As hunters, cats should be able to work out where objects that have gone out of sight are likely to be. No wild cat would give up hunting a mouse immediately after it disappeared from sight under the mistaken impression that it had ceased to exist. As expected, cats do seem to remember where prey has disappeared, although they store this information for only a few seconds, in "working" memory; not until the cat actually makes contact with the prey does the memory become longer-lasting. Presumably, it is not worthwhile for a cat to continue to search a particular location for highly mobile prey for much longer than this; by that point, the prey has either made its escape or gone to ground.

Scientists recently demonstrated that cats do indeed remember the last place they saw a mouse, rather than merely keeping their eyes fixed on it or even just heading in that general direction. In the apparatus ill.u.s.trated, scientists allowed the cat to watch a food treat being pulled behind a small barrier on a piece of cord, through the transparent section of the screen. The cat was then allowed to walk into the apparatus, but-as the remainder of the screen was opaque-this action temporarily blocked the cat's view of the location of the food. Nevertheless, the cat usually picked the right place to look for the food. Interestingly, many of the cats tested sometimes went the long way around-they set off and entered the apparatus from the "wrong" side, but then immediately crossed to the correct side to retrieve the food. This is similar to a common hunting tactic: if cats are in hot pursuit of a mouse or rat, they will often briefly take a more roundabout route, possibly to make their prey think that they have mistaken where it has hidden.16 Cats sometimes take the most direct route to the place where food has disappeared (left), but other times they seem to deliberately take a more indirect route (right), as if they were hunting and wanted to confuse their prey.

The cat's mental abilities are specifically tuned to their hunting lifestyle, rather than being part of a more general spatial intelligence. Cats perform remarkably badly in tests designed to track the development of such abilities in human infants. Many children of eighteen months can understand that if an object is hidden in a container, and then the container itself is hidden, the object ought to be in the container when it reappears. If they then find that it isn't, they look next in the place where the container disappeared. They not only understand that an object they have seen must still be somewhere even when they can't see it, but can also use their imagination to guess where it might be. Cats cannot do this at all, probably because it is not a situation that their ancestors encountered while hunting. Mice hide, certainly, but they do not hide inside objects that are themselves capable of moving around.

The cat's ability to reason seems limited, especially when determining cause and effect. They rely on simple a.s.sociations built up through conditioning, and can easily be "fooled" by our manipulations of their surroundings, which must seem arbitrary in their view of the world (as if they were thinking "How did that bag arrive in the middle of the kitchen floor?" or "Why does my owner talk into that thing in her hand?"). Nevertheless, it is entirely possible that scientists have yet to design experiments that allow cats to demonstrate their true abilities. It could be that the small number of situations in which cats have been tested happen to be those in which their evolution has favored reliance on simple learning and short-term, rather than long-term, memory. Scientists studying canine intelligence, which has received considerable attention over the past two decades, have only recently begun to find ways of testing dogs that fit their particular way of interpreting the world. Cats, with their enigmatic reputation, may still be hiding the true extent of their brainpower.

Cats are masters at concealing their thoughts, and are even better at hiding their emotions. Several cartoons show an array of cats with identical expressions, each labeled with a different emotion, ranging from "Frisky" to "Content" to "Sad," and in one instance, with irony, "Alive." One version that I treasure, by British cartoonist Steven Appleby, has no less than thirty cat faces, of which twenty-nine have identical expressions (the captions range from "About to do nothing at all" to "Slightly irritated but concealing it well"); the thirtieth, "Asleep" differs from the others only by its closed eyes.17 Biology provides good reasons why most animals keep their emotions to themselves. Dog owners may find this idea absurd, since both dogs and humans express their emotions spontaneously. Indeed, we must often suppress our feelings when social mores dictate. Nevertheless, we humans have evolved a very sophisticated ability to detect tiny flickers of emotion in others, telltale signs that help us predict what that person is likely to do next. Dogs, in their way, are similar to us in this respect; not only have they evolved the ability to guess our intentions from our body language, but they also express their emotions openly, partly because we generally respond in ways that are beneficial to them-backing away when they growl, or giving them a pat if they're wagging their tails. Dogs and humans are both social species that usually live in stable groups, and such stability means that emotional honesty is unlikely to be penalized.

Cats are descended from a species with a solitary lifestyle, and therefore much of their behavior is guided by the need to compete, not to collaborate. In the wild, a male cat will live alone. The only way he can be sure of leaving any descendants is first to convince a female to accept him as a mate, and second to convince any rival males to back off. Macho behavior, laced with a generous portion of bluff, is therefore essential to their success. Although female domestic cats do cooperate when raising kittens, this habit, which may have evolved during domestication, seems to have had little effect on their capacity to express their emotions.

Historically, scientists have changed their mind several times over whether animals' emotions should be used when discussing their behavior. In the nineteenth century, researchers often ascribed human emotions to cats. For example, in his 1886 book Animal Intelligence, physiologist George S. Romanes wrote: The only other feature in the emotional life of cats which calls for special notice is that which leads to their universal and proverbial treatment of helpless prey. The feelings that prompt a cat to torture a captured mouse can only, I think, be a.s.signed to the category to which by common consent they are ascribed-delight in torturing for torture's sake.18 By the beginning of the twentieth century, such anthropomorphism had been abandoned, and the guiding principle in animal psychology was Morgan's canon: "In no case may we interpret an action as the outcome of the exercise of a higher psychical faculty, if it can be interpreted as the outcome of the exercise of one which stands lower [that is, simpler] in the psychological scale."19 For a while, indeed, scientists conceived of animals as if they were robotic stimulus-response machines, leaving no room for any consideration of emotion. Recently, however, we have come to realize that it is difficult to explain some animal behavior without invoking the idea of emotion. In addition, MRI scanning has enabled us to see where in the human brain emotions are generated, and the simpler, "gut-feeling" emotions occur in parts of our brains that we share with other mammals, including cats.

The current view holds that emotions are a necessary component of the mechanisms that drive animal behavior-and indeed our behavior as well. They can be prompts to shortcuts, enabling our brains to choose the best response to a situation when fast action is called for. In this, cats are no different from ourselves. A cat that sees another, larger, unfamiliar cat approaching will immediately become alert, crouch down, and prepare to flee. The anxiety it feels at seeing a potential adversary enables it to take these actions immediately, without having to think the situation through and evaluate all possible strategies and outcomes.

