Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission.

Transcription

1 Vigilance and Foraging Substrate: Anti-Predatory Considerations in a Non-Standard Environment Author(s): Steven L. Lima Source: Behavioral Ecology and Sociobiology, Vol. 30, No. 3/4 (1992), pp Published by: Springer Stable URL: Accessed: 22/01/ :26 Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We work with the scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that promotes the discovery and use of these resources. For more information about JSTOR, please contact support@jstor.org. Springer is collaborating with JSTOR to digitize, preserve and extend access to Behavioral Ecology and Sociobiology.

2 Behav Ecol Sociobiol (1992) 30: Behavioral l Ecology and Sociobiology? Springer-Verlag 1992 Vigilance and foraging substrate: anti-predatory considerations in a non-standard environment Steven L. Lima Department of Life Sciences, Indiana State University, Terre Haute, IN 47809, USA Received June 10, 1991 / Accepted October 1, 1991 Summary. The commonly studied "standard" anti-predatory environment presents animals with spatially-distinct feeding sites and refuges from attack, neither of which necessarily obstructs predator detection. In contrast, tree-trunks provide animals with a markedly "nonstandard" environment in which the foraging substrate itself may be a refuge from attack that unavoidably obstructs predator detection. Thus anti-predatory behavior in this environment should be influenced not only by a perceived risk of attack, but also by the nature of the refuge/foraging substrate itself. Downy woodpeckers (Picoides pubescens) are a common tree-trunk foraging animal, and an experimental analysis of their behavior suggests that they respond appropriately to their nonstandard anti-predatory environment. In particular, anti-predatory vigilance varies strongly with changes in tree trunk diameter. Two modes of vigilance were apparent. In "stationary" vigilance, woodpeckers maintained the position of their feet while rotating their bodies sideto-side to peer around the trunk; "mobile" vigilance involved movement around the trunk itself. Both the frequency and angle of rotation of stationary vigilance increased with trunk diameter, as did the frequency of mobile vigilance. The woodpeckers also held their heads farther away from the trunk surface as diameter increased. All of these measures of vigilance increased under a greater perceived risk of predation. As might be expected given these results, downy woodpeckers avoided thick trunks; they did not, however, prefer the thinnest (least obstructive) available trunks. These preferences may reflect the influence of trunk diameter on thermo-ecological and/or anti-predator considerations not related to vigilance. Overall, this arboreal environment provides an unusual perspective on anti-predator decision-making with several implications for tree-trunk foraging animals in general. Introduction The physical structure of the environment and its behavioral implications are an important focus (implicitly or explicitly) in many studies of animals feeding under the risk of predation. In the majority of such studies, the foraging environment is a mosaic of food-free refuges to which animals flee when attacked, and adjacent unsafe areas in which animals must feed (see Sih 1987; Lima and Dill 1990 for reviews). Animals typical of such "standard" environments act as though they perceive a higher risk of predation as they feed farther from a refuge (but see Lima 1990 a), and understanding this perception is critical to understanding behavior. This is apparent in escape behavior (e.g., Dill and Houtman 1989; Helfman 1989; see Ydenberg and Dill 1986 for a review), choice of feeding site (e.g., Grubb and Greenwald 1982; Schneider 1984; Newman and Caraco 1987; Newman et al. 1988; Bowers 1988; Kotler and Brown 1988; Brown et al. 1988; Piper 1990; Todd and Cowie 1990), and anti-predatory vigilance (e.g., Barnard 1980; Caraco et al. 1980; Holmes 1984; Ekman 1987; Hogstad 1988; see Elgar 1989 for a review). This standard environmental paradigm and associated perceptions of risk break down when applied to a tree-trunk foraging environment typical of animals such as woodpeckers. First, this environment cannot be divided neatly into safe refuges and unsafe feeding areas. In fact, the very substrate upon which animals feed is potentially a refuge from attack; for instance, woodpeckers escape predatory attack by jumping to the opposite side of the tree trunk upon which they are feeding (Sullivan 1984a, b; pers. obs.). Second, this arboreal environment is also unique from the viewpoint of predator detection: the refuge/foraging substrate unavoidably obstructs the visual detection of predators. This might seem behaviorally irrelevant given that the visual obstruction is the refuge itself. However, escape requires getting to the other side of the trunk, and that is the side on which a woodpecker never finds itself when attacked. Thus a tree-trunk forager must both remain vigi-

