Cognitive Ethology and the Empirical Analysis of Nonhuman Social Behavior*

Etolog(a, 3:23-39 (1993)

Cognitive ethology and the empirical analysis of nonhuman social behavior*

M. Bekoff**

Departmentof Environmental, Population and OrganismicBiology, University of Colorado Boulder, Colorado 80309-0334 USA

ABSTRACT. Cognitive ethology and the empirical analysis of nonhuman social behavior.- As a relatively new interdisciplinary science, cognitive ethology is under attack with respect to its scientific status. However, there also are many supporters of cognitive ethological research. In this paper I consider (1) the sorts of topics in which cognitive ethologists are interested, (2) some connections between cognitive analyses of social behavior and philosophical concepts including intentionality, representation, and folk psychology, (3) different views of cognitive ethology, (4) recent work on social play and antipredator vigilance that seem to benefit from taking a cognitive perspective, and (5) what is gained by taking a cognitive approach to the study of social behavior and what is lost by not doing so. Cognitive ethology has a bright future, and has much to gain froma broad interdisciplinary perspective.

KEYWORDS: Cognitive ethology, Animal cognition, Animal minds, Social play, Vigilance

Cognitive ethology: a preliminary (1992, 1993a, b), Beer (1992), Caro & Hauser agenda (1992), Whiten & Ham (1992), Allen & Hauser (1991, 1993), Bekoff (1993a, b), Jamieson & Bekoff (1993), Heyes (1993a, b), Wemelsfelder There is a lot of interest in cognitive ethology (1993), and Bekoff & Allen (1994). Renewed by people who come to their views from vastly interest in cognitive ethology has signified a return different perspectives. Many examples and/or to many of the ideas of CharlesDarwin and early detailed arguments can be found in Griffin (1976, anecdotalcognitivists (Jamieson & Bekoff, 1983), 1984, 1992), Wyers (1985), Byrne & Whiten especially their appeals to evolutionarytheory, their (1988), Cheney & Seyfarth (1990, 1992), Burghardt close association with natural history, and their (1991), deWaal (1991), Ristau (1991a), Allen reliance on anecdote and anthropomorphism to

*Parts of this paper were presented in the plenary session on comparative cognition at the XXIII International Ethological Conference held in Torremolinos, Spain, September 1993 **Mailing address: 296 Canyonside Drive, Boulder, Colorado 80302 USA Telephone: 303-443-6857; Fax: 303-443-2275; e-mail: [email protected]

23 Bekoff

inform and motivate more rigorous study. select representatives of limited taxa, (4) favor Comparative cognitive ethology is also an observations and experiments (only those that meet important extension of classical ethology because it the strictest guidelines for animal welfare) in explicitly licenses hypotheses about the internal conditions that are as close as possible to the natural states of animals in the tradition of Tinbergen environment where selection has occurred, and (5) (1951, 1963) and Lorenz (1981). maintain that field studies of animals that include In the present essay, research on non primates careful observation and experimentation can inform will primarily be considered. Many people inform studies of animal cognition, and that cognitive their views of cognitive ethology by using the same ethology will not have to brought into the studies over and over again --most are on nonhuman laboratory to make it respectable. While some are of primates-- and they often forget that there are many the opinion that advanced cognition is confined to other nonhuman animals (hereafter animals) who the laboratory (e.g. Premack, 1988, pp. 171-172), also show interesting patterns of behavior that lend those who have studied animals in the wild disagree themselves to cognitive analyses (Beck, 1982). (e.g. de Waal, 1991, p. 311; McGrew, 1992, pp. Furthermore, people often fail to recognize that in 83ff). many instances sweeping generalizations about the Two notions to which cognitive ethologists pay cognitive skills ( or lack thereof) of species of a lot of attention are the needs of individual primates (and not of individuals) are based on small organisms in the habitats in which they evolved or data sets from a limited number of individuals in which they currentlyreside, and also the animals' representing fewtaxa (Platt, et al., 1991; Swartz & perceptual worlds, or timwelts. A concentration on Evans, 1991; Bekoff, et al., 1993; Heyes, 1993a; individuals and not on species should form an Wemelsfelder, 1993). important part of the agenda for future research in Although I am enthusiastic about the field of cognitive ethology. There is a lot of individual cognitive ethology, I do not believe that all variation in behavior within species and sweeping behavior patterns will benefit from a cognitive generalizations about what an individual ought to oo analysis, but I feel that some surely will benefit. It because she belongs to a given species must be is narrow-minded both to think that no analyses of taken with great caution. In contrast to cognitive behavior will benefit or that all analyses will benefit ethologists, cognitive psychologists typically work from a cognitive approach. I also strongly favor on related topics in laboratory settings, and do not empirical research and the careful use of anecdotal emphasize comparative or evolutionary aspects of data --anecdotes are data-- toinform these endeavors. animal cognition. When cognitive psychologists oo make cross-species comparisons, they are usually interestedin explaining differentbehavior patterns in What are cognitive ethologists interested terms of common underlying mechanisms; in? ethologists, in common with other biologists, are often more concernedwith the diversity of solutions that living organisms have found for common Cognitive ethologists (1) are interested in problems. comparing thought processes, consciousness, Given their interest in the evolution of beliefs, and rationality in animals, (2) are concerned cognition, cognitive ethologists also pay attention with claims about the evolution of cognitive to mental continuity between humans and other processes in animals, (3) emphasize broad animals. They note that evolutionary biologists talk taxonomic comparisons and do not focus on a few about the evolution of hearts, lungs, kidneys, and