Emotions also explain spontaneous and apparently functionless behavior. Kittens engage in play during most of their waking moments, and explaining why they do so is not straightforward. In the wild, play is a mildly risky activity, exposing the kittens to danger and possibly attracting the attention of a predator-surely it seems safer for kittens to stay quietly in their nest and wait for their mother to return with food. Furthermore, kittens do sometimes nip each other when playing, but this doesn't seem to put them off playing with that same kitten again, which it should do if they are simple stimulus/response machines.

The simplest explanation for why kittens feel they ought to play, and why they continue to play even following a slight mishap, is that play is fun. Neuroscientists have found in young rats that when they play, the neurohormone profile of the brain changes. Moreover, these changes are not the consequences of play, they appear to be its cause: they occur as soon as the rats are given the signal that it's time to play. As such, the mere sight of a sibling ready to play is likely enough to make a kitten want to join in, because the brain signals "fun" before the playing has even begun.

Of course, hormones are not the same as emotions, but changes in certain hormones are often a sign that emotions are being experienced. We are all aware of the racing heartbeat, hyperventilation, hyperalertness, and sweaty palms induced by adrenalin or epinephrine, the "fight-or-flight" hormone a.s.sociated with feelings of fear and panic. Some of us are familiar with the elation we sometimes feel after strenuous exercise, caused by the release of endorphins and other hormones into the brain. Although not all hormones are so closely connected to emotions, many are, and can provide an indicator of immediate emotion or underlying mood.

We can therefore conceive of animal emotions as manifestations of the brain and the nervous system and their a.s.sociated hormones, sometimes enabling decisions to be made quickly, other times directing learning. Sometimes, information coming into the brain from the cat's senses has to trigger an immediate reaction. A cat that slips while walking along the top of a fence must correct its balance immediately: the emotional panic that almost certainly follows will help train the cat to be more careful next time. On other occasions, the emotion triggers the behavior. Indeed, the sight of its owner coming home will cause a cat to feel affectionate, and as a consequence it will raise its tail upright and begin to walk toward her.

Some people who don't like cats-and even some who do-would claim that love, especially towards its human owner-is not part of the cat's behavioral repertoire: as the popular adage goes, "Dogs have masters, cats have staff." True, the average cat does not outwardly demonstrate love for its owner in the same way as, say, a Labrador retriever would. Still, that tells us little about what is going on inside the cat's head.

In the animal world, lavish outward displays of emotion are usually manipulative. Consider, for example, the incessant chirping of baby birds in a nest, each essentially calling out, "Feed me first, feed me first!" Evolution has ensured that where such displays occur, they work: the baby bird that makes the least fuss is ignored by its parents, which often produce more offspring than they can comfortably feed, and it may perish as a result. Wildcats are solitary and self-sufficient except for a few brief weeks at the beginning of their lives, and so have little need for sophisticated signals. Although domestic cats now depend on us for food, shelter, and protection, they have not done so for long enough to have evolved the average dog's effusive greeting. That doesn't mean that cats are incapable of love, merely that their ways of showing love are somewhat limited.

Cats become extravagantly demonstrative only when they are angry or afraid. A fearful cat either will make itself look as small as possible by crouching, and then slink away, or, if it judges that running away may provoke a chase, will make itself look as large as possible, arching and raising the fur on its back. The raw emotion does not therefore provoke an automatic and invariable reaction; instead, the brain selects the more appropriate response based on other information available to it.

The angry cat will not only try to look as large as possible, but will also stand head-on toward the threat (usually another cat) with its ears forward, either yowling or growling loudly and lashing its tail side to side. We may be inclined to read these postures as expressions of emotion, but they are fundamentally expressions of intention, as well as attempts to manipulate the animal the cat is confronting.

Even when a cat is demonstrably trying to manipulate his opponent's behavior, such manipulation need not be conscious. We can explain the cat's behavior by its adhering to a set of rules that have served its ancestors well; which have, in the past, achieved the best possible outcome from an aggressive encounter. We must not forget, however, that the bluff is being directed at an animal with similar brainpower, and therefore evolution has also favored animals capable of "mindreading" the other's intentions. One cat confronting another will expect one of two possible reactions: either a sign that the other cat is fearful and will probably back down, or that it is not fearful and so a fight is imminent. In this way, the behavior, whether indicating fear or anger, becomes ritualized: any cat that adopts neither posture, or does something different, is likely to be attacked.

To explain their behavior, we must thus allow that cats feel joy, love, anger, and fear. What other emotions do they possess? Can they feel the full gamut of emotions that we can? To answer these questions, we must consider which are likely the product of human consciousness, and therefore unknown to animals.

People do not agree on the range of emotions their cats can feel. A 2008 survey of British cat owners revealed that almost all think their cat could feel affection, joy, and fear.20 Nearly one-fifth of these owners-perhaps those who own very timid cats-were unsure whether anger is within the cat's repertoire.

We all know the old saw "Curiosity killed the cat," and indeed, most owners acknowledge the cat's characteristic nosiness. The original version of this proverb, from its first appearance in the sixteenth century until the end of the nineteenth, was "Care killed the cat"-"care" in the sense of worry, anxiety, or sorrow. Apparently, the idea that cats could become so anxious that they could even die from it was once common (and is now being revisited by veterinarians). Despite this, about a quarter of the owners surveyed in 2008 thought their cat incapable of feeling anxiety or sadness.

If asked, scientists would now agree that the old version of the proverb contained a germ of truth: anxiety does const.i.tute a serious and real affliction for many cats. Anxiety, if simply defined as a fear of something that is not currently happening, has a reliable basis in physiology. Some anti-anxiety drugs developed for humans have also been found to reduce symptoms of anxiety in cats, so although we cannot be sure that cats experience anxiety in precisely the same way we do, we know that they feel something similar.

