ebook img

The Cognitive Face of Avian Life Histories. The 2003 Margaret Morse Nice Lecture PDF

2004·11 MB·English
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview The Cognitive Face of Avian Life Histories. The 2003 Margaret Morse Nice Lecture

THE WILSON BULLETIN A QUARTERLY JOURNAL OF ORNITHOLOGY Published by the Wilson Ornithological Society VOL. 116, NO. 2 June 2004 PAGES 119-196 Wilson Bulletin, 116(2), 2004, pp. 119-133 THE COGNITIVE FACE OF AVIAN LIFE HISTORIES The 2003 Margaret Morse Nice Lecture ROBERT E. RICKLEFS’ — ABSTRACT. Cognition includes the acquisition, processing, retention of, and acting upon information from the environment. Avian cognition has been investigated by the approaches of experimental psychology and in the context of specific tasks, such as spatial memory. However, the costs and benefits ofcognitive ability have not been considered in a life-history context. I explore possible relationships between behaviors that might indicate cognitive function and other attributes, particularly brain size, rate ofdevelopment, age at maturity, and life span. Large brain size and prolonged development are seen as potential costs of intelligent behavior. Long life span may permit the extended learning periods that support experienced-based cognitive function. Play behavior, which plausibly supports the development ofmotor and social skills, and, to a lesser extent, foraging innovations, are related to brain size. The challenge of foraging in a spatially and temporally varying environ- ment, experienced for example by pelagic seabirds, is associated with prolonged embryonic development. Al- though these connections lack mechanistic foundations, they suggest that cognition can be considered as a part of the life history of the individual and that potential costs ofcognition might provide guidelines for directing the comparative study of intelligent behavior. Received30April 2004, accepted 7June 2004. Over the course of her remarkable career, relationship with its environment. The envi- Margaret Morse Nice produced many original ronment of every organism is varied in space contributions to ornithology, including her pi- and changes constantly through time, often oneering studies on the life histories of birds unpredictably, but also with regularities that (e.g., Nice 1937, 1943, 1957) and the devel- can be learned over time. Individuals respond opment of behavior (Nice 1962). For many of to their environments in a variety of ways. us who were students during the 1960s, which Many behaviors are “hard-wired” into the was a period oftransformation in ecology and nervous system and express evolved respons- behavior, her work laid the foundation for all es, available from birth, to consistent features that we set out to accomplish. It is this inter- of an individual’s surroundings. At the other section of life history and behavior that I extreme, organisms occasionally encounter would like to address in this contribution. I novel situations for which they must devise am grateful to the Wilson Ornithological So- novel solutions, what many of us would con- ciety, and particularly to Jed Burtt, for giving sider intelligent behavior. We suspect that spe- me the opportunity (and a captive audience) cies vary widely in terms of what we think of for exploring ideas that have fascinated me for as intelligence; most ornithologists, if asked, many years, but which have not come into would put parrots and corvids at the top ofthe clear focus until now. a\ian intelligence scale and relegate doves Behavior provides the individual a flexible and sparrows to lower pt)sitions. although one could argue that individuals of every species ' Dept, of Biology. Univ. of Missouri-St. Louis. .St. handle very well the tasks necessary for their Loui.s. MC) 63121. USA; c-mail; ricklcts^Himsl.cclu survival and reproduction. 120 THE WILSON BULLETIN • Vol. IJ6, No. 2, June 2004 I assume that most behavior is adaptive and life-history context. Clearly, this is a crude be- contributes to the lifetime reproductive suc- ginning to a highly complex and difficult task. cess, or fitness, of the individual. Variation in I hope, however, that this essay will encourage capacities for certain types of behavior pre- readers to regard intelligent behavior as inte- sumably reflects different balances between gral to the life history of the individual and the costs and benefits of such behavior in the subject to selection that weighs its costs and context ofdemands of the environment. I also benefits. alesasstumien ttheartmstheofcatphaecidteyvetolotphmiennktisancodstmlya,ina-t WHAT IS COGNITION? tenance of the hardware required, and that if According to Shettleworth (2001), cogni- it is not necessary to think in a particular way, tion, broadly defined, “includes perception, individuals should not bear the cost of that learning, memory and decision making, in particular ability. This simple idea constitutes short all ways in which animals take in infor- a life-history approach, which considers the mation about the world through the senses, conflicting contributions