24 Etolog(a, Vol. 3, 1993

stomachs, and even brains, and cognitive ethologists behave intentionally when she has propositional recognize that if brains are associated in some way attitudes such as beliefs about something, or desires with minds and thinking then there is something to of something. One of the major problems and be gained by looking at possible continuity between challenges in cognitive ethology and the study of human and nonhuman animals. Of course, this view cognition in general is getting from behavior to does not mean that there is continuity between all mental content (Allen, 1992, 1993a). taxa, but it does mean that it would be wrong to While Lorenz, Tinbergen, and others have used look at continuity between structureand function in terms such as "intention movements," they areused some systems and not in others. quite differentlythan in the philosophical sense. Lorenz and Tinbergen were referring more to preparatory movements that might communicate Some philosophical underpinnings what individuals were likely to do next, and not of cognitive ethology necessarily their beliefs and desires, although one might suppose that the individuals did indeed want to fly and believed that if they moved their wings, Cognitive studies lead us to consider questions they would fly. This distinction is important to such as "What does it mean to claim that Jethro(my mark because the use of these terms do not companion dog) behaves intentionally? "What does necessarily add a cognitive dimension to classical it mean to claim that he has a belief?" "What does it ethological notions, although they could. mean to have a representation of something?" For Representation is generally is taken to mean that example, in my work on social play I am interested something stands for something as being in a in questions such as "Does Jethro want Henrietta to certain way (for detailed discussions see Sterelny, play with him?" Does Jethro believe that Henrietta 1990 and Perner, 1991). Thus, in a cognitive wants to play with him?" "Does Jethro believe that ethologicalinquiry about antipredatory vigilance, we Henrietta believes that he wants to play with her?" can ask if a bird represents his flock to himself as And, in my work on antipredatory vigilance I am being in a certain geometric distributionor as being interested in questions that center on if and how of a certainsize (see below). birds represent their individuals in their flocks to Finally, folk psychology can be thought of as a themselves. These are questions for empirical common sense theory of mind embeddedin the cognitive studies to answer. wisdom of the common folk. Folk psychological Three major areas in philosophy of mind are explanations appeal to intentionality-beliefs and important to cognitive ethology, namely, desires. intentionality, representation,and folk psychology While there is some concern about the (Millikan, 1984, 1993; Brand & Harnish, 1986; shortcomings of folk psychological explanations Bogdan, 1986; Sterelny, 1990; Godfrey-Smith, and how these may influence research in cognitive 1991; Clark, 1993; Goldman, 1993). One major ethology (Beer, 1992), there are more and more of question in philosophy of mind concerns "how those who feel very comfortable using folk thoughts and other mental states get their psychological explanations as aids to understanding intentional or representational character" (Goldman, (Cling, 1991; Saide], 1992; Christensen & Turner, 1993, p. 76). Intentionality, in the limited 1993; Goldman, 1993). Eliminativists, who deny philosophical sense, means that mental states have that there areany propositional attitudes, providethe content; they are about things, and may be noted as hardest line against folk psychology. These x believes that p. Thus, an individual can be said to individuals take the stand that neural events and

25 Bekoff

neural properties are all there is. However, those would now accept that, no matter how difficult they who have studied not only philosophy of mind but may be to study, it is no longer possible to deny the also animal behavior disagree with their position. existence of true thought processes and even For example, Bennett (1993, p. 385) notes that consciousness of some type in some mammals and "attributions of beliefs and desires are nothing birds." (Manning & Dawkins, 1992, p. 143) unless they help us to explain behavior. The The major problems that cognitive ethology intentional concepts that we ordinarily use, the ones faces are those that center on methods of data that are defined by folk psychology, simply are collection and analysis, and on the description, explanatory, and attempts to analyze them come to interpretation, and explanation of behavior (Bekoff grief if they don't give a central placeto that fact." & Jamieson, 1990a, b; Jamieson & Bekoff, 1993). Kim Sterelny (personal communication) has It is of interest to note that while some people feel suggested that perhaps we should concentrateon the very comfortable with evolutionaryor adaptationist mental lives of the animals themselves, and not try explanations that can be rather tenuous or vague, to apply human-based folk psychology to they also feel comfortable rejecting outright nonhumans; rather, for example, we might try to cognitive explanations that are not necessarily any characterizethe folkpsychology of dogs. more questionable. In a recent attempt to analyze Saidel (1992) also points out that in many how cognitive ethology is viewed, we have instances, a folk psychological explanation does the identified three major groups whose members have work it is supposed to do, whereas an appeal to differentperspectives of cognitive ethology (but also neural events does not. For example, it is unlikely among some of whom there are blurred that Eric and Marc arein the same neural state when distinctions), namely, slayers, skeptics, and they are running out of a burning building, but it is proponents (for discussion see Bekoff & Allen, highly likely that they will give the same 1994). The positions that members of each of these explanation of why they ran out of the burning categories usually take can be summarized as building--they didn't want to get burned; This follows: argument that it is highly improbable that a given Slayers: Slayers deny any possibility of success mental property has the same neural form in all in cognitive ethology. Sometimes slayers conflate organisms is called the argument from multiple the difficultyof doing rigorous cognitive ethological realizability (Goldman, 1993). investigations with the impossibility of doing so. Of course, there are many other points of Slayers also often ignore specific details of work by connection between cognitive ethology and cognitive ethologists and frequently mount philosophy of mind. For example, I have not said philosophically motivated objections to the anything about consciousness or information possibility of learning anything about animal processing. These will be considered below, but cognition. Slayers do not believe that cognitive only brieflyand indirectly. ethologicalapproaches can lead, or have lead, to new and testable hypotheses. They often pick out the most difficult and least accessible phenomena to Some different views on animal study (e.g. consciousness) and then conclude that minds: slayers, skeptics, and because we can gain little detailed knowledge about proponents this subject, we cannot do better in other areas. Slayers also appeal to parsimony in explanations of animal behavior, but they dismiss the possibility "Probably a majority of animal behaviour workers that cognitive explanations can be more