Owners' opinions about their cats' emotional capabilities The most common cause of anxiety in cats is probably the worry that their territory is likely to be invaded by other cats in the neighborhood, or even by another cat in the same household. When I surveyed ninety cat owners in suburban Hampshire and rural Devon in 2000, they reported that almost half of their cats regularly fought with other cats, and two out of five were fearful of cats in general. My colleague Rachel Casey, a veterinary surgeon specializing in cat behavioral disorders, regularly diagnoses anxiety and fear as the main factors driving cats to urinate and defecate indoors, outside the litter tray. Some cats spray the walls or furniture with urine, possibly to deter other cats from entering their owner's house believing it to be cat-free; others find the point in the house farthest from the cat-flap and urinate there, seemingly terrified of attracting the attention of any other cat. Some will defecate on the bedsheets, desperately trying to mingle their own odor with their owner's to establish "ownership" of the core of the house. When the conflict is between two cats that live in the same house, one may spend much of its time hiding, or obsessively groom itself until its coat becomes patchy.21 The stress of being forced to live with cats it does not trust can often be severe enough to affect the cat's health. One illness now known to be closely linked to psychological stress is cyst.i.tis, referred to by veterinarians as idiopathic cyst.i.tis because no disease or other medical cause is apparent. As many as two-thirds of cats taken to vets for urination problems-blood in the urine, difficult or painful urination, urinating in inappropriate places-have no obvious medical problems, other than inflammation of the bladder and intermittent blockage of the urethra by mucus thereby displaced from the bladder wall. The factors triggering such episodes of cyst.i.tis are therefore psychological, and research has identified conflict with other cats living in the same household as being perhaps the most important of these. Less easy to quantify, but possibly just as important, are conflicts with neighbors' cats: certainly cats p.r.o.ne to cyst.i.tis usually run away from cats they meet in their own gardens, rather than standing up to them, which suggests that they find contact with other cats particularly stressful. Idiopathic cyst.i.tis is less common in female cats than in males: the conventional medical explanation is that the tube leading out of the bladder, the urethra, is generally narrower in males and therefore more p.r.o.ne to blockage. However, male cats are generally more territorial and less sociable than females, so the latter may find it easier to resolve or avoid conflicts with other cats before stress affects their health.

Colleagues of mine at Bristol University Veterinarian School doc.u.mented one case involving a five-year-old male cat that was having great difficulty urinating, and when he did, his urine was b.l.o.o.d.y. He also groomed his abdomen excessively, but otherwise was entirely healthy. The cat was one of six within the household, but was friendly with none of the other cats. Moreover, cats from neighboring households had recently attacked him. His symptoms gradually disappeared once his owners had implemented the changes recommended by the clinic: his own exclusive area within the house, his own food bowl, and his own litter tray that the other cats could not access. At the same time, his owners blocked his view of the garden by covering over the gla.s.s at the bottom of the windows in his part of the house, so he could not see any cats that were coming into the garden. Six months later his symptoms returned, but on investigation, it turned out he had been accidentally shut in with the other cats a couple days previously. His owner vowed never to allow this to happen again, and the cat soon recovered.22 Anxiety, a useful emotion if experienced for a few minutes, can become the bane of a cat's existence if prolonged for weeks or months, leading to chronically elevated levels of stress hormones, (presumably) a nagging and ever-present sense of dread, and eventually to deterioration in health.

In the same survey, cat owners were also asked about their cats' more complex emotions-jealousy, pride, shame, guilt, and grief. Almost two-thirds believed that their cat could feel jealous or proud. Only shame, guilt, and grief were ruled out by a majority of the owners.

Basic emotions such as anger, affection, joy, fear, and anxiety are "gut feelings" that appear spontaneously. The most primitive part of the cat's brain produces these emotions-the same part that evolved hundreds of millions of years ago before there were any mammals, let alone cats. More complex emotions, such as jealousy, empathy, and grief, require the cat to have some understanding of the mental processes of animals other than themselves, and hence psychologists sometimes refer to them as relational emotions.

Take jealousy, for example. When we experience jealousy, we are not only aware that whoever we are feeling jealous of is another human being, we can also guess what that other person is feeling; we have what psychologists call theory of mind, the idea that other humans have their own thoughts that may be different from ours. We are also capable of becoming obsessively jealous, thinking about the incident that triggered the original feeling long afterward, even when the person we were jealous of is no longer present. We have little evidence that cats have either the brainpower or the imagination to do either of these.

Cats undoubtedly recognize other cats as cats, and can evidently react to what they see them doing. However, even dogs, which are much more highly evolved socially than cats, show no evidence of understanding what other dogs are thinking, so it's unlikely that cats can either. Moreover, cats seem to live in the present, neither reflecting on the past nor planning for the future. Still, at its heart, jealousy is an emotion first experienced in the here and now; it does not require the cat to understand what its rival is thinking, or even that it is capable of thinking at all. All that jealousy requires is that the cat merely perceive that another cat is getting more of something than it should. Thus, cats are almost certainly capable of feeling jealous, even if not quite as demonstratively, or as commonly, as dogs. Although not something that my cats have ever indulged in, countless owners have regaled me with stories of how one of their cats would always intervene when they tried to stroke the other.

Many people think that cats are capable of grief, because they behave oddly when another cat that they have known disappears. What they actually feel is probably a temporary anxiety, which disappears once all traces of the missing cat have disappeared. A mother cat may search for her kitten for a day or two after it has been homed. She probably has a memory of that kitten, and may even count the remaining kittens to check that one is missing. This behavior would be the same if that kitten had temporarily gotten lost; in the wild, it would be in the mother cat's interest to seek it out and continue to look after it until it was old enough to become independent of her. She cannot "know" that it has gone to a good home where it will be well cared for, as nothing in her evolution will have prepared her to embrace that concept. For a few days, the mother is reminded of that particular kitten by the lingering traces of its individual odor, the kind of cue that is often meaningless to us. We know that the kitten has gone because we can no longer see or hear it. Once the kitten's odor has faded below her threshold, the mother cat probably forgets all about the departed kitten. While she can still smell it, she may well feel an anxiety that drives her to continue to look for it. This, however, is not the same as grief.