3 284 lant to predatory attack while feeding, and deal with the visual obstruction provided by its refuge. As part of a broader study on the influence of predators on the avian trunk-foraging guild, this study addresses whether the behavior of downy woodpeckers (Picoides pubescens) suggests that they perceive such a "non-standard" anti-predator environment. The response of downy woodpecker vigilance to experimental manipulations of tree trunk diameter and the perceived risk of predation indicate that they do in fact perceive their foraging substrate as an obstruction to predator detection. Despite this, however, these birds do not prefer to feed on the least visually obstructive trunks under controlled conditions. Methods Study site and woodpeckers. This study was conducted from December 1989 through February 1990 in Kiewig Wood, an Indiana State University property in western Vigo County, Indiana. Kiewig Wood is a 16-ha, unusually mature second-growth forest, in which downy woodpeckers are by far the most abundant tree-trunk foraging animal. All downy woodpeckers were individually recognizable due to substantial differences in markings on the top and back of their heads (Kilham 1983). Throughout this study, a stable group of nine downy woodpeckers (five males and four females) visited the study site each day. Food and the feeding apparatus. Woodpeckers fed on solid, purified beef fat while perched on cut sections of tree trunks of varying diameter (see below). Beef fat tissue was obtained from a local market, minced, and then heated until its entire fat content was liquified. The liquid fat was then strained, poured into holes drilled into the cut logs, and allowed to cool and solidify. At temperatures < 15? C, the purified fat has the consistency of candle wax. All woodpecker species readily consumed this food, as did Carolina chickadees (Parus carolinensis), tufted titmice (P. bicolor), and white-breated nuthatches (Sitta carolinensis). The cut logs (hereafter referred to as patches) were of three diameters: "thin" (3-4 cm); "medium" (8-10 cm) and "thick" (18-22 cm). All were derived from sassafras trees (Sassafras albidum) and were 1.5 m in length. Into each patch were drilled holes 1.1 cm in diameter and 2 cm deep, which were spaced 6 cm apart in a single, straight line along the entire length of the patch. This essentially forces a woodpecker to move straight up a patch while feeding, which is a common pattern of movement in these birds under natural conditions (pers. obs.). Patches were presented to woodpeckers by hanging them vertically from the horizontal cross beam of a support structure. This cross beam, 2 m in length and 3 m above the ground, was held 2 m away from a main vertical support pole by an additional wooden beam. Metal hooks at 1-m intervals in the cross beam held the patches 1.5 m off the ground (measuring from the bottom of a patch). The bottom of each patch was fastened securely to a hook at the base of the support structure by a strong elastic cord; this prevented swaying. The cross beam was oriented parallel to, and 5 m from the edge of the woods. Depending on the experiment in question, either two (placed at opposite ends of the cross beam) or three patches (placed at 1-m intervals) were secured to the support structure. Patches were in place from 20 min before sunrise until the end of the day (when they were often depleted of food). An experimental session began at dawn with the arrival of the first woodpecker and lasted 2.5 h; sessions took place on consecutive days except during those with heavy rain. All woodpecker feeding bouts were video-taped from within a blind 20 m from the support structure. Experiment 1: vigilance and patch diameter This experiment addressed the influence of patch diameter on two aspects of downy woodpecker vigilance behavior: side-to-side body movements (see Fig. 1 a, b) and head movement directly away from the patch ("pull-backs"; Fig. Ic). The influence of the presence of other birds on downy woodpecker vigilance was also examined. This experiment consisted of 18 sessions. The first and last 6 sessions examined side-to-side (STS) vigilance. In order to estimate the angle of body movements during STS vigilance (see below and Fig. 1), the line of holes in each patch was oriented toward the blind such that woodpeckers were video-taped directly from behind. The middle 6 sessions examined pull-backs, and during this segment the patches were oriented such that the birds were video-taped in silhouette. All three patch diameters were used during experiment 1. Three patches of the same diameter were available during a given session. Patch diameters always differed between two consecutive sessions, and all three diameters were used over three consecutive sessions (in random order) before repeating a given diameter. Thus, each six-session segment consisted of two "cycles" through the three patch diameters. Analysis of video tapes. Two modes of side-to-side (STS) vigilance were recognized: "stationary" and "mobile". During a stationary vigilance event (or scan), a woodpecker keeps its feet stationary and peers around the patch after rotating its body from the vertical feeding position (Fig. la); mobile vigilance involves movement through a given angle around the patch itself (Fig. 1 b). The angle of body rotation for a stationary scan was determined directly from a large-screen video monitor using a transparent protractor. For mobile scans, the angle of movement (measuring movement of the feet) was estimated to the nearest 30?. The number of scans of each mode was recorded from the video tapes for each feeding bout. However, it was not practical to measure the angle of each scan in a feeding bout. Thus, a sampling procedure was adopted which involved measuring the first scan during a bout, and then measuring subsequent scans after 20 s from the previously measured scan (during which there were usually several scans). Data were included in the following analyses only if taken from bouts exceeding 60 s in length without any change in the composition of birds feeding at the apparatus; thus a i THIN MEDIUM X THICK Fig. l a-c. Diagrammatic representation of angles (0) measured during scans. a Stationary vigilance: downy woodpeckers drawn to scale on all three patch diameters. b Mobile vigilance (on thick patch). c Pull-back vigilance (on thick patch). Woodpeckers are drawn to indicate the average angle observed for solitary downy woodpeckers on the patch shown