26 Etologfa, Vol. 3, 1993

parsimonious than noncognitive alternatives, and Bekoff & Allen (1994). Two examples of the they deny the utility of cognitive hypotheses for slayer's position that are of interest to cognitive directing empirical research. Slayers are also ethologists are Heyes's (1987) and Wasserman's troubled by folk psychological, anthropomorphic, (1993) positions. Heyes, who is a laboratory and cognitive explanations. psychologist, denies that evidence gained by Skeptics: Skeptics are often difficult to observing animals in natural settings is particularly categorize. They are a bit more open-minded than relevant to understanding animal minds. She advises slayers, and there seems to be greater variation cognitive ethologists to hang up their field glasses among skeptical views of cognitive ethology than and turn to laboratory research if they want to among slayers' opinions. However, some skeptics understand animal cognition. She writes: recognize some past and present successes in "It is perhaps at this moment that the cognitive cognitive ethology, and remain cautiously ethologist decides to hang up his field glasses, become optimistic about futuresuccesses; in these instances a cognitive psychologist, and have nothing furtherto they resemble moderate proponents. Many skeptics do with talk about consciousness or intention." appeal to the future of neuroscience, and claim that (Heyes, 1987, p. 124) when we know all there is to know about nervous systems, cognitive ethology will be superfluous. Heyes (1993c), in a well-argued paper, has Like slayers, skeptics frequently conflate the extended some of her claims more forcefully. Not difficulty of doing rigorous cognitive ethological only does she believe "that there is still no investigations with the impossibility of doing convincing evidence of mental state attribution in so. Skeptics also find folk psychological, non-human animals," but she also maintains "that anthropomorphic, and cognitive explanations to be most current methods of investigation do not have problematic. the potential to provide such evidence" (p. 177). Her Proponents: Proponents recognize the utility of latest paper is a challenge that should be accepted cognitive ethological investigations. They claim with open arms by those who embrace cognitive that there are already many successes and they see ethology. In her criticisms, Heyes concentrates that cognitive ethological approaches have provided heavily on laboratory studies. One possibility that new and interesting data that also can inform and Heyes needs to consider in more detail is that there motivate further study. Proponents also accept the areevents that occurin the wild, that while difficult cautious use of folk psychological and cognitive to study, may help to provide some of the sorts of explanations to build a systematic explanatory evidence that she claims is lacking. framework in conjunction with empirical studies, Wasserman's (1993) position on cognitive and do not find anecdotes or anthropomorphism to ethology is poorly argued and ill-informed when be thoroughly off-putting. Some proponents are as compared to Heyes's. Wasserman makes no attempt extreme in their advocacyof cognitive ethology as to discuss critically current literature and feels some slayers are in their opposition. But most comfortable with the claim that "Griffin's (1978) proponents are willing to be critical of cognitive call for a cognitive ethology appears to be a ethological research without dooming the field throwback to a prescientific analysis of behavior in prematurely; if cognitive ethology is to die, it will terms of conscious experience" (p. 223; my be of natural causes and not as a result of hasty emphasis). Furthermore, Wasserman concludes slayings. that "There is simply no clear or necessary role Representative views of people who have been for subjectiveexperience to play in behavior ... " (p. placed in each group are discussed in detail by 223) and that the study of mental experience "might

27 Bekoff

not be a fitting topic for scientificinquiry .." (p. are likely to be more effective in convincing other 223; my emphasis). While there is equivocationin scientists of the significance of the abilities of these claims, Wasserman does not tell us about the nonhuman animals." sources underlying this uncertainty. Nonetheless, Proponents are more optimistic in their views cognitive ethology is dismissed as an unsuitable about the contributions that the field of cognitive candidate for a viable field of scientific inquiry. ethology and its reliance on field work and on Other slayers, like Williams (1992), argue against comparativeecologica l and evolutionary studies can the study of animal cognition on the basis of a make to the study of animal cognition in terms of philosophical view about the privacy of the mental opening up new areas of research (e.g., Ristau, or by the related "other minds" problem. These 1991b, p. 102) and reconsidering old data. A critics typically do not give specific critiques of summaryof their views is as follows: actual empirical investigations carried out by "At this point, however, cognitive ethologists can cognitive ethologists; rather they try to dismiss console themselves with the knowledge that their such investigations on philosophical grounds alone. discipline is an aspect of the broader fieldof cognitive Still others, such as Rosenberg (1990), pay careful studies andconceptually may not be in any worse shape than highly regarded, related fields suchas cognitive attention to actual work in behavioral biology, psychology. We area long way from understandingthe specificallythe study of play, but they belittle the natural history of the mind, but in our view this idea that it might be possible to attribute human amounts to a scientific challenge rather than grounds concepts to nonhumans. Rosenberg's argument for depression or dismissal." (Jamieson & Bekoff against a science of intentionality is based on 1992a, p. 81) considerations that seem to conflatethe difficultyof specifying the content of intentional states with the impossibility of doing so (Allen, 1992; Allen & Social play and antipredatory Bekoff, 1993). Specifically, Rosenberg appeals to vigilance: what might individuals his inability to imagine what organisms could do to know about themselves and others? allow the attribution of certain concepts; the difficulty of thinking of suitable experiments is a Social play and antipredatory vigilance are two challenge, not necessarily a barrier to concept very differentbehavior patterns that lend themselves attribution (Allen & Hauser, 1991). to cognitive studies. Both areas lend themselves to In general, skeptics are a bit more open-minded empirical analysis, and both are connected to than slayers. With respect to the sorts of different notions such as intentionality and explanations that are offeredin studies of animal representation. Furthermore, in studies of social cognition, many slayers and some skeptics favor play and vigilance, folk psychological explanations noncognitive explanations because they believe have been useful for informing and motivating them more parsimonious and more accurate than furtherresearch, and these sorts of explanations have cognitive alternatives, and less off-putting to others also turned out to be very good predictors of who do not hold the fieldof cognitive ethology in behavior. high esteem. For example, Snowdon (1991, p. 814) claims that: "It is possible to explore the cognitive capacities The communication of play intention of nonhuman animals without recourse to mentalistic concepts such as consciousness, intentionality, and deception. Studies that avoid mentalistic terminology Some believe that play may provide more