of adaptations to fit- process, retain and decide to act on it.” Vau- ness resulting from the allocation of limited clair (1996:10) sees cognition more narrowly time, energy, tissue, and otherresources. I find as allowing an individual “to adapt to unpre- it remarkable that the evolution of intelligent dictable changing conditions in its environ- behavior is rarely considered within a life-his- ment. Thus, behaviors that would aid adapta- tory framework. Indeed, one of the most tion would reflect several characteristics, such widely respected textbooks in animal behavior as flexibility, novelty, and generalization. that uses an evolutionary approach (Alcock Flexibility of behavior designates the possi- 1998) does not list cognition in its index and bility of constructing an adapted response to under intelligence refers only to a single page unusual external conditions. The response also devoted to the heritability of IQ. must be novel in the sense that it does not In this essay, I have endeavored to treat express the existence of a pre-wired program. thinking, or the capacity to think, in a life- Finally, the novel behavior, established to history context. The fact that some birds ap- solve a novel problem, must be susceptible to pear to do less “thinking” than others sug- generalization to situations that differ partially gests that thought might have substantial costs or totally from those in which they were ini- or that it might be highly constrained by other tially acquired.” aspects of the life history. Of course, thought Many psychologists distinguish “between is hard to measure. It is possible, however, cognition, a possible means to an end, and in- that some of the costs it incurs are not. These telligence, an assessment of performance costs might involve maintaining a large brain judged by some functional criteria” (Mc- and prolonging development to build a large, Farland 1989:130). Intelligent behaviors are complex nervous system and acquire the ex- often regarded as specific adaptations to spe- perience necessary to perceive pattern in en- cific problems (Rozin 1976). For example, the vironmental variation. If this inference were sophisticated navigation abilities of pigeons correct, could we not use these presumed appear to be highly intelligent (Wiltschko and costs to indicate certain mental capacities? I Wiltschko 1993, 1998; Walcott 1996; Wallraff shall use data on brain mass and the length of 2001), but because they are mostly controlled the incubation period for a large number of by hard-wired systems, these abilities cannot birds to construct a two-dimensional space be generalized to other kinds of behavior. representing some presumed costs of cogni- Thus, by Vauclair’s definition, they cannot be tion. I will then determine whether certain considered as an indication of general cogni- kinds ofbehavior that are associated with cog- tive ability. It is entirely possible that intelli- nitive ability (sociality, appearance of novel gence and cognition defined in this manner behaviors, play, experience-based foraging, represent points on a continuum and that the for example) bear a relationship to presumed distinction, although perhaps heuristically use- costs (some do!). Of course, patterns do not ful, is artificial. Cognitive abilities themselves tell the whole story. However, we can use pat- are certainly specialized in many respects. For terns to guide our thinking about thought in a example, Clark’s Nutcrackers (Nucifraga col- Ricklefs • MARGARET MORSE NICE LECTURE 121 Limbiana) have a high degree of spatial cog- Chappell and Kacelnik 2002, Hunt and Gray nition across many types oftasks, but perform 2002, Weir et al. 2002), have received consid- less well with non-spatial information, such as erable and well-deserved attention in the me- color (Olson et al. 1995). dia. A third approach is to infercognitive abil- I shall adopt a broader dehnition of cogni- ities from the kinds of problems that animals tive ability here to include the acquisition and have to solve in their daily lives. That is, we processing of information from the environ- can ask what kinds of mental function are ment as a basis for behavior. I hasten to add needed for an individual to behave the way it that the study of animal cognition is not the does in its particular environment. For exam- study of animal consciousness, although con- ple, tracking nectar sources by hummingbirds sciousness (self-awareness) should be consid- (Garrison and Gass 1999, Bateson et al. 2003) ered as a type of cognitive behavior (Rozin and cache retrieval byjays (Baida et al. 1996; 1976; Griffin 1984, 1985; Terrace 1984; Roit- Griffiths et al. 1999; Clayton et al. 2001, blatt 1987; Bekoff 2000). 2003; Baida and Kamil 2002) may require Many researchers regard as central to cog- more sophisticated spatial and temporal rep- nition the concept of a cognitive map (Roit- resentations than leaf gleaning by warblers blatt 1982, 1987; Gallistel 1990; Gervet et al. and ground foraging by doves. 