Emotions such as guilt and pride would require cats to possess a further level of cognitive sophistication, the ability to compare their actions with a set of rules or standards that they have worked out for themselves. When we feel guilty, we compare the memory of something we've just done with our sense of what is wrong. Such feelings are sometimes referred to as self-conscious emotions, because they require a degree of self-awareness to be experienced. So far, science has yet to reveal any evidence for self-awareness in cats, or even in dogs. Dogs are widely believed to display a "guilty look" when their owners discover that they've done something that's forbidden, but a clever experiment has shown that this is all in the owners' imagination.23 The researcher asked the owners to command their dogs not to touch a tempting food treat, and then leave the room. Then, unbeknown to the owners, she encouraged some of the dogs, but not others, to eat the treat. When the owners came back into the room, they were all told that their dog had stolen the treat-whereon all of the dogs immediately began to look guilty, whether or not they actually had something to feel guilty about. The "guilty look" was nothing more than each dog's reaction to their owners' body language, which had changed subtly as soon as they had been told of the dog's misdeed, whether real or invented. If dogs' "guilty looks" are a figment of their owners' imaginations, it follows that they-and by extension, cats-are incapable of feeling guilt. The same is probably true for pride, but no scientist seems to have studied it in any relevant animal.

Cats' emotional lives are more elaborate than their detractors would have us think, but not quite as sophisticated as the most ardent cat-lover would probably like to believe. Unlike dogs, cats hide their emotions-not primarily from us but from one another, a legacy of their evolutionary history as solitary, compet.i.tive animals. We have every reason to believe that they possess the basic range of emotions, the gut feelings, shared by all mammals, because having these empowers them to make quick decisions, whether that is to run away (fear), play with a ball of string (joy) or curl up on their owner's lap (love). However, cats are not as socially sophisticated as dogs are: they are undoubtedly intelligent, but much of that intelligence relates to obtaining food and defending territory. Emotions that relate to relationships, such as jealousy, grief and guilt, are probably beyond their reach, as is the ability to comprehend social relationships with any great sophistication. This leaves them ill-equipped for the demands of living closely with other cats, as domestication has progressively required them to do.

CHAPTER 7.

Cats Together Cats can be very affectionate, but they are choosy about the objects of their affection. This apparent fastidiousness stems from the cat's evolutionary past: wildcats, especially males, live much of their lives with no adult company, and regard most other members of their species as rivals rather than as potential colleagues. Domestication has inhibited not only the wildcat's intrinsic distrust of people, but has also tempered some of their wariness of other cats.

The bond between cat and owner must have its origins in the bond between cat and cat; such behavior has no other plausible evolutionary source. Although the cat's immediate ancestor, the wildcat, is not a social animal, adult felids of other species, such as lions, do cooperate. As such, cats of any species could possibly become more sociable in the right conditions. We may find the source of the domestic cat's affection for its owners through a brief survey of the social life of the entire cat family.

Both male and female tigers are solitary, exemplifying the pattern that almost all members of the cat family, big and small alike, live alone. The females hold non-overlapping territories, which they defend from one another; each territory is large enough to provide food not just for that female, but also for the litters of cubs she rears. Young males are usually nomadic, and when they become mature, they try to set up their own large territories. These will contain far more prey than the male will ever need to satisfy his hunger, but that is not its purpose. The male is trying to achieve exclusive access to as many females as possible: especially successful males may hold territories that overlap those of up to seven females.

Cheetah males, especially brothers, are a bit more sociable than tiger males. Female cheetahs are as solitary as female tigers, but male cheetahs sometimes band together in twos or threes to seduce females, many of which are migratory, as they pa.s.s nearby. Even though only one of the brothers will be the father of the resulting litter, biologists have shown that over his lifetime each brother will father more cubs than if he had tried to attract females on his own. Male cheetahs do occasionally try to hunt together as well, but are rarely successful, apparently lacking the skills to coordinate their efforts.

The best-known exception to the standard felid pattern is the lion, the only member of the cat family to have several males and several females living together. In Africa, the lion pride is usually constructed from one family of female lions, with the males originating from a different family (thereby preventing inbreeding). While still young, related males band together, sometimes adding unrelated males to swell their ranks, until their numbers are sufficient to challenge and eject the resident males from a pride. Once they have taken over the pride, they may kill all the cubs, and by doing so bring all the females into season within a few months. The males must then keep control of the pride until the females have given birth to their progeny and raised them to independence. For their part, the females not only have to raise the cubs, but they also do most of the hunting, while the males do little but protect the females from other bands of males. Thus, the superficial image of lion society as generally harmonious is something of a myth; rather, it is the result of tension between the benefits of cooperation and compet.i.tion, with each individual lion employing tactics that maximize its own breeding success.

Biologists still debate why lions live in such groups. In India, lions are often solitary animals, males and females meeting only for mating, so members of this species can evidently choose whether or not to live together. Although the females in a pride generally hunt together, they do not cooperate as selflessly as their "brave" image suggests: if the prey is large and potentially dangerous, the more experienced females usually hold back and let the younger, more impetuous lionesses take the risks. The main benefit of numbers, and especially the presence of the fierce males, may come after the kill, when the valuable meat must be defended from other animals, particularly hyenas.

Scientists once considered lions and cheetahs the only social cats, and added the domestic cat to that exclusive list only recently. It had long been clear that wherever there was a regular source of suitable food, a group of feral cats would spring up, but these groups were originally considered mere aggregations of individuals that had somehow agreed to tolerate one another, as happens when animals of many species come together to drink at a waterhole. Cat breeders also knew that their queens would sometimes suckle one another's litters, but scientists dismissed this behavior as the result of the artificial conditions under which humans usually keep pedigree cats. In the late 1970s, however, David Macdonald's doc.u.mentary about cats on a farm in Devon, England, showed that this was in fact natural behavior-that free-living females, especially related females, will spontaneously cooperate to raise their kittens together.1 At the start of the study, the colony consisted of just four cats: a female, Smudge; her daughters, Pickle and Domino; and their father, Tom. When not in the farmyard, they kept to themselves-domestic cats, unlike lions, do not hunt together-but when their visits coincided, they usually curled up together in apparent contentment. They evidently regarded the farmyard and the food and shelter it provided as "theirs," since the three females would join together to drive away three other cats that lived nearby-a female, Whitetip, her son Shadow, and her daughter Tab. Tom, though, was aggressive only toward Shadow, presumably regarding him as a potential rival, whereas he might need to court the two females should he lose his own "pride" at some point in the future.