28 Etolog(a, Vol. 3, 1993

prom1smg evidence of animal minds than many beginning or towards the middle of playful other behavior patterns, and analyses of play present encounters. Thus, the latter two messages--"this is a great challenge to researchers (Mitchell, 1990; still play," or "let's play again"--may be sent aftera Bekoff & Allen, 1992; Allen & Bekoff, 1993; very short break or after rough play has occurred. Wemelsfelder, 1993). It would have been Supporting evidence concerning the importance of unfortunateif people decidedthat just because play play signals for allowing cooperativesocial play to was difficult to study, it was impossible to study. occur comes from studies in which it is shown that When animals play, they typically use actions play-soliciting signals show little variability in patterns that are also used in other contexts such as form or temporal characteristics, and that they are predatory behavior, antipredatory behavior, and used almost solely in the context of play (Bekoff, mating (Bekoff & Byers, 1981; Fagen, 1981). 1977). These features of bows can be related to the Because play is typically composed of motor fact that when engaging in social play, canids patterns that are also used in a variety of other typically use action patterns that are also used in contexts, an individual needs to be able to other contexts such as predatory behavior, agonistic communicate to potential play partners that he is encounters, or mating, where misinterpretation of not trying to dominate them, eat them, or mate with play intention could be injurious. them. Analyses of play sequences may also inform In most species in which play has been future studies of social play. For example, in described, play-soliciting signals appear to foster intraspecific comparisons, it has been found that some sort of cooperation between players so that sequences of social play are usually more variable each responds to the other in a way consistent with than sequences of nonplay behavior (Bekoff & play and different from the responses the same Byers, 1981). Is it possible that animals rea:l actions would elicit in other contexts. For example, differences in behavioral sequences that are in coyotes, the response to a threat gesture after a performed duringplay and in other contexts? Might play bow had immediately precededthe threat or a increased (or consistent variations in) variability in play signal had been performed in the beginning of sequences also (along with play signals) that convey an interaction, is different from the response to the message "this is play" and enable individuals to threat in the absence of any preceding play signal predict what is likely to occur or to understand what (Bekoff,1975). The play bow somehow altered the has already occurred. We can also ask if there are meaning of a threat signal by establishing (or differencesin the variability and structure of play maintaining) a "play mood." There is much room sequences that are preceded by a play signal when here for a theory for mental attribution and compared to play bouts that are not? explanations that appeal to intentionality (Bekoff, There might also be important species 1993a; Jamieson & Bekoff, 1993). differences,especially when comparingwild canids. Let's consider in more detail the question of For example, the play sequences of infant coyotes whether or not signals that appear to be used to were considerably more variable than those of infant communicate play-intention could foster the wolves, and one wonders if this is because infant cooperation among participants that is necessary for coyotes are more aggressive towards one another play to occur. It generally is assumed that such than are infantwolves. What is also very interesting play-soliciting signals transmit messages such as is that the coefficients of variation for duration and "what followsis play," "this is still play," or "let's form for play bows performed at the beginning of play again." In canids, play bows also occur play sequences are also lower for infant coyotes throughout play sequences, but usually at the when comparedto infant wolves, young dogs, and

29 Bekoff

adult dogs, and the coefficient of variation for the dogs or coyotes looked surprisedwhen, on a very form of the play bow is also lower for bows rare occasion, a bow resulted in the recipient perfonnedduring a play bout for infant coyotes attacking the signaler. The soliciting animal's eyes when compared to other canids for which there are opened widely, her tail dropped, and she rapidly similar data. The differencebetween infant coyotes turned away from the noncooperating animal to and the next closest group (infant wolves) is also whom she directed a play-soliciting signal, as if larger than the differencebetween the other groups. what happened was totally unexpected and perhaps Recall also that the play sequences of infant coyotes confusing. After moving away the surprised animal were much more variable than the play sequences for often lookedat the other individual, cockedher head the infant wolves. to one side, squinted, and furrowedher brow, and Here then, at least, is some compelling evidence seemed to be saying "I want to play, this is not that play bows might actually influenceon-going what I wanted to happen." play, so that it remains play. Young coyotes, being With respect to the solicitor's beliefs about the more aggressive than other young canids, use play future, detailed analyses of movie film also show bows to keep a cap on aggressive tendencies so that that on some occasions, a soliciting animal begins they can engage in social play. Colin Allen to perform another behavior before the recipient (personal communication) has suggested that as commits himself. The solicitor behaves as if she animals become fatigued, they may have lower expects that something specific will happen and thresholds for certain sorts of play spilling over into commits herself to this course of action. A major serious aggressive encounters. question, then, is how to operationalize these questions; what would be convincing data?How ch we know when we have an instance of a given Future research behavior(s)?Thus, we need to ask what is frustrated, what is the goal, and what is the belief about? In attempts to answer these and other questions, there A cognitive perspective will be very useful in simply is no substitute for detailed descriptions of future analyses of social play. Comparative subtle behavior patterns (Bekoff, 1992; Golani, observations strongly suggest that individuals 1992). expect that play will ensue if they perfonn a bow; they act as if they want play to occur. On many occasions it looks as if animals are frustrated or Antipredatory vigilance: the importance surprisedwhen their bow is not reciprocated in a of failure way that is consistent with their belief about what is most likely to occur, namely, social play. Frustration also may be inferredfrom the common In studies of vigilance it generally is assumed, observation that canids and other mammals are very for simplicity's sake, that individuals compromise persistent in their attempts to get others to play their ability to detect predators when feeding with with them and they often engage in some sort of their heads down, and compromise their food intake self-play such as tail-chasing, after a bow or other when scanning for predators with their heads up. play invitation signal is ignored, or they rapidly run Thus, it has been argued that there are good reasons over to another individual and try to get them to for individuals to live, or at least to foragefor food, play. in groups, if doing so increases the probability of Surprise is more difficult to deal with, but often detecting a predator or reduces the time spent

30 Etolog{a, Vol. 3, 1993

scanning forpredators, thus permitting more time to are concerned with what an individual might know be spent doing other things. about himself and others would involve asking at A very popular question in the comparative study least the following questions, all of which are of vigilance is "How does the behavior of interconnectedand all of which lend themselves to individuals vary in groups of different sizes?" empirical study. While some of the following Another, and perhaps more important question to questions are not directly related to cognitive which very little attention has been directed, is inquiry, all can inform and motivate such an "Why does the relationship between group size and approach. Some are also very basic, but this return scanning rates where it is not expected to fail?" to basics is necessary. These data should not be viewed as negative data or (1) What is a group? What does it mean to say noise. Rather, they should be used to inform and that an individual is a member of a group and is our motivate new researchas well as reanalyses of old conception of group the same as that of the data. Generally, it has been found that there is a animals? In studies of vigilance, variations in group negative relationship between group size and rates of size are often used to explain variations in other scanning by individuals and a positive relationship patterns of behavior, and precise measurements of between group size and the probability of predator group size are essential. Questions that inform the detection. This is because there are more eyes and conception of group membership include "Is there a perhaps other sense organs (e.g., Sullivan, 1984) critical distance between individuals below which we that can be used to scan foror to detect predators. can say with some degree of certainty that they are Elgar (1989) notes that although the negative members of the same group?" and "Do individuals relationship between group size and individual have to spend a certain amount of time together scanning rate is quite robust and is approaching the within a certain distance to justify calling them a status of dogma (Lima, 1990), few studies have group? With respect to studies of vigilance, Elgar, actually controlledfor confounding variables, such Burren, & Posen (1984) found that a house sparrow as variation in the density and type of food who was in visual contact with other house resources, group composition, ambient temperature sparrows but separated by 1.2 meters scanned as if and time of day, proximity to a safeplace and to the she was alone. observer, visibility within the habitat, and group (2) Does the size of a group or the geometric composition (see also Lima, 1990 and Lima & Dill, distribution or orientation of individuals influence 1991). With respect to whether or not an individual individual vigilance? is really being vigilant, Lazarus ( 1990, p. 65) notes It is important to keep in mind that there are that "... researchershave simply assumed that the confounding variables such as the geometric behaviour in question is vigilance, and have then relationships among group members and how sought its function" (see also Valone, 1989, 1993). individuals are oriented in space--in a circle, a Lima (1990) also claims that there seem to be no triangle, or in a straight line, for example--that studies that have directly examined the question of might influence scanning rates of individuals. Little whether foragerspay any attention to the behavior attention has been paid to group geometry. Some of other group members. He also concludes that authors write about visual obstructions but do not very little is known about the perceptions of the consider the actual geometry of the group. For animals being studied and that many models of example, Elgar (1989) does not directly refer to vigilance reflect mainly the perceptions of the geometry as a variable influencing scanning for modelers themselves (p. 262). predators, but he does write about visual A cognitive analysis of vigilance in which we obstructions, in terms of how they might influence

