home, after midnight, he suddenly developed a case of gastroenteritis. The next time he sat down to a dinner that included sauce béarnaise, its taste made him queasy.

  Linking the stomach pains to the sauce béarnaise eaten six hours earlier made no sense according to the traditional canons of learning theory. Organisms are supposed to associate punishment (in this case, the extreme one of feeling ill) with the events that immediately precede it. The hours of operatic music were far more recent than the dinner; by all logic, Wagner, not the food, should have nauseated Seligman. But Seligman and others have noted that not all such associations are made with equal ease. Instead, to apply the principle to the case at hand, one is far more likely to link stomach pains with a gustatory event (the dinner) than with an auditory one (the opera), even if the gustatory stimulus took place much earlier than the auditory one.

  The sauce béarnaise syndrome, as originally formulated, dealt with making associations about bad events. But it should operate just as readily for associations about feeling better. Thus an ailing animal might happen upon a curative diet or a medicinal plant, and its bad feelings might make it try something new. If the new food then works reasonably quickly to make the ailment go away, the animal could associate the new food with the relief of its ailment.

  It is by no means clear, however, whether that style of learning can explain all the purported cases of animal self-medication—primates and their diets, rats and their clay eating, dogs and cats and their grass eating. Consider the reported cases of self-medication with fertility plants. Even the sauce béarnaise syndrome could not account for an association between pregnancy and a particular food. What woman can remember what she ate for breakfast two weeks before she either did or did not have her menstrual period?

  Building on Rozin’s work, Bennett G. Galef Jr. and Matthew Beck, who were both psychologists at McMaster University, may have provided some answers. They noted that many of Rozin’s thiamine-deprived rats, left to their own devices, were not great at learning which foods are effective remedies. But the social dimension makes a critical difference. Galef and Beck observed that a rat is far more likely to learn to eat a curative food if the animal is surrounded by other rats that already prefer the food. Moreover, a rat strongly conditioned to avoid a given food becomes less phobic about the food if other rats are happily eating it. In animals, as in people, social observation plays a powerful role in learning what to eat.

  Social learning may go far to explain the phenomenon of zoopharmacognosy. How many people independently figure out how antibiotics work? How many learn to build an internal combustion engine on their own? Instead, people generally learn how to use medicine or technology by watching others who are proficient at it. Perhaps some chimpanzee with a miserable stomachache, and with a metabolism that responded quickly to thiarubrine-A, stumbled onto a rich vein of the stuff in some Aspilia leaves. All that remained was for the rest of the group to mimic the discoverer. And both primates and rodents are very good at mimicking behavior.

  Notwithstanding such speculation, it remains true that in only one of the cases I have discussed here—that of nauseated rats eating clay—have experiments shown that animals make a concerted effort to eat a medicinal food. Furthermore, in no putative case of primate self-medication have the animals been shown to eat enough of the compounds to cure their ills. Finally, if rigorous experiments in the future demonstrate cases of effective self-medication, some pretty fancy learning theories will probably be needed to explain how the animals pull it off.

  Somewhere down the line—after some essential and difficult research has been done—it may turn out that all the reports of zoopharmacognosy constitute legitimate animal self-medication. I am fairly skeptical about that prospect, but if it comes to pass, it will be a delightful result. Practically speaking, it might lead to the discovery of some new medicine for people; maybe somewhere out there there really is a chimp who knows the cure for cancer. And more fundamentally—despite my own occasional cynicism—it’s hard not to be moved by evidence that other species can learn to be wise about the world.

  But even if none of the reports of zoopharmacognosy turn out to be true, the work of Rozin, Galef, and Beck demonstrates that animals can be pretty wise about the world. They know that if something ain’t broken, don’t fix it, and if something does need fixing, the best course is to keep an eye on individuals who are more experienced. Above all, they know that if things aren’t working, it’s a good idea to keep an open mind for new solutions. Those are useful lessons for human beings to learn—even from animals that might not turn out to be trained pharmacists.

  EPILOGUE

  Not surprisingly, some of the individuals whose work I discussed in this piece were not at all pleased. I reproduce their letter published in The Sciences and my response.

  To the Editor:

  Robert Sapolsky’s review of work on medicinal plant use by animals implies in the most charming manner that he’s a lone skeptic looking on a field of suckers. Mr. Sapolsky dismisses completely “the myth that animals inherently know all sorts of wise and wonderful things.” Guess what: so do we. Contra Mr. Sapolsky’s implications, we know of no one working in zoopharmacognosy who suggests that animal self-medication has been proved, or that primates have innate knowledge of medicinal plants. But we say Nonsense! to the idea that zoopharmacognosy is inherently stupid. We agree with Mr. Sapolsky that any medicinal knowledge gained by primates is likely to be culturally transmitted. But we haven’t decided exactly how the practice develops, because there is no experimental ev
idence.

  Mr. Sapolsky thinks the psychological mechanisms underlying the use of medicinal plants by apes are exclusively concerned with learning. Perhaps. But it might also be true that physical condition influences the degree of aversion to toxins. The theoretical biologist Marjorie Profet, now at the University of Washington in Seattle, pulled together the evidence that aversion to toxins increases in women when they are pregnant. When Jane Goodall was dosing sick chimpanzees, she put tetracycline into bananas. The sick chimpanzees accepted it, but the healthy ones didn’t. As the sick chimpanzees improved in health, they stopped accepting the laced bananas. Furthermore, one of us (Michael Huffman) has noted that chimpanzees with the worst parasite loads tend to chew the bitterest leaves. A health-mediated aversion to noxious toxins would rank as a fascinating example of the principle that perception can be influenced in adaptive ways by “inherent” mechanisms laid down through natural selection. We’re not saying it’s been established, just as we haven’t said animal self-medication has been proved to occur. But if it’s too early to know whether apes self-medicate, it’s certainly too soon to have a closed mind about how they do it!

  Mr. Sapolsky caricatures the evidence of the role of thiarubrine (the putative medicinal in Aspilia) in self-medication by equating it with finding a rare anticholera chemical in a bag of potato chips. That’s not even a cheap shot: it’s simply wrong. First of all, we eat chips because they taste good and because they provide nutrients. But nothing suggests that Aspilia tastes good or that the plant provides nutrients for chimpanzees. The leaves are swallowed whole; they pass through the gut with so little damage that it takes a scanning electron microscope to detect rupture of the surface cells and trichomes (“hairs”). Aspilia leaves could provide food only if they had extraordinarily high concentrations of soluble nutrients—and work by the anthropologist Nancy Lou Conklin of Harvard University shows they do not. In short, Aspilia leaves are nothing like potato chips. Ripe fruits, for a chimpanzee, are like potato chips. So are chewed leaves. But a whole leaf is like a chunk of bark, or a bit of old leather, or a page from Mr. Sapolsky’s article. You don’t expect anyone to swallow it!