Pickle and Domino were the first to reveal how female cats come together to help one another. Early in May, Pickle produced three kittens in a nest in a straw stack. For the first couple weeks, she looked after her kittens on her own, just as any other mother cat would. Then, suddenly, her sister Domino appeared in the nest and gave birth to five more kittens-ably a.s.sisted by Pickle, who helped with their delivery and with cleaning them up. Then, despite their difference in ages, all eight kittens settled down together and were nursed and cared for indiscriminately by both their mothers. Sadly, all eight kittens subsequently succ.u.mbed to cat flu, a common scourge of litters born outdoors in the UK. However, when their grandmother Smudge produced a single male kitten a few weeks later, the aptly named Lucky, both Domino and Pickle helped with caring for him, including playing with him and bringing back mice they had caught, to save Smudge leaving him alone while she went out hunting for herself.

Domino and Lucky playing Subsequent studies have revealed that such cooperation between related female cats is the rule, not the exception. I observed this in my own home when my cat Libby gave birth to her first litter: Libby's own mother, Lucy, shared their care, grooming them and curling around them to keep them warm. Indeed, when they grew old enough to move around the house, Libby's kittens generally preferred their grandmother's company to their mother's.

Cat society is based on females from the same family. In feral or farm cats, it rarely involves more than two (sisters and their kittens) or three generations (mother, daughter[s], and kittens). However, we see little indication that the partic.i.p.ants are consciously helping one another. Rather, many female cats, especially those with kittens, simply seem not to distinguish between their own offspring and those of other cats with whom they are already friendly; in the wild, these are most likely their own daughters or sisters-cats they have known and trusted all their lives. Some mother cats that have recently given birth will accept almost any kitten they are presented with, and some humane organizations make use of such cats as wet nurses for motherless litters.

Larger colonies of feral cats usually consist of more than one family, and while these families continue to cooperate among themselves, they also compete with one another. The size of a cat colony is determined by the amount of food available on a regular basis, and where this is plentiful-for example, in a traditional fishing village where the catch is processed in situ-colonies can build up until several hundred cats are all living in close proximity to one another. Cats are prolific breeders, and numbers can grow quickly until food becomes scarce, at which point the cats on the edge of the colony will either leave or succ.u.mb to disease brought about by malnutrition.

Each family group strives to monopolize the best sites for dens in which kittens can be born, and to stay as close as possible to the best places to find food. However, as the most successful families grow, tensions increase among its members, even when there is enough food. Cats appear to be incapable of sustaining a large number of friendly relationships, even when all their neighbors are close relatives. Squabbles break out, and eventually some members of the family are forced out-and because all the prime s.p.a.ce in the colony will already be occupied by other cats, they may have to find a place alongside the new arrivals on the very fringes of the colony, where pickings are slim.

Cat colonies are therefore far from being well-regulated societies: rather, they are spontaneous gatherings of cats that occur around a localized concentration of food. If the food supply is limited, a single family may monopolize it. When food is especially plentiful, several families compete for the best and biggest share, although warfare between the various clans is generally conducted through threats and careful skirting around one another, punctuated only occasionally by overt violence. In such situations, being able to call on family to a.s.sist is essential to holding on to prime territory: a female cat on her own, especially if she has kittens to feed, is unlikely to thrive.

The cooperation within families that undoubtedly occurs in larger colonies is based on the same bonds of kinship that form in much smaller colonies consisting of a single family. Cats seem incapable of forming alliances between family groups, unlike, for example, some primates; negotiation skills of this sophistication lie beyond their capabilities.

Biologists are uncertain about the precise origins of these family ties. They could be accidental, caused by female cats' inability to distinguish their own kittens from others. Looking back to their wildcat ancestors, every female holds her own territory, which she defends against all other females, so the chance that two litters could ever be born in the same place is virtually zero. A female cat, wild or domestic, probably follows a simple rule of looking after all the kittens that she finds in the nest she has made: she sees no need to sniff each kitten carefully to check that it is not an interloper before settling down to nurse it. However this is unlikely to form the sole basis for cooperation among adult cats. The thousands of generations over which domestic cats have evolved away from their wild ancestors provides enough time for more sophisticated social mechanisms to have evolved.

Cats' social behavior probably started to evolve as soon as humankind's invention of food storage first made concentrated sources of food available. Any cats that maintained their natural antagonism toward all members of their own species could not have exploited this new resource as efficiently as those able to recognize their relations, and both give and receive help from them.

Biologists distinguish two different ways in which cooperative behavior can be beneficial to both parties. One is reciprocal altruism, continuing to do favors only for those who have done favors for you. This could theoretically take place between any two animals that live near each other, irrespective of whether they were related. However, if they are related, a second reason why it would be a good idea to cooperate is kin selection. Cats that are sisters share half their genes, a much greater proportion than two unrelated female cats would. Their kittens, even if fathered by different tomcats, each share a quarter of their genes with their aunt. Thus, whichever of those kittens goes on to have kittens of his or her own (it's an inevitable fact of life that not all of them will), each kitten shares genes with both their mother and their aunt.2 Since neither sister knows which of their offspring is most likely to survive to maturity, they should, all other things being equal, try to raise both litters. Thus, the genes that favor cooperation between cats that are related can flourish, at the expense of rival genes that promote antagonism, even between sisters.3 Reciprocal altruism and kin selection are useful mechanisms preventing selfish behavior, but cooperative behavior itself will evolve only if its benefits outweigh its costs. For the first domestic cats, the initial advantage of living in family groups would have been that the abundant food-whether prey infesting food stores, sc.r.a.ps provided by people, or a mix of the two-could be shared without constant strife. However, putting several litters into one nest means that if one kitten gets sick, they all do; this can be fatal, as Domino and Pickle's experience ill.u.s.trates. When in 1978 South African scientists introduced a virus to exterminate cats that were causing havoc among the ground-nesting seabirds on Marion Island in the Indian Ocean, cats that had retained their wild ancestors' habit of nesting on their own mostly survived, but family groups perished. Elsewhere, this disadvantage may be outweighed by the benefit of having several mothers on hand to protect kittens from predators: a solitary mother must leave her litter from time to time to find food, or her milk will dry up.