31 Bekoff

vigilance and risk of predation. Likewise, in his varies seasonally), as it becomes more difficult (at review of vigilance in mammals, Quenette (1991) least on my view) for a bird to see other members of writes about visual obstructions and their effecton a flock, (I) the birds move their bodies more to vigilance because they influencehow information is orient themselves to other birds, (2) they change received from the environment. Elgar et al. (1984, their relative positions more frequently, (3) their p. 221) report data that strongly suggest that in head movements are more random, and ( 4) they house sparrows, "it is necessary forthem to be able spenda greater percentage of time being vigilant. to continuously see their flockmates."Further, they Questions such as "How does a bird or other review literature that shows that in general, nonhuman assess group geometry?"also need to be scanning rates in small passerines do not decrease considered.;it is not known if and how individuals significantly with flocks larger than eight or nine actually assess the geometry of the group in which birds. They write: "It is possible that sparrows they are a member. simply cannot estimate the number of birds in larger (3) Do changes in group size or geometry flocks ... " Of course, it also is possible that birds influence patterns of social interactions? It is and other animals cannot estimate the number of possible that as group size and geometry change, birds in flocks that are organized in a way such that either singly or together, there is also a change in visual inspection is difficultor impossible. There is how individuals interact. If this is the case, then it a lot of work still to be done. might be possible for an individual to gain One important point about which I could find information about these variations from changes in very little information involves how the geometry encounter patterns without having to read them of the flock of potential prey influences predator directly, as has been suggested for ants (Gordon, detection. There may well be trade-offssuch that Paul, & Thorpe, 1993). To the best of my although it is easier to see what other flock knowledge, there are no data forbirds or mammals members are doing when the birds are arrangedin a that can be used to answer these questions with any specific geometric array, it also is easier for degree of certainty. potential predators to see them. Answers to the question "How does the geometric distribution of individuals influence Future research individual vigilance?" will likely have something to say about animal cognitive abilities. Thus, while it is known that the location of an individual in her Some other interesting questions that may be group (center or periphery) can influence her pattern informed by a cognitive approach include: "Do of vigilance, it remains to be studied how the individuals change their relative position in a group geometry of the whole group influences the ease to make it more likely that they could feedmore with which an individual is able to assess what efficiently and/or detect potential predators more others are doing by seeing or hearing them. For easily?" example, it seems that it would be easier to see "Is this a cooperative endeavor; do individuals what others are doing if individuals were organized move synchronously?" Also, "Does one's position in a circle rather than in a straight line, but this is in a group influence influencewhether he can assess not known. My working hypotheses are that in changes in group size or geometry?" Here I am flocks of birds (including mixed flocks containing asking if and how the location of an individual in a house finches, steller's jays, juncos, cowbirds, group makes it easier or more difficultto know how evening and black-headed grosbeaks; composition many other individuals are there and how they are

32 Etologfa, Vol. 3, 1993

distributedin space. It might be very useful for an that we possess, many of which are not used very individual to be able to see what others are doing, often at all. for while scanning, an individual might also pick up Now, what are some reasons for advocating and store information about what individuals in a cognitive analyses and explanations of animal particular part of the group are most likely to be behavior? Why might cognitive explanations be the doing, or she might generalize from her own best explanations to which we can appeal to help us previous experience in that partof the group to what come to terms with the study of animal minds. others are most likely to be doing when they are in There are many reasons why people are interested in that position. If we can get answers to these sorts of the study of animal cognition. That the explanatory questions, we might be able to assess if it is power of our theorizing is increased applies without possible that the inverse relationship between group question. Further, it is obvious that a cognitive size and individual scanning rate levels offor fails approach will allow forthe generating and testing of because of the inability of individuals to monitor new ideas, and perhaps, the reviewing and the behavior of "too many other animals" who reanalyzing of old data. might also be difficultto see. If thereare "too many While each of the following reasons does not animals," it is possible that large groups may break necessarily warrant a cognitive approach, taken up into smaller subgroups (Valone, 1993). together they justify the currentinterest in cognitive Furthermore, it is possible that birds may also be ethology. Further, these are not the only reasons using auditory (or other) cues to obtain information that one might offer, just some suggestions that about other flock members (Sullivan, 1984). take into account a number of factors. These include the following (in no order of importance). ( 1) Many models in ethology and behavioral ecology presuppose cognition. Thus, it would be The advantages of an evolutionary, especially useful to have informed ideas about the comparative, and ecological types of knowledge that nonhumans have about cognitive ethology? their social and nonsocial environments and how they use this information (Allen & Hauser, 1991). (2) It may be more economical or parsimonious Two important questions need to be given to assume that not everything that an individual serious attention. These are "What is gained by needs to be able to do in all situations in which he appealing to the possibility of cognitive finds himself is preprogrammed. While general rules explanations?" and "What is lost by dismissing the of thumb may be laid down genetically during possibility of cognitive explanations?" Answering evolution, specific rules of conduct that account for the second question first, it seems clear that we all possible contingencies are too numerous to be would lose a lot of information about the possible hard-wired (Griffin, 1984). Behavioristic learning richness of animals' lives if we ignored the schemes can account for some flexibility in possibilities that they behaved intentionally on organisms, but learning at high degrees of some occasions. Even if we discover that some abstraction from sensory stimulation seems less animals do not appear all that cognitive, this does amenable to behavioristic analysis (Allen & Hauser, not mean that they are not cognitive at all. Perhaps, 1991). Cognitive models of learning provide because of their relatively impoverished cognitive explanatory schemes for such cases. We can also abilities, the few skills that they have are more addthat it might actually be more parsimonious to important to them than many of the cognitive skills appeal to cognitive explanations in terms of