1996). Animals acquire and remember land- Representation, which might be thought of marks, attach values to them, and incorporate as the formation ofmental images, can be ver- changes in these landmarks over time. Cog- ified in its simplest form by various tests of nitive maps may represent physical space (Ka- memory (e.g., Griffiths et al. 1999, Clayton et mil and Jones 1997, Gibson and Kamil 2001), al. 2001). Beyond this, experimental compar- or they may relate to social “space,” predator ative psychology investigates problem-solving “space,” weather “space,” and the like. A abilities associated with concepts ofsets, iden- cognitive map is a representation of features tity and oddity, perceptual categories, serial of the world in the brain itself, and neurobi- learning, and imagery (Vauclair 1996). In an ologists are beginning to discover the corre- experimental setting, identifying the formation spondence between the two (e.g., Jarvis and of sets, for example, involves a subject’s abil- Mello 2000, Bingman and Able 2002). ity to generalize the concept of similarity to HOW DO WE ASSESS COGNITIVE novel objects; serial learning is revealed by the ability to construct a correct sequence out ABILITIES? of a subset of sequentially presented stimuli; Corvids and parrots are generally regarded and so on. Many excellent reviews detail the as intelligent and sociable (Takahashi and Kel- results of experimental psychology (Pearce ler 1994; Hunt 1996; Marler 1996; Cohen et 1987; Gallistel 1989; Ristau 1991; Byrne al. 1998; Pepperberg 1999, 2002; Bugnyar et 1994; Baida et al. 1996, 1998; Vauclair 1996; al. 2001; Baida and Kamil 2002; Bugnyar and Shettleworth 1998; Griffiths et al. 1999; Pep- Kotrschal 2002; Heinrich 2002; Hunt et al. perberg 1999; Heyes and Huber 2()()(); Clay- 2002; Knudsen 2002). What does this mean? ton et al. 2001; Wynne 2001; Bekoff et al. We have various tools for probing the cogni- 2002). tive abilities of animals. The most important In a comparative, life-history framework, of these comprise the methods of experimen- the approaches of experimental psychology tal psychology, in which mental capacities are become laborious and context dependent. For assessed by performance in various kinds of example, it is difficult to compare the results behavior tests. A second approach, which may of psychological tests on species with \aricd have broader application in comparative stud- communication modalities and different prt)b- ies, is to record behavioral correlates of cog- lem solving requirements in their lives, riicrc- nitive ability. Recently, the frequency of be- fore, the analyses in this discussion rely on havioral innovations has received attention in various presumed behavioral correlates of this regard (Lefebvre et al. 1997, 1998; Ni- cognitive function and inferences concerning colakakis and Ixfebvrc 2()()0) and particular cognitive function from the bcha\ioral de- cases, such as tool making in New C’aledonian mands posed by tasks that organisms accom- Crows (Corvus nu)ncduloiclcs\ Hunt 1996, plish in their tlaily lives. 122 THE WILSON BULLETIN • Vol. 116, No. 2, June 2004 COGNITION AS A LIFE-HISTORY TRAIT in the brain may be prolonged development. cogTnhietiobnenmefuistts ionfvoclovgeniftuinocnti.—onTihngeinvaaluceomo-f Tofheitsbagsriocwtahrchoictceucrtsurbeefoofrethbeirbtrhaiinn maandmmmaolsst plex, variable environment where decisions (Pagel and Harvey 1988), and this is also true in birds (Portmann and Stingelin 1961; Starck conditioned by experience or reasoning are 1993, 1998; Ricklefs and Starck 1998). Birds crucial. As explained by W. J. Smith (1990), cognitive function enables prediction and exhibit tremendous va—riation in the length of the incubation period more than a factor of shapes expectation about the environment, in- three when incubation period is normalized by cluding social interactions (see also Stephens egg size (Ricklefs 1993). Costs of prolonged 1989). Examples include food caching (Krebs incubation include increased exposure to time- et al. 1996, Gibson and Kamil 2001, Kamil dependent mortality, higher energy require- and Cheng 2001), risk assessment associated ments ofembryonic growth, and increased re- with variable rewards (Real 1991, 1993; Ka- productive stress for parents. The benefits of celnik and Bateson 1996; Bateson and Kacel- long incubation periods have not been iden- nik 1997; Garrison and Gass 1999; Marsh and tified, although presumably they reflect a Kacelnik 2002; Schlick-Paim and Kacelnik higher quality chick (Ricklefs 1992, 1993). 