Two or more mother cats that pool their kittens can guard them much more effectively than can a lone mother, who has to leave her kittens alone in order to hunt. This benefit must outweigh the increased risk of disease wiping out the whole litter, or the ability to cooperate like this would never have evolved. Domestic cats living among humans have probably always had two main enemies: stray dogs and other cats. Twenty years ago, while on holiday in a Turkish village with my wife and our youngest son, our apartment was visited by a heavily pregnant calico stray we christened Arikan. After disappearing for a couple of days, she returned, much thinner and starving hungry, so we knew that she must have given birth nearby. We found a supermarket that sold cat food (receiving puzzled looks from the cashier, who seemed not to expect tourists to make such a purchase). When we followed Arikan after she ate, she disappeared into some derelict farm buildings farther up the road. After that, we fed her morning and evening, until one night we were awoken by pitiful wailing: Arikan was outside our door, carrying a dead kitten that showed signs of having been mauled. Arikan immediately ran off, and moments later returned with another dead kitten, which she deposited alongside the first on our doorstep.

A dog may conceivably have been the culprit. Small packs of "latchkey" dogs did roam the village in the evenings, badgering diners at outdoor restaurants for table sc.r.a.ps and chasing cats. If a pack of dogs found a cat's nest, they would probably dismantle it, but possibly not eat the kittens-although on one occasion, I did witness a dog playing with a dead kitten. Under other circ.u.mstances, dogs have been effective predators of cats: in Australia, dingoes-originally domestic dogs, gone wild-keep feral cats in check, enabling local small marsupials to thrive.4 Still, despite a long history and reputation for being the cat's prime nemesis, dogs were likely not to blame for these kittens' deaths.

Because Arikan's loss occurred at night, the culprit was more likely another cat, the dogs having returned to their owners' homes at dusk. Infanticide is a regular occurrence in lions, but only a few instances have been doc.u.mented in domestic cats.5 Male lions kill cubs that are not their own because doing so brings the mother lion back into season almost immediately; otherwise, the males must endure a nineteen-month interval between births, and by that time those males might have lost control of the females.

Female cats are often ready to mate as soon as their kittens are weaned, or earlier if they do not survive infancy, so the intruding male cats that have carried out all recorded instances of infanticide probably gained little from their cruel act, at least in terms of increasing their opportunities for mating. Infanticide appears to be most common in small, single-family cat colonies on farms, rather than-perhaps surprisingly-in the large multi-group colonies where aggression is much more generally evident. In those larger colonies, females often mate with more than one male, thereby making it much more difficult for those males to work out which kittens are theirs and which are not. In this way, males should kill only the kittens of females they can be sure they have never met before.

If the mother is present, she defends her kittens with all her might, so males are wise to target unguarded litters. A pooled litter with two or more females in attendance will be better protected against marauding males than litters kept in separate nests, even though they are less well-isolated from infections. Perhaps the unfortunate Arikan had no surviving sisters with whom she could have joined forces.

Family life provides opportunities for cats to learn from one another, rather than working everything out for themselves. As we have seen, mother cats instruct their kittens on how to handle prey by bringing it back to them. We have no evidence that she actively teaches them: she simply provides them with opportunities to learn what prey is like, but in the safe environment of the nest. Also, although kittens naturally pay great attention to everything their mothers do, it is unlikely that they deliberately imitate her. True imitation involves complex mental processes; the animal must first know what the relevant actions are, and then translate what it has seen into movements of its own muscles. Because we ourselves find it easy to imitate, we tend to a.s.sume that other animals can do the same, but research indicates that true imitation-deliberate copying of another animal's actions-may be restricted to primates.

We know of simpler ways that kittens learn from their mothers, without directly imitating her actions. Instead, the mother draws the kittens' attention to an appropriate target, and they then direct their own instinctive behavior toward it. In a 1967 experiment, scientists demonstrated that mothers teach their kittens best when challenging them to perform an arbitrary task to get food. The experiment consisted of allowing a kitten to explore a box that had a lever sticking out of one wall.6 The kitten would usually ignore the lever, apart from giving it a sniff, unless it had watched its mother pawing the lever and being rewarded with food, in which case it paid great attention to the lever and quickly learned that pawing it produced a reward. Usually, kittens will spontaneously pat an object that moves, but only rarely pat something that appears inanimate and fixed. As other experiments have shown, that both mother and kitten ended up pawing the lever was probably not due to direct imitation-the lever was designed to work best when pawed-but instead to the kitten's realization that the lever it had seen its mother manipulating looked the same as the one in its own box, and was therefore something it should investigate closely. Kittens invariably use their paws to investigate something novel, so they didn't need to imitate their mother's actions; they simply did what came naturally. In a similar experiment, other kittens were allowed to watch unfamiliar female cats performing the task. These kittens either took a long time to learn to paw the lever or didn't learn it at all, showing that it is specifically the mother that kittens feel comfortable watching. They are probably just too inhibited by the presence of an unfamiliar adult cat to learn anything.

Mother cats can pa.s.s some of their hard-won expertise on to their kittens by providing them with novel experiences, but this is equally easy for solitary mothers and those that live socially, in families. Can adult cats benefit from living socially by learning from one another? We know little about this intriguing possibility, but one obscure experiment done more than seventy years ago suggests that they can.7 Several six-month-old cats were given the opportunity to obtain a bowl of food placed on a turntable out of their reach, as shown in the ill.u.s.tration. A ratchet under the turntable ensured that a single swipe of a paw was not sufficient to bring the food within reach, and it took many sessions before the cats had learned that careful manipulation with one paw was needed to pull the food within reach. However, two cats that had never used the apparatus themselves, but had watched their sisters successfully rotate the turntable and eat the food, both solved the problem in less than a minute. It was probably no coincidence that it was sisters, not unrelated cats, that found it easy to learn from one another. Transmission of skills between family members may be mutually beneficial and give family groups an advantage over solitary cats, that can learn only by trial and error.

The turntable used to test cats' ability to learn from other cats. The cat can get the food by gradually rotating the table until the bowl pa.s.ses through the gap.