33 Bekoff

accounting for complex patterns of behavior with go about doing our research. If it is because fewerexplanations. philosophers do not have the experience with (3) The assumption of animal minds leads to empirical work that allows them to make realistic more rigorous empirical analyses of behavioral suggestions for experimental design, then it would plasticity and flexibility in the many and diverse be useful for philosophers to watch ethologists at situations--both social and nonsocial--that many work (Dennett, 1987, 1988). This experience might nonhumans regularly encounter. Further, Yoerg allow philosophers to gain a better understanding of (1991) argues that considerations of cognitive what cognitive ethology is all about. Even then, it function can lead to original ideas about behavioral may be the case that ethologists are ill-advised to adaptation. look to philosophers for a crisp and empirically (4) By providing different perspectives on rigorous definition of intentionality (for example), behavior, cognitive ethology can raise new even if some philosophers promise to provide one questions that may be approached from other levels (Colin Allen, personal communication). of analysis. For example, appeals to detailed Philosophical theories are only starting points for descriptive information about subtle behavior empirical investigation, so no one should be too patterns (Bekoff, 1992; Golani, 1992) and to wedded to any particular theory. Nonetheless, neuroethological data (e.g. Dennett, 1991; starting points are necessary, so the philosophical Dusenbery, 1992; Griffin, 1992; Bullock, 1993; theories are useful (Allen, 1993b). Hauser, 1993; Lahav, 1993; Stein & Meredith, Obviously, taking a strong cognitive stance will 1993; Wemelsfelder, 1993) may be important for not be a deterrent to learning more about animal informing further studies in animal cognition, and behavior and animal minds. Detailed observational might also be useful for explaining data that are and experimental researchin cognitive ethology is already available. Such analyses will not make possible and should be encouraged; field work cognitive ethological investigations superfluous(as should be pursued wherever possible. Cognitive some eliminativists think}, because behavioral ethologists should not hang up their field glasses evidence is primary over anatomical or and have nothing to do with talk about nonhuman physiological data in assessments of cognitive intentional behavior. Rather, cognitive ethologists abilities (Allen, 1990). should welcome the fact that they are dealing with (5) Welfare issues are tightly connected to views difficult and important questions. Subjectivity, on the cognitive abilities of nonhumans (Bekoff& indeterminacy, and intractability present stumbling Jamieson, 1991; Jamieson & Bekoff, 1992b; blocks, but not impenetrable barriers. As Lawrence Bekoff, 1994) and humans (Dresser, 1993). Shapiro (personal communication) points out, the To these we can also add that we will be able to "privacy" of mental states does not necessarily test Charles Darwin's and others' ideas about mental present more or less of a problem for cognitive continuity. ethology than the invisibility of electrons does for Interdisciplinary efforts are essential in our quest chemistry. Claims about minds and electrons are for knowledge about animal minds. In these joint posited abductively because they make the most efforts, keeping an open mind would also be useful sense of the data that have been collected. To allow at this stage of the game. Philosophers need to be such posits in chemistry but not in cognitive clear when they tell us about what they think about ethology is to adopt a double standard, which is animal minds and those who carefully study the unfair to cognitive ethology. Shapiro recommends behavior of nonhumans need to tell philosophers that cognitive ethologists no longer worry about what we know, what we are able to do, and how we subjectivity, because this characterizationserves to

34 Etolog{a, Vol. 3, 1993

stigmatize their subject matter, distinguishing it in Millikan, Susan Townsend, John Lazarus, Cecelia a way that leads to an undeserved skepticism. Heyes, Daniel Dennett, Gordon Burghardt, Charles In conclusion, while many of the problems Snowdon, and especially Thomas Valone and Colin facing cognitive ethologists are difficult and do not Allen for discussing many of these issues with me. submit to easy answers, they are not necessarily Thomas Valone and Colin Allen provided useful impossible problems with which to deal. The comments on an ancestral version of this essay. challenges put forth by slayers and skeptics should be accepted by proponents of cognitive ethology. In the future, cognitive ethologists most likely will be References pursuing the challenging questions that confront them, rather than looking for other work. Allen, C., 1990. Animal consciousness. Paper presented at the Central Division of the Resumen American Philosophical Association. Allen, C., 1992. Mental content and evolutionary explanation. Bio. Phil., 7:1-12. Etolog{a cognitiva y el analisis empfrico del Allen, C., 1993a. Mental content. Brit. J. Phil. comportamiento social no humano. Sci., 43:537-553. La etologfa cognitiva, como ciencia Allen, C., 1993b. Intentionality: Natural and interdisciplinar relativamente nueva, esta siendo artifical. In: Comparative Approaches to atacada con respecto a su estatus cientffico. Sin Cognitive Science. (J.-A. Meyer & H. Roitblat, embargo, existen tambien muchos apoyos a la Eds.), Cambridge: MITPress. investigaci6n cognitiva etol6gica. En este trabajo se Allen, C. & Bekoff, M. 1993. Intentionality, social consideran (1) los tipos de problemas en los que play, and definition.Bio. Phil., In press. estan interesados los et6logos cognitivos, (2) Allen, C. & Hauser, M.D., 1991. Concept algunas conexiones entre analisis cognitivo del attribution in nonhuman animals: Theoretical comportamiento social y conceptos filos6ficos and methodological problems in ascribing incluyendo intencionalidad, representaci6n y complex mental processes. Phil. Sci., 58:221- psicologfa popular, (3) visiones diferentes de la 240. etologfa cognitiva, ( 4) trabajos recientessobre juego Allen, C. & Hauser, M.D., 1993. Communication social y vigilancia antidepredadora que parece se and cognition: Is information the connection? beneficiande la perspectiva cognitiva, y (5) que se Philosophy of Science Association, 2: In press. gana considerando una aproximaci6n cognitiva al Beck, B.B., 1982. Chimpocentrism: Bias in estudio del comportamiento social y que se pierde no cognitive ethology. J. Human Evol., 11:3-17. haciendolo. La etologfa cognitiva tiene un brillante Beer, C., 1992. Conceptual issues in cognitive futuro y mucho que ganar desde una amplia ethology. Adv. Study Behav., 21: 69-109 perspectiva interdisciplinar. Bekoff, M., 1975. The communication of play intention: Are play signals functional? Semiotica, 15:231-239. Acknowledgements Bekoff, M., 1977. Social communication in canids: Evidence for the evolution of a stereotyped I thank Dale Jamieson, Larry Shapiro, Eric mammalian display. Science, 197: 1097-1099. Saidel, David Rosenthal, Kim Sterelny, Ruth Bekoff, M., 1992. Description and explanation: A