4 285 a minimum of three scans were measured per bout. Overall, STS vigilance during each bout was characterized by six variables for statistical analysis (3 behavioral measures x 2 vigilance modes): average scanning rate (number scans/bout length), average scan angle, and average scan duration (elapsed time) for both stationary and mobile vigilance. The above 20-s sampling procedure was used when analyzing video tapes of pull-back vigilance. Note that these pull-backs are often a part of STS vigilance, and angles were measured at their maximum only when the birds were in an upright position. Each bout during these sessions was characterized by only one variable for statistical analysis: the average pull-back angle. The details of STS scans could not be distinguished reliably from video-tapes of the birds in silhouette, and thus scanning rates, etc., were not determined. Statistical analyses. Each behavioral variable was analyzed via a blocked ANOVA (Wilkinson 1988). Data were blocked among nine individual downy woodpeckers, and the factors in the model were patch diameter and social condition; the latter factor had two levels, according to whether downy woodpeckers were solitary or feeding in the presence of other birds. Post-hoc orthogonal contrasts examined sexual differences in vigilance behavior. The first and last six sessions of experiment 1 were performed under the same conditions, and thus data from these two segments were pooled for the analyses of STS vigilance. Experimental designs following a set of animals over several treatments are useful, often unavoidable, and statistically problematic; strictly speaking, the data gathered do not meet the assumptions of any available statistical procedure. A particular problem concerns the possibility of only partial independence between behavioral observations, which makes it difficult to interpret P values produced by standard tests (such values may be artificially depressed). Two steps were taken to mitigate this problem. First, a maximum of only one datum per individual per session was included in a given ANOVA; data obtained over several sessions were assumed to yield estimates of daily behavioral variability within a given individual. Second, a conservative level of significance was set at rather than the typical value of Note also that the statistical inferences from such designs, strictly speaking, apply only to the given set of birds and not the population in general. There is, however, no reason to believe that the nine woodpeckers in this study were unique in any way. Experiment 2: patch choice This brief experiment examined whether patterns in vigilance shed light on patch diameter preferences in downy woodpeckers. During an experimental session, woodpeckers chose between two patches of different diameters. There were three possible pairwise patch combinations given the three available patch diameters. Each of the three combinations was presented during two sessions, first with one diameter on the left and the other on the right, and then with the positions reversed during the second session. Thus, there were six different situations in which to test woodpecker patch preferences, and each was available in random order over a six-session period, with the exception that a particular combination not be used during two consecutive sessions. A maximum of one choice per individual downy woodpecker per session was included in the analysis. This was the choice made by a given individual the first time during which it was the only bird at the study site. Data were pooled over the two sessions during which a particular pairwise combination of patches was available. Experiment 3. vigilance and the perception of risk This experiment examined whether downy woodpecker STS vigilance responded to an increase in the perceived risk of predation as per an earlier study by Sullivan (1984a); a lack of such a response would suggest that vigilance at the study site is not antipredatory in nature. Altering a downy woodpecker's perception of predation risk presented certain problems. In order to control adequately for the effects of the passage of time during this experiment, "high risk" and "low-risk" sessions had to be alternated in some fashion. Thus, a relatively low-level, "reversible" threat was used to insure that the perception of increased risk during a high-risk session did not carry-over to the next low-risk session. The use of a dangerous natural predator, such as a trained hawk, would be effective in increasing perceived risk, but this change would likely persists for several days. For these reasons, the