2002), context-dependent responses to social A more cognitive life style might also be signaling (W. J. Smith 1990), social negotia- associated with a long learning period to ac- tion based on shared information (Smith 1997, quire information, leading to delayed repro- 1998), and reciprocal altruism depending on duction and reduced lifetime reproductive suc- long-term association with identifiable indi- cess. For example, increased cognitive ability viduals (Trivers 1971, Axelrod and Hamilton might carry with it the opportunity to exploit 1981). — a type of food supply that is not available, or The costs of cognitive ability. The asso- is less efficiently used, without acquiring ex- ciation of well-developed cognitive abilities tensive information about temporal and spatial with a large brain may be a particularly hu- distribution of prey or of suitable foraging man conceit, but many studies of differences conditions. Complex social settings may re- in behavior among species are generally con- quire learning appropriate responses to indi- sistent with such a relationship (Jerison 1973, vidual variations in behavior and signaling. Clutton-Brock and Harvey 1980, Macphail 1982, Dunbar 1992) and certain measured CAN WE RECOGNIZE COGNITIVE cognitive abilities are related to the size of ABILITIES BY THEIR ASSOCIATED relevant parts of the brain. I am referring in COSTS? particular to the relationship between spatial I believe that the most obvious candidates memory and the size of the hippocampus for costs are brain size and embryonic devel- (Healy and Krebs 1996, Biegler et al. 2001), opment period. Postnatal growth rate and age but similar structure-function connections are at maturity are also reasonable choices in that evident with respect to other behaviors, such they refleet a substantial portion of brain as singing, and related brain regions (Bren- growth and development and the most signif- owitz et al. 1985, Brenowitz and Arnold 1986, icant part of an individual’s learning period. Devoogd et al. 1993, Bernard et al. 1996), al- In fact, these measures are generally correlat- though one must exercise caution in general- ed with embryonic development. izing such connections (Aboitiz 1996, 2001). The relationship between brain mass and The brain is thought to be an expensive or- body mass in birds is shown in Figure 1. The gan metabolically (Field et al. 1939, Martin residuals from the logarithmic regression pro- and Fuhrman 1955, Aiello and Wheeler 1995) vide a body mass-specific index to relative and large brain size presumably also applies brain size. Averaged over larger taxonomic architectural stresses on morphology and as- groups of birds (Sibley and Monroe 1990), pects of animal function (Aboitiz 1996). these residuals identify several families with Flight itself may impose some limits on brain larger-than-average brain mass: Psittacidae, size. Another cost of a larger brain or more Strigidae, Corvidae, Picidae, Bucerotidae. It is precise and complex neural connections with- not surprising to find the parrots and corvids Ricklefs • MARGARET MORSE NICE LECTURE 123 (/) C/5 05 E c '03 CD Body mass (g) EIG. 1. Relationship between brain mass and body mass in birds based on data for 837 species in 104 sfalmoipleie=s (0f.o6r25det±ail0s.,0s0e8e,Nientaelrecnepatn=d R-i0c.k9le9f7s;2t0h0u1s),.thReegerqeusastiioonnwiisthloign obrradienrs=(n-=0.2939)7E,+yg0o.6=2565X90,loPg b<od0y..001, = 0.952, RMSE (within orders) = 0.114. Order effect; F22JS0 ~ 48.8, P < 0.001. Residuals are calculated from this regression line. Among families represented by more than 10 individuals, those with the highest residual brain masses were Psittacidae (0.27 ± 0.10 SD, n = 49), Strigidae (0.25 ± 0.17, n = 24), Corvidae (0.24 ± 0.10, n = 21), Picidae (0.22 ± 0.11, n = 13), and Bucerotidae (0.15 ± 0.07, n = 13). Those with the lowest 0r.e0s9i,duanls=w6e4r)e, PShcaoslioapnaicdiadeae(-(0-.03.012±±0.01.10,5,nn==392)7,),CoalnudmTbriodcaheili(d—a0e.2(2-0±.120.1±0,0.n06=, n21)=, 2A7n)a.tidae (-0.14 ± near the top of the list, and hornbills are also sion provide an index to relative incubation known for complex social behavior (Kemp period. In this case, taxa with exceptionally 1995), which might be indicative of well-de- long incubation periods include Procellariifor- veloped cognitive abilities. Owls have very mes, Accipitridae, Spheniscidae, Psittacidae, sophisticated foraging methods involving Strigidae, Bucerotidae, and Falconidae. Three complex processing of auditory information ofthese families also have exceptionally large (Takahashi and Keller 1994, Cohen et al. relative brain size. In addition, the group in- 1998, Knudsen 2002). I must have underesti- cludes two lineages of seabirds and most rap- mated woodpeckers in the past, although torial birds. On the low end of the scale are Acorn Woodpeckers (Melcinerpes formicivo- the Picidae, Columbidae, Fringillidae, Mus- rus) certainly exhibit complex social behavior cicapidae, and Passeridae, all of which have (e.g., Koenig and Mumme 1987, Koenig et al. extremely altricial development and tend to 1998). Families on the bottom of the list of have small body sizes. relative brain size are Phasianidae, Columbi- Residuals from the brain mass and incuba- dae, Anatidae, Scolopacidae, and Trochilidae. tion period regressions define a space within It is notable that three of these groups have which any particular group of birds can be precocial offspring whose brains are well de- placed. Within this space, the presumed cost veloped at hatching (Ricklefs and Starck of cognition increases with larger rclati\c 1998). It may also be signihcant that four of brain size and longer relati\e embryonic de- the.se groups constitute the bulk of bird