35 Bekoff

plea forplurality. Behav.BrainSci., 15:269-270. cognitive ethology. Behav.BrainSci., In press. Bekoff, M., 1993a. Cognitive ethology and the Bennett, J., 1993. Comments on Dennett from a explanation of nonhuman animal behavior. In: cautious ally. Behav.BrainSci.,16:381- 385. ComparativeApproaches to Cognitive Science. Bogdan, R.J. (Ed.), 1986. Belief:Form, content awl (J.-A. Meyer & H. Roitblat, Eds.), Cambridge: function. New York: OxfordUniversity Press. MIT Press. Brand, M. & R.M. Harnish (Eds.), 1986. The Bekoff, M., 1993b. Review of Griffin 1992. representationof know ledge and belief Tucson: Ethology, In press. University of ArizonaPress. Bekoff, M. 1994. Cognitive ethology and the Bullock, T.H., 1993. How aremore complex brains treatment of nonhuman animals: How matters of different?Brain Behav. Evol., 41:88-96. mind informmatters of welfare. Submitted. Burghardt, G.M., 1991. Cognitive ethology and Bekoff, M. & Allen, C., 1992. Intentional icons: critical anthropomorphism: a snake with two towards an evolutionary cognitive ethology. heads and hognose snakes that play dead. In: Ethology, 91:1-16. Cognitive ethology: The minds of other Bekoff,M. & Allen, C., 1994 Cognitive ethology: animals. Essays in honor ofDonald R. Griffin: Slayers, skeptics, and proponents. In: 53-90 (C.A. Ristau, Ed.). Hillsdale, New Jersey: Anthropomorphism, Anecdote, andAnimals:The LawrenceErlbaum. Emperor's New Clothes?(R.W. Mitchell, N. Byrne, R. & Whiten, A. (Eds.), 1988. Thompson & L. Miles, Eds.). University of Machiavellian intelligence: Social expertise m1 Nebraska Press: Lincoln the evolution of intellect in monkeys, apes, m1 Bekoff, M. & Byers, J.A., 1981. A critical humans. New York: OxfordUniversity Press. reanalysis of the ontogeny of mammalian social Caro, T.M. & Hauser, M.D., 1992. Is there and locomotor play: An ethological hornet's teaching in nonhuman animals? Q. Rev. nest. In: BehavioralDeve lopment: TheBielefeld Biol.,61: 151-174. Interdisciplinary Project: 296-337 (K. Cheney, D.L. & Seyfarth, R.M., 1990. How lmmelmann, G.W. Barlow,L. Petrinovich & M. monkeys see the world: Inside the mind of Main, Eds.). New York: Cambridge University another species. Chicago: University of Chicago Press. Press. Bekoff, M. & Jamieson, D., 1990a. (Eds.) Cheney, D.L. & Seyfarth, R.M., 1992. Precis of Interpretationand explanation in the study of How monkeys see the world: Inside the mind of animal behavior: Vol. 1, Interpretation, another species. Behav.BrainSci., 15:135-182. intentionality, and communication. Boulder, Christensen, S.M. & Turner, D.R. (Eds.), 1993. Colorado:Westview Press. Folk psychology and the philosophyof mind. Bekoff, M. & Jamieson, D., 1990b. (Eds.) Hillsdale, New Jersey: Lawrence Erlbaum. Interpretationand explanation in the study of Clark, A., 1993. Mice, shrews, and animal behavior: Vol. 2, Explanation, misrepresentation. J. Phil., 90:290-310. evolution, and adaptation. Boulder, Colorado: Cling, A.D., 1991. The empirical virtues of beliefs. Westview Press. Phil. Psycho[., 4:303-323. Bekoff, M. & Jamieson, D., 1991. Reflective Dennett, D.C., 1983. Intentional systems in ethology, applied philosophy, and the moral cognitive ethology: The "Panglossian paradigm" status of animals. Persp.Ethol., 9:1-47. defended.Behav. Brain Sci., 6:343-390. Bekoff,M., Townsend, S.E. & Jamieson, D., 1993. Dennett, D.C., 1987. Reflections: Interpreting Beyond monkey minds: Towards a richer monkeys, theorists, and genes. In: The