Whether young cats ever learn much from cats that are not members of their family is doubtful. Underlying distrust probably focuses their minds exclusively on staying out of trouble, overriding any curiosity about what another cat is doing. However, within a close-knit family group, younger members may benefit from watching how older, more experienced members of the group solve everyday problems. Because cats hunt alone, this is most likely to occur when the cats are in their shared core territory, perhaps when scavenging for food or interacting with people.

Although we may logically a.s.sume that cats began to live in family groups only since they started to a.s.sociate with humans, wildcats also may have (or may have had) this ability. The formation of a group of cats seems to have only one essential requirement: a reliable source of food that can feed more than one cat and her litter of kittens. The only felid that has consistently achieved this leap is the lion, which has adopted group hunting as a way to prey on large animals. However, other felids might once have lived in small groups even without developing this additional skill.

Before man's domination of the environment in the twentieth century led to the depletion of the small cats and their favorite prey, wildcats might occasionally have lived in colonies. Several accounts left by early twentieth-century European explorers of Africa give tantalizing glimpses of this possibility. Willoughby Prescott Lowe, one of the last noted collectors of animals for the British Natural History Museum, describes a specimen he took from near Darfur in the Sudan in 1921: I trapped an interesting cat near Fasher. Something like a domestic cat-but v. different in coloration. The curious thing is that they live in colonies in holes in the open plain-all the holes are close together-just like a rabbit warren. I'm told they are v. local-Anyhow a cat with these habits was quite new to me! They feed on gerbils which swarm everywhere and the ground is always a ma.s.s of holes.8 Ten years later, on an expedition in the Ahaggar Mountains in the center of the Sahara desert, Lowe again recorded colonies of wildcats, living in burrows previously dug by fennec foxes.

Both these cats looked like typical African wildcats, but their sociability need not have come from their wildcat ancestry. The DNA of apparent wildcats from farther south in Africa, and also from the Middle East, reveals extensive interbreeding between domestic and wild cats. What Lowe saw may have been colonies of hybrids, which had retained the domestic cat's ability to live in family groups while outwardly appearing to be wildcats. That social groups of these Felis lybica have been recorded so infrequently suggests that, when they do occur, their social skills may have originated in previous interbreeding with domestic cats.

The switch from solitary animal to social living requires a quantum leap in communication skills. For an animal as well-armored and suspicious as a cat, a simple tiff between sisters might well escalate into a family bust-up-unless, that is, a system of signaling evolved that allowed each cat to a.s.sess the others' moods and intentions. And this, it seems, is precisely what happened.

For domestic cats, my own research has shown that the key signal is the familiar straight-up tail. In cat colonies, when two cats are working out whether to approach each other, one usually raises its tail vertically; if the other is happy to reciprocate, it usually raises its tail also, and the two will walk up to each other.9 If the second cat does not raise its tail and the first is feeling especially bold, it may approach nevertheless, but obliquely. If the second cat then turns away, the first cat occasionally meows to attract its attention-among the very few occasions when feral cats meow. Otherwise, the first cat lowers its tail and heads off in another direction, presumably judging that the other is not in the mood to be friendly. Hesitation can be risky. My research team doc.u.mented instances when a cat moving in the wrong direction, even with its tail upright, was chased off by another, usually larger, cat determined to be left alone.

Observations such as these do not prove conclusively that the upright tail is a signal; it could conceivably be something that happens when two friendly cats meet, with no meaning for either. To isolate the raised tail from everything else that a real cat might do to indicate its intentions, we cut life-size silhouettes of cats from black paper and stuck them to the baseboard of cat owners' houses. When the resident cat saw an upright-tail silhouette, it usually approached and sniffed it; when the silhouette had a horizontal tail, the cat backed away.10 The tail-up signal has almost certainly evolved since domestication, arising from a posture kittens use when greeting their mothers. Adults of other cat species raise their tails only when they are about to spray urine, simply for hygiene. A few individuals of Felis lybica in zoos do raise their tails when about to rub on their keepers' legs, but these, of course, could have some domestic cat in their ancestry. Adults of the other races of Felis silvestris do not raise their tails in greeting, but their kittens do hold their tails upright when approaching their mother; no one has tested whether this holds any significance for the mother, so we do not know whether kittens do this as a signal or whether it's purely incidental. Therefore, it seems most plausible that the upright tail evolved from a posture into a signal during the early stages of domestication. This would have required two changes in how cats organize their behavior: one, for adult cats to perform the kitten's raised tail posture when approaching other cats, and two, for other cats to recognize instinctively that a cat with its tail raised is not a threat.11 Once both these changes had occurred, the posture would have evolved into a signal, one that enabled adult cats to live in close proximity to each other with less risk of quarreling.12 Once an exchange of tail-ups has established that both cats are happy to approach each other, one of two things occurs; which of the two seems to be influenced by both the cats' moods and the relationship they have with each other. If the cats are in the middle of doing something else, and/or one cat is significantly older or larger than the other, they usually walk up to or alongside each other. Then, keeping their tails upright, they come into physical contact and rub their heads, flanks, or tails-or a combination of all of them-on each other, before separating and walking on. Any two cats from the same group will perform this occasionally, but it is typically performed by female cats greeting males, and young cats of either s.e.x greeting females.

The precise significance of this rubbing ritual is still unclear. The physical contact in itself may reinforce the friendship between the two partic.i.p.ants, and thereby keep the group together-a ritual that counteracts the natural tendency of cats to regard others as rivals, not allies. The act of rubbing together also inevitably transfers scent from one cat to the other, so repeated rubbing could cause a "family odor" to build up. We know that some of the cat's carnivore relatives exchange scent via rubbing rituals: for example, badgers from the same sett create a "clan odor" by rubbing their back ends together, exchanging scent between their subcaudal glands, wax-filled pockets of skin that lie just beneath their tails.13 Cats may not deliberately exchange odor when they rub on one another; if they were doing so, they would likely concentrate their rubs on scent-producing areas of their bodies, such as the glands at the corners of their mouths, which they use to scent-mark prominent objects in their territories-but they generally don't. As such, the rubbing ritual may be mainly tactile, a reaffirmation of trust between two animals, which by acc.u.mulation reduces the likelihood of the group splintering apart.