spe- velopment, and this is where (uie expects to cies that have been hunted commercially ;ind find birds with the most de\elopeil cognitise for sport. abilities. Because we ha\e little tlirect com- The relationship between incubation period parative information on the cogniti\e capaci- and egg mass in birds is shown in Figure 2. ties of birtls. with the exeeption of pigeons, Again, residuals from the logarithmic regres- parrots, and cor\ids. I shall eonsider several 124 rHK Wll.SON liULLHl'IN • Vol. 116, No. 2, June 2004 0.1 1 10 100 1000 10000 Egg mass (g) FIG. 2. Relationship between incubation period and egg mass in birds based on 799 species in 19 orders (lor details, see Ricklel's 1993). Regression within orders {n = 19) f’l7V6 = 203, P < 0.001, slope = 0.160 ± 0.008, intercept = 1.099 ± 0.073; thus, the equation is log incubation period = 1.099 + 0.160 X log egg mass, IP = 0.833, RMSE (within orders) = 0.073. Order elYeet: E18.776 = 35.1, P < O.OOl. Residuals are calculated from this regression line. Among taxa with exceptionally long relative incubation periods are Procellariiformes (0.323), Sulidae (0.239), Accipitridae (0.174), Trochilidae (0.172), Apodidae (0.156), Spheniscidae (0.153), Psit- tacidae (0.146), Strigidae (0.137), Bucerotidae (0.135), and Falconidae (0.135). Those with exceptionally short incubation periods include Sturnidae (-0.1 19), Picidae (-0.088), Columbidae (-0.067), Fringillidae (-0.053), Muscicapidac (—0.048), and Passeridae (—0.031). behavior indices that might plausibly be re- cognition. In general, however, the proportion lated to cognition: (1) cooperative breeding; of species with helping behavior is unrelated (2) sociality, or group living; (3) play behav- to either relative brain mass or relative incu- ior; (4) foraging innovations; and (5) chal- bation period. In a stepwise multiple regres- lenging foraging situations. I shall also com- sion of the proportion helping (SAS PROC pare the putative costs to life span (which is GLM), neither relative brain mass (F, = 20 also closely related to age at maturity) to as- 0.46, P = 0.50) nor relative incubation period sess the idea that some types of cognitive be- (F| 2o = 2.96, P == 0.10) were signiheant ef- havior require extensive l—earning periods. fects (total = 0.142). In retrospect, helping Cooperative breeding. My criterion for behavior is probably not a good cognition in- cooperative breeding for each family or other dex because, in essence, helping merely com- large taxonomic group is the proportion of bines failure to disperse with what all birds species exhibiting helping behavior, reported by Jerram Brown (1987:table 3.1). The ratio- with altricial development do naturally, that is, feed offspring. dnaelxetfoorcougsniintgiocnooispetrhaattiivnembarneyedifnagmilasy aorn eixn-- A more pertinent index might be the capac- tended family groups, individuals discriminate ity to develop complex interactions within and the recipients of helping behavior on the basis among extended family groups in species with of relationship, which requires the learning of social breeding, where reproductive success kin relationships (Emlen et al. 1995). The tax- may hinge on personal knowledge of, and onomic groups with the highest proportion of long-term association with, other individuals helping, according to Brown’s summary, are in the group. For example, among the species the hornbills, other coraciiforms, grebes, cor- of cooperatively breeding birds detailed in vids, and mousebirds. Hornbills and corvids Stacey and Koenig (1990), those engaging in also get high scores for the putative costs of colonial breeding belonged to coraciiforms Ricklefs • MARGARET MORSE NICE LECTURE 125 25 Corvidae# Corvidae• 20 •Accipitridae Acclpltridae' o c o I0055S-r^5o: 15 Psittacidae# ^0-CQD 1o- C>D- 10 • •• O1- >CD, E2 • «• • E 2 z • Columbifo•rm•es•#•#•^ ^ ••Piciformes Z ••••P•IcIfor•mesColumbMifor•mes -0.4 -0,2 0.0 0.2 1.0 1,2 1.4 1,6 Brainsizeresidual(log^ounits) Log^olifespan(years) EIG. 3. Relationship between number ofpublications describing play behavior and relative brain size (left) and life span (right). Erom data compiled by Eagen (1981). and corvids (J. N. M. Smith 1990:table 2), assign different scores, but the result probably taxa with relatively large brains. would not change. Again, the problem with — Sociality. Group living balances benefits group living as an index to cognition is that of group defense and social foraging against many associated behaviors may require little the costs of local competition for resources more cognitive capacity than the kinds of co- and social strife. Social behavior is thought by operative and antagonistic interactions that all many authors to go hand in hand with cog- birds engage in, whether social or not. A bet- nitive behavior and brain size in primates (Sa- ter understanding of more complex behavior waguchi 1990, 1992; Dunbar 1992, 1993). in social species based on individual knowl- The “social complexity hypothesis” states edge and association might lead to a better that living in large groups selects forenhanced index (e.g., W. J. Smith 1990, 1998), but this cognitive abilities with respect to recognizing is beyon—d my understanding of bird societies. individuals and assessing social relationships Play. Play is a more promising indicator (Cheney and Seyfarth 1990, Byrne and Whit- ofcognitive abilities because play presumably en 1997, Kummer et al. 1997). Support for represents practice behaviors that refine phys- this hypothesis has recently come from ex- ical and social skills (Fagen 1981, Byers and perimental studies on cognitive abilities in Walker 1995, Bekoff and Byers 1998). Un- jays (Bond et al. 2003). My criterion forgroup fortunately, there is no widely accepted deli- living was the tendency to form groups with nition ofplay in birds and comparative studies complex social structure. Both parrots and ofavian play behavior are largely lacking. We corvids place high on the list of such taxa, but all know play when we see it, but many re- to keep things simple 1 subjectively assigned ports of play in the ornithological literature taxa a score of either 0 or 1. Among the less are similarly anecdotal. 1 have taken as my social taxa are ducks, doves, cuckoos, quail, index to play the number of publications de- oscine passerines, and raptors; among the scribing play behaviors listed in fables 3-26 more social taxa are parrots, corvids, many to 3—28 of Fagen ( 1981 ). The taxa that come seabirds, and many coraciiforms. Again, as in out on top of this list are falcons, passcrids. the case of helping behavior, group living was parrots, accipiters, and corvids, fhe sample is not a significant effect in an analysis of vari- undoubtedly biased by the large number of ance (ANOVA: SAS PROC GEM) for either studies on these groups; however, it is note- relative brain mass (/j 33 = 2.04, P = 0.16, /C worthy that each of these li\e taxa has above- 0=.905.,05PS)=or0.r3e4l,atiRv~e =inc0u.b0a3t1i)o.n Npoerrioddid(/a| 3d,,is=- awveerreagaenarleylzaetid\ebybraaisntespi/weis(efimgu.lt3i).plefhreegdraetsa- criminant analysis with body size, relative sion, with body mass, relative brain mass brain mass, and relative incubation period dis- (rbrain). relative incubation period, and life tinguish social versus non-social species (/'v^h span as iiuleiKMulent \ariables. In this analysis, = 0.69, = 0.57). Different observers would brain size (/ | .„ 7.0. P 0.014) and life 126 THE WILSON BULLETIN • Vol. !16, No. 2, June 2004 span {F^2b ^ 5.85, P - 0.023) explained 37% Mettke-Hofmann 2001, Mettke-Hofmann et of the variance in number of citations of play al. 2002). — behavior. Relative brain size by itself ex- The challenge offindingfood. Feeding is plained 23% of the variance (/^i2? = 8.1, P = a behavior common to all birds, however, dif- 0.008, play - 4.42 [±1.00] + iv.29 [±6.08] ferent prey present widely different behavioral rbrain). — challenges. I presume that the most challeng- Foraging innovations. Louis Lefebvre and ing types of prey resources are those that ex- his colleagues at McGill University have re- hibit extreme temporal and spatial heteroge- cently tabulated reports of foraging innova- neity, such as the prey of most pelagic sea- tions from the literature. Lefebvre et al. (1997) birds, or those that have well-developed abil- define foraging innovation as “either the in- ities to sense and evade predators, such as the gestion ofa new food type or the use ofa new prey of many raptorial birds (see also, Glut- foraging technique,” generally a behavior re- ton-Brock and Harvey 1980, Milton 1988, ported either for the first time or as being Dunbar 1992). Other birds handle special highly unusual for a given species. These in- challenges, such as those which feed by trap- clude such diverse behaviors as an American lining on changing arrays of flowers (hum- Kestrel (Falco sparverius) drowning a Red- mingbirds) or widely dispersed fruiting trees winged Blackbird (Agelaius phoeniceus), a (some tropical frugivores). I regarded species Hooded Merganser {Lophodytes cucullatus) that feed on fine-grained food resources, depredating an adult meadow vole (Microtus where success is proportional primarily to pennsylvanicus), and sparrows searching car searching time rather than special searching radiator grills for insects. Here, I use as an strategies (foliage gleaners, most seed-eaters, index the proportion of papers on a particular for example), as not requiring well-developed taxonomic group that describe foraging inno- cognitive abilities. I classified foraging as vations, compiled for North American and very challenging 2 raptors, swifts, seabirds), ( : Australian birds by Lefebvre et al. (2001). moderately challenging (1: corvids, parrots, Thus, this index is corrected for research ef- hummingbirds, plovers, flycatchers, wood- fort. Incidentally, more than half of the re- peckers, several tropical fruit-eating groups), ported innovations in North American birds and less challenging 0 most opportunistic ( : come from the pages of The Wilson Bulletin, ground feeders, waterfowl, most passerines). which remains one of the few ornithological Analyses of variance with