36 Etolog{a, Vol. 3, 1993

intentional stance: pp. 269-286. Cambridge: Heyes, C.M., 1987. Cognisance of consciousness MIT Press. in the study of animal knowledge. In: Dennett, D. C., 1988. Out of the armchairand into Evolutionary Epistemology: 105-135 (Callebaut the field.Poetics Today, 9:205-221. W & Pinxten R, Eds). D. Reidel: Boston. Dennett, D.C., 1991. Consciousness explained. Heyes, C.M., 1993a. Reflections on self Boston: Little, Brown. recognition in primates. Anim. Behav., In press. Dresser, R., 1993. Culpability and other minds. Heyes, C.M., 1993b. Imitation, culture, and Southern CaliforniaLaw J., 2:41-88. cognition. Anim. Behav., In press. Dusenbery, D.B., 1992. Sensory ecology: How Heyes, C.M., 1993c. Anecdotes, training, trapping organisms acquire and respond to information. and triangulating. Do animals attribute mental San Francisco: W. H. Freeman. states? Anim. Behav., 46:177-188. Elgar, M.A., 1989. Predator vigilance and group Jamieson, D. & Bekoff, M., 1992a.Some problems size in mammals and birds: A critical review of and prospects for cognitive ethology. Between the empirical evidence. Biol. Reviews,64:13-33. The Species, 8:80-82. Elgar, M.A., Burren, P.J. & Posen, M., 1984. Jamieson, D. & Bekoff,M., 1992b. Carruthers on Vigilance and perception of flocksize in foraging nonconscious experience., 52:23-28. house sparrows Passerdomesticus L. Behaviour, Jamieson, D. & Bekoff, M., 1993. On aims and 90:215-223. methods of cognitive ethology. Philosophy of Fagen, R., 1981. Animal Play Behavior. New Science Association, 2: In press. York: Oxford University Press. Lahav, R., 1993. What neuropsychology tells us Godfrey-Smith, P., 1991. Signal, decision, action. about consciousness. Phil. Sci., 60:67-85. J. Phil., 88:709-722. Lazarus, J., 1990. Looking for trouble. New Golani, I., 1992. A mobility gradient in the Scientist, 125:62-65. organization of vertebrate movement: The Lima, S.L., 1990. The influenceof models on the perception of movement through symbolic interpretation of vigilance. In: lnterpretationard language. Behav.Brain Sci., 15:249-308. Explanation in the Study of Animal Behavior: Goldman, A., 1993. PhilosophicalApplications of Vol. II, Explanation, Evolution, andAdaptation: Cognitive Science. Boulder, Colorado: Westview 246-267 (M. Bekoff & D. Jamieson, Eds.). Press. Boulder, Colorado: Westview Press. Gordon, D.M.; Paul, R.E. & Thorpe, K., 19 What Lima, S.L. & Dill, L.M., 1990. Behavioral is the function of encounter patterns in ant decisions made under the risk of predation: A colonies? Anim. Behav., 45:1083-1100. review and prospectus., 68:619-640. Griffin, D.R., 1976. The question of animal Lorenz, K.Z., 1981. New York: Springer-Verlag. awareness: Evolutionary continuity of mental Manning, A. & Dawkins, M.S., 1992. Animal experience. New York: The Rockefeller behaviour. New York: Cambridge University University Press. Press. Griffin, D.R., 1984. Animal thinking. Cambridge: McGrew, W.C., 1992. Chimpanzee material HarvardUniversity. Press. culture: Implications for human Evolution. New Griffin, D.R., 1992. Animal minds. Chicago: York: Cambridge University Press. University of Chicago Press. Metcalfe,N.B., 1984a. The effectsof habitat on the Hauser, M.D., 1993. Right hemisphere dominance vigilance of shorebirds: Is visibility important? for the production of facial expression in Anim. Behav., 32:981-985. monkeys. Science, 261:475-477. Metcalfe, N.B., 1984b. The effectsof mixed-species

37 Bekoff

flocking on the vigilance of shorebirds: Who oo Jamieson, Eds.). Boulder, Colorado: Westview they trust? Anim. Behav., 32:986-993. Press. Millikan, R.G., 1984. Language, thought, and other Saide!, E., 1992. What price neurophilosophy? biological categories. Cambridge: MITPress. Philosophy of Science Association, 1 :461-468. Millikan, R.G., 1993. White queen psychology aid Snowdon, C.T., 1991. Review of Ristau 1991. other essays forAlice. Cambridge: MIT Press. Science, 251:813-814. Mitchell, R.W., 1990: A theory of play. In: Stein, B.E. & Alexander,M.A., 1993. Themerging Interpretation and Explanation in the Study of ofthe senses. Cambridge: MITPress. Animal Behavior: Vol. I, Interpretation, Sterelny, K., 1990. The representational theory of Intentionality, and Communication: 197-227 mind: An introduction. Cambridge: Basil (M. Bekoff& D. Jamieson, Eds.). Boulder, Blackwell. Colorado: Westview Press. Sullivan, K.A., 1984. Informationexploitation by Perner, J., 1991. Understandingthe representational downy woodpeckers in mixed-species flocks. mind. Cambridge: MIT Press. Behaviour, 91:294-311. Platt, M.M., Thompson, R.L. & Boatright, S.L., Swartz, K.B. & Evans, S., 1991. Not all 1991. Monkeys and mirrors: Questions of chimpanzees (Pan trogodytes) show self methodology. In: The monkeys of Arashiyama recognition. Primates, 32:483-496. (L.M. Fedigan & P.J. Asquith, Eds.). Albany, Tinbergen, N., 1951. The study ofinstinct. Oxford: New York: SUNY Press. OxfordUniv. Press. Premack, D., 1988. "Does the chimpanzee have a Tinbergen, N., 1963. On aims and methods of theory of mind?" revisited. In: Machiavellian ethology. Z. Tierpsychol., 20:410-429. Intelligence: Social Expertise and the Evolution Valone, T.J., 1989. Group foraging, public of Intellect in Monkeys, Apes, and Humans: information, and patch estimation. Oikos, 160-179 (R. Byrne& A. Whiten, Eds.). New 56:357-363. York: Oxford University Press. Valone, T.J., 1993. Patch information and Quenette, P .-Y ., 1990. Functions of vigilance estimation: A cost of group living. Oikos, In behaviour in mammals: A review. Acta press. Oecologica , 11:801-818. Waal, de F.B., 1991. Complementary methods and Ristau, C. (Ed.), 1991a. Cognitive ethology: The convergent evidence in the study of primate minds of other animals. essays in honor of social cognition. Behaviour, 118:297-320. Donald R. Griffin. Hillsdale, New Jersey: Wasserman, E.A., (1993). Comparative cognition: Lawrence Erlbaum. Beginning the second century of the study of Ristau, C., 1991b. Aspects of the cognitive animal intelligence. Psycho/. Bull., 113:211- ethology of an injury-feigning bird, the piping 228. plovers. In: Cognitive ethology: The minds of Wemelsfelder,F., 1993. Animal boredom: Towards other animals. Essays in honor of Donald R. an empirical approach of animal subjectivity. Griffin: 91-126 (C.A. Ristau, Ed.). Hillsdale, DoctoralThesis, Rijksuniversiteit te Leiden. New Jersey: Lawrence Erlbaum. Whiten, A. & Ham, R., 1992. On the nature and Rosenberg, A., 1990. Is there an evolutionary evolution of imitation in the animal kingdom: biology of play? In: Interpretation a,d Reappraisal of a century of research.Adv.Study explanation in the study of animal behavior: Behav., 21:239-283 Vol. I, Interpretation, intentionality, aid Williams, G.C., l 992. NaturalSelection: Domains, communication: 180-196 (M. Bekoff & D. levels, and challenges. New York: Oxford

38 Etolog(a, Vol. 3, 1993

University Press. Yoerg, S. I. 1991. Ecological frames of mind: the Wyers, E.J., 1985. Cognitive-behavior in role of cognition in behavioral ecology. Q. sticklebacks. Behaviour,95: 1-10. Rev.Biol., 66:287-301.

(Recibido: agosto 1993)