The other social exchange that can follow the tail-up signal is mutual licking, or allogrooming. Cats spend much time licking their coats, so it is hardly surprising that when two cats lie down side by side, they often lick each other. Moreover, they tend to groom the top of the other cat's head and between the shoulders. These are areas that the supplest of cats finds hardest to groom for itself-not impossible, of course, since cats with no grooming buddy use their wrists to wipe those areas, and then in turn lick their wrists-and all cats use this method to clean their mouths after eating.

One interpretation of allogrooming holds that it is entirely accidental: two cats sitting together groom those areas that smell the least clean, oblivious that those areas belong to another cat. However, we know that allogrooming has a profound social significance in many other animals, especially in primates, in which it has been linked with pair-bonding, the building of coalitions, and in reconciliations between family members who have recently quarreled. In cats, allogrooming likely performs the same function as mutual rubbing: cementing an amicable relationship. Consistent with this is the observation that in large groups of cats consisting of more than one family, most allogrooming takes place between relatives.14 Libby grooming Lucy Some evidence shows that allogrooming does reduce conflict. In artificial colonies, such as those established by cat rescue organizations, aggression is often much less prevalent than might be expected from tensions caused by forcing unrelated animals to live together; significantly, however, allogrooming is common. Moreover, the most aggressive cats often do most of the allogrooming, implying that licking another cat may be an "apology" for a recent loss of temper. Alternatively, a cat that allows itself to be groomed does so because it remembers that it was recently attacked by the same cat, and being groomed is much more pleasant than being bitten. This latter interpretation conceives allogrooming as an alternative to aggression, so placing it in a "dominance" framework, whereby one animal controls another's activities.

Some scientists have proposed that cat societies are indeed structured according to dominance hierarchies, with larger, stronger, more experienced and more aggressive cats imposing themselves on those that are smaller, younger, or more timid. The dominance concept has long been applied to domestic dogs and their ancestor the gray wolf, but has recently been the source of much controversy. Most biologists now agree that while groups of dogs (and wolves) may sometimes appear to establish and maintain hierarchies using aggression and threat, they do so only under extreme circ.u.mstances, when their natural tendencies to form amicable relationships have been thwarted.15 As with dogs, the apparent formation of a dominance hierarchy in cats may be a result of external pressures. Social tensions arise when unrelated cats live together, either in one of the large outdoor colonies that form around large concentrations of food, such as fishing villages, or in a household with many unrelated cats obtained at different times from multiple sources. No hierarchies are apparent in small, one-family colonies, or indeed within family groups that form part of a larger colony.

Cat society is not as highly evolved as canid society. Domestic cat society is matriarchal: each unit begins with one female and her offspring, and if enough food is available on a regular basis, her daughters will stay with her; when they produce litters of their own, care of the kittens will be shared between them. This situation is more equitable-and probably less highly evolved-than occurs in wolf society, in which juveniles a.s.sist their parents in raising the next generation of cubs but refrain from breeding themselves that year.16 Moreover-and in contrast to the situation in the wolf pack, which typically consists of roughly equal numbers of males and females-male cats do not help with raising kittens. In some colonies, female cats have been observed as super-affectionate toward the resident male-presumably the father of their most recent litters-possibly regarding him as a first line of defense against infanticide by other males. A typical small cat colony might thus consist of a mother, her grownup daughters, their most recent litters of kittens, and one or two tomcats.

Family groups cannot expand ad infinitum, of course, because they will inevitably outstrip their food supply. Young males start to leave their colonies at about six months of age, sometimes maintaining a shadowy existence around the fringes for a year or two, but eventually forging out on their own to look for females elsewhere-thereby incidentally preventing inbreeding. Among females, relationships become increasingly strained as they must compete for s.p.a.ce and food. Any major event, perhaps the death of the matriarch, can trigger a breakdown in the relationships between some of the cats; aggression increases, and the once-peaceful colony may irrevocably split into two or more groups. The members of the minority group may be forced to leave the area entirely, something that will have serious consequences for them; thereafter, deprived of the resources around which the original colony formed, they are unlikely to raise many offspring to maturity. In this way, over the years the central family group will usually persist, but the more successful it becomes, the more likely that some of its female members will become outcasts. The composition of the central group is also likely to be disrupted, both by other cats trying to get access to the food and shelter that the original group are monopolizing, and also by humans, who often attempt to cull the cats as their numbers rise. As a result, cat society rarely remains stable for more than a few years at a stretch.

The benefits of cooperation, while not powerful enough to produce sophisticated social behavior, have evidently been sufficient for a limited range of social communication to have evolved: the tail-up posture, mutual rubbing, and allogrooming. In the absence of evidence to the contrary, we can reasonably a.s.sume that this change did not start until cats began their a.s.sociation with humankind, some 10,000 years ago. If so, it has happened remarkably quickly, but not impossibly so. Although we usually conceive of evolution by natural selection operating over time scales of hundreds of thousands or even millions of years, examples of exceptionally rapid change or "explosive speciation" have recently been doc.u.mented in wild animals that have stumbled on new and hitherto unexploited environments, such that completely new species can emerge in just a few hundred generations.17 Furthermore, if the tail-up signal has evolved from a posture performed by kittens into a sign whose meaning is recognized by every adult cat, then it is the only doc.u.mented example of a new signal having evolved as a consequence of domestication; every other doc.u.mented domestic species communicates using a subset of signals performed by its wild ancestor.

Male domestic cats, in contrast to females, appear largely untouched by domestication, apart from their capacity to become socialized to people when they are kittens. Each one precisely resembles Rudyard Kipling's "cat that walked by himself."18 Unlike lions and cheetahs, male domestic cats do not form alliances with one another, remaining resolutely compet.i.tive for the whole of their lives-which, as a result, are often both eventful and rather short. Female cats (and neutered males) try to avoid one another when they can, but when two males meet, and neither wishes to back down, their fighting can be brutal (see box on page 176, "Bluff and Bl.u.s.ter").

Tomcat spraying Bluff and Bl.u.s.ter When two rival cats meet, each will try to persuade the other to back down without either having to resort to physical contact. Cats are too well armed to risk fighting unless this is unavoidable; they therefore resort to adopting postures that attempt to persuade their opponent that they are bigger than they really are.

Each cat will draw itself up to its full height, turn partially sideways, and make its hair stand on end, al