foraging chal- journals that publishes natural history obser- lenge as the main effect were not significant vations. for body mass (F232 = 2.82, P — 0.075), but Lefebvre and his colleagues have shown were highly significant for relative brain size that the incidence of foraging innovations is (7^2,32 = 8.27, P = 0.001, = 0.34) and rel- positively related to brain size (Lefebvre et al. ative incubation period (F229 — 7.06, P = My = 1997, 1998, 2001). analysis also reveals 0.003, /?2 0.33) (Fig. 4). Relative incubation such a relationship, although it is weak (r = period increased from foraging class 0 (0.017 0.38, P < 0.05; arcsin-transformed: r = 0.46, ± 0.072 SD, n = 17) to class 1 (0.073 ± P = 0.016, n = 27). The taxa with the highest 0.084, ^ = 11) and class 2 (0.149 ± 0.071, n incidence of reported foraging innovations = 7); relative brain size increased from for- were the cranes and their relatives (Gruidae), aging class 0 (-0.153 ± 0.141, n = 17) to falcons (Falconidae), parrots (Psittacidae), class 1 (0.062 ± 0.129, n = 11), but was not hummingbirds (Trochilidae), rails and their significantly higher in class 2 (—0.023 ± relatives (Rallidae), and swifts (Apodidae). 0.152, n = 7). Because mode of development Multiple regression showed that incidence of is associated with both brain size (precocials foraging innovations is unrelated to body smaller) and incubation period (precocials mass, relative incubation period, mode of de- longer), I analyzed the data again only fortaxa velopment, and life span. How foraging in- with altricial development. Precocial species, novation might be related to inherent accep- except plovers, which actively pursue mobile tance or avoidance ofnew stimuli (neophobia) prey, were placed in challenge class 0. In this is an interesting, but unexplored problem second analysis without precocial species, rel- (e.g., Marples et al. 1998, Greenberg and ative brain size was no longer significant (F2jg Ricklefs • MARGARET MORSE NICE LECTURE 127 Incubation Low Foraging 0.3 Moderate period High residuals difficulty Low,precocial g03 Mod,precocial 0.2 Body Play o mass7 Life span' behavior g 0 0.1 Cl V c Brain mass^ Foraging o o o OO ro o o Aq residuals innovation 0.0 o c: FIG. 5. Connections between life-historyvariables and indicators ofcognitive function revealed by anal- -0.1 yses in this study. Dashed arrows indicate weak cor- relations. -0.3 -0.2 -0.1 0.0 0.1 0.2 Brain massresiduals EIG. 4. Groups ofbirds exhibiting different levels osqfufaorreasg:inhgigchh)alasleangfeun(ccitricolneso:florewl,attirviaengblreasi:nmmoadsesraatned, muImusleodngaesviatymefaorsuaretaoxfonloifmeicspgarnoutpherempaoxrit-- relative incubation period. Groups with altricial devel- ed in the compilation of Carey and Judge opment are shown as filled symbols, those with pre- (2000). Among the taxa sampled in this anal- cocial development as open symbols. ysis, life span is positively correlated with body mass (r = 0.43, P = 0.017, n - 30), = 1.96, P — 0.17), body size was marginally relative incubation period (r = 0.48, P = asingdnifrieclaanttiv(eF2ijgnc=ub3a.t7i,onP =per0.i0o4d6,expl=ain0e.3d0),a r0e.p0o0r9t,inng =pla2y9)b,ehaanvdionru(mrb=er0.o6f5,puPbl=ic0a.t0i0o6n,s larger proportion ofthe variance (^ ^ 12.7, n = 16), but not relative brain mass (r = 0.25, P < 0.001, = 0.60). Mean rela2,t1i8ve incu- P = 0.19, n = 30). Neither the proportion of bation periods for the foraging classes were species exhibiting cooperative breeding {r = n0(c.l0=a5s9s60±: 0-.00.840,42n ±= 09;.0c7l0asSsD2,: 0n.1=725;±cl0a.s0s421,: no-f0=.r2e52p1o6r,)tsPwao-sf is0in.gn2no7i,vfaintcia-ontnl2y(1r)re=nlao0tr.e2dt4h,teoPflrief=equs0ep.na2cn3y., ). Life span also did not differ among taxa in DEMOGRAPHY AND COGNITION different foraging challenge classes (F227 = Long life span is strongly associated with a 0.85, P = 0.44). Taxa with higher tendencies low reproductive rate, delayed maturity, and to form social groups had marginally longer slow rate of senescence (Ricklefs 1973, 2000; life spans than less social taxa (/-Y28 “ Ricklefs and Scheuerlein 2001). Long life = 0.042, H = 0.14). span may also promote the ability of an in- Connections between behavior and life-his- dividual to make use of well-developed cog- tory variables found in this analysis are dia- nitive capacities owing to long learning peri- grammed in Figure 5. Play beha\ior and for- ods, including apprenticeships with older in- aging innovations are pc^sitively related to dividuals, and extensive experience with tem- large brain mass and, in the case of play, h)iig poral and spatial variation in the environment. life. Challenges of foraging are iikuc closely Most species present evidence of improved re- associated with relatively long incubati(ui pe- productive success with experience and age riod, which itself is correlated v\ith life span, (CoLilson 1966, Ollason and Dunnct 1978, riius, several life-history traits might be as- Newton 1985, Perrins and McCleery 1985), sociated with well-developed cogniti\e abili- which may be attributed to the acquisition of ties: large brain si/e, com|')lcx brain structure, information rather than the development of and slow development as costs (or enabling physical skills. Thus, it is reasonable to ask adaptations): high parental investment, de- whether species with long life spans tlemon- layed maturity, and long life span as assoei- stratc higher cognitive skills or costs associ- atetl traits; and complex foraging aiul social ated with cognition. behavit>rs as benefits. 