predatory threat was a human sitting motionless 25 m from the patches on the side opposite the food holes. This threat was apparently reversible, and the woodpeckers' response was similar to that exhibited towards the coyotes (Canis latrans) in the area (pers. obs.). This experiment covered eight alternating high-risk (threat present) and low-risk (threat absent) sessions. Only the thick patches were used, as vigilance on these patches was most readily apparent, and the rapidly warming winter weather (and subsequent increase in territorial behavior) made uncertain the completion of a more extensive experiment involving all patch diameters. All woodpecker bouts were video-taped and analyzed as described for the STS vigilance sessions of experiment 1. Results Experiment. vigilance and patch diameter Patch diameter, as well as the social environment, clearly affected the rate of stationary side-to-side (STS) scanning by downy woodpeckers (Fig. 2a). Scanning increased with increasing patch diameter, and decreased in the presence of other birds. A blocked ANOVA indi , v b O ' Ix 0.04 < THIN MEDIUM THICK z " >- 0.5 >- 25 TI MI TI TI MIU 0.4- Z 0.3 C215 - <o10 H 0.2 PATCH DIAMETER THIN MEDIUM THICK Fig. 2a-d. Aspects of downy woodpecker body movements during PATCH DIAMETER side-to-side vigilance in experiment 1. Shown are averages (SE) for each patch diameter given solitary woodpeckers (solid bars) and woodpeckers feeding in the presence of other birds (open bars). Rate a stationary of scanning b Rate of mobile scanning. c Angle of stationary scanning. d Duration (elapsed time) of stationary scans

5 286 Table 1. ANOVA results for experiment 1: Table entries are P values for the significance of the factor in question Behavioral variable Source n Indivi- Patch Soc. env D x S duals diameter (S) (D) Stationary scans Rate < <0.001 <0.001 Angle <0.001 <0.001 <0.001 Duration Mobile scans Rate < Pull-backs Angle < Soc. Env., social environment cated that both of these effects were significant, as was the patch diameter - social environment interaction (Ps<0.005, Table 1). Also indicated is a significant difference in scanning rates among individuals (Table 1). These differences were largely sexual, with females scanning less frequently than males (post-hoc contrast, F1,140=9.41, P<0.005). The rate of mobile scanning also increased with patch diameter and decreased in the presence of other birds (Fig. 2b), but only the former effect was statistically significant (Table 1). Mobile scanning rates were over an order of magnitude lower than stationary rates. The angle through which downy woodpeckers rotated during stationary scans increased significantly with patch diameter and decreased in the presence of other birds (Fig. 2c, Table 1). There was also a significant diameter - social environment interaction (Table 1). Average angles ranged from 19? to 37?, but some scans on thick patches occasionally exceeded 50?. Despite the wider scans on thicker patches, the average duration of scans was apparently uninfluenced by either patch diameter or social environment (Fig. 2d, Table 1). This indicates that scans were made at a higher angular velocity as patch diameter increased, as was qualitatively apparent when viewing the video tapes. There were no apparent individual differences in either scan angle or duration (Table 1); only stationary scanning rates showed individual differences in experiment 1. Mobile scans were too few in number for the above statistical analyses of angles moved. Pooling over all patch types, however, the average (n=17) angle "jumped" during mobile scans was 59?, and the average duration was 2.1 s. Patch diameter also influenced the angle at which woodpeckers pulled back away from patches while scanning ("pull-backs"). Average pull-back angles increased significantly with patch diameter, and were unaffected by social environment (Fig. 3, Table 1). Average pullbacks varied from 13? to 32?, but occasional pull-backs on thick patches exceeded 50?. The above analyses used a convenient but crude measure of the social environment: the presence or absence (D25y ~Fig. * 3. o Average (SE) pull- 25 back angles for downy y20 H woodpeckers during o experiment 1 as a function m 15 * - i * of patch diameter and social 10 o - environment (solid bars: - CZ a z solitary woodpeckers; open 5 d d I dj bars: woodpeckers in the 0o _ B _ presence of other birds) THIN MEDIUM THICK Gn8 X o Cn) 9 PATCH DIAMETER z o 0 8 < 0.2 -? A 8 A? (1n? A NO. ADDITIONAL BIRDS Fig. 4. Rate of stationary scanning as a function of the number of birds present in addition to the "focal" woodpecker. Each point represents the average rate from a single bout for a solitary downy woodpecker (open circles), or a woodpecker feeding in the presence of only dominant woodpecker species (closed circles) or subordinate species (triangles). All data are from medium patches; data from the other two patch diameters are quantitatively very similar of other individuals. Figure 4 shows that stationary scanning rates in focal downy woodpeckers decreased continuously with increasing sociality. Furthermore, scanning decreased in the presence of both subordinate species (chickadees, titmice or nuthatches) and dominant species (mostly hairy woodpeckers P. villosus, red-bellied woodpeckers Melanerpes carolinus, and northern flickers Colaptes auratus); the effects of these two groups