128 THE WILSON BULLETIN • Vol. 116, No. 2, June 2004 DIFFERENT WAYS OE BEING Do certain preconditionsfacilitate the evo- — INTELLIGENT lution ofenhancedcognitive abilities? Ifcer- tain types ofthinking are associated with such The association of challenging foraging with embryonic development and the associ- life-history traits as large body size, long life span, and prolonged development, the evolu- ation of play and foraging innovation with tion of these traits for other reasons might brain size raises the question of whether there are different ways ofbeing intelligent, each of facilitate the evolution of cognitive ability. Distinguishing preconditions from correlated these relationships representing different com- ponents of cognitive ability. Many authors evolution oftraits requires analysis ofthe dis- have made the distinction between special and tribution of traits on a well-supported phylo- genetic hypothesis. Information concerning general intelligence, the difference essentially this issue might also be obtained from the la- between experience-based decision-making bility of traits within a phylogeny. Conserva- and reasoning (Rozin 1976, McEarland 1989, tive traits are more likely to have preceded the Vauclair 1996). The first might be thought of evolution of more labile traits owing to their as being retrospective, building on the accu- longer histories. However, disparity in the la- mulation of information about the environ- bility oftraits also would signal an uncoupling ment and processing it in ways to make pre- oftheirevolution. Both relative incubation pe- dictions based on past experience. The ability riod and relative brain size are conservative, of seabirds to find their way over thousands with most of their variance occurring on the of kilometers of ocean and locate quality for- level of families within orders or even higher aging areas (Jouventin and Weimerskirch (Fig. 6). Because we do not have adequate W1i9l90s,onPr2i0n0c0e)etmaayl. r1e9q9u2i,reWtehiemearcsckuimruclhatainodn measures of cognitive abilities, it is difficult to determine whether these have a comparable of experience with correlations between distribution of variance or are more labile. weather, oceanographic conditions, and for- Where groups of birds have been looked at aging success. If this were true, waiting up to closely with regard to behavior, researchers 10 years to achieve sexual maturity (see Rick- have tended to emphasize differences between lefs 1973, 2000) might represent a learning closely related species rather than their simi- period necessary before an individual can feed even a single chick successfully. Why such a tlawreiteineslaorr,geralttearxnoatniovmeilyc, gtrhoeupdsif(feDreevnocoesgdbee-t capacity for learning might be related to the al. 1993, Healy and Krebs 1996, Baida et al. length of the embryonic development period 1997, Baida and Kamil 2002). Perhaps differ- rather than brain size is unclear. If this type ences between closely related species repre- of information accumulation and processing sent the evolutionary elaboration ofmore gen- required an unusually large number of con- erally shared abilities where species are chal- nections per neuron ratherthan a large number lenged to perform disparate specific tasks. of neurons, the time element might represent Consideration of the evolution of cognition the difficulty of making so many connec- leads to another question, namely whether tions. variation in cognition is graded or exhibits The second kind of cognition might be thresholds across which abilities change rap- thought of as prospective, the ability to work idly. This question frequently arises in discus- out a novel solution to a novel problem, per- sions about the evolution of human intelli- haps involving the generation of predictive gence. Some authors suggest that at some scenarios based on accumulated experience point in our evolutionary lineage brain size and detailed observation. However, why this and intelligent behavior became self-acceler- kind of thinking might require a large brain ating and our cognitive abilities increased rap- rather than a complex brain is unclear. None- idly, creating a substantial gap between hu- theless, the analyses presented here indicate mans and otherprimates in both brain size and that those birds with relatively higher capac- intelligent behavior (e.g., Dunbar 1993, Aboi- ities for reasoning and problem solving, such tiz and Garcia 1997). Similar thresholds at as corvids and parrots, tend to have large lower levels of cognitive ability might also brains. lead to the creation of gaps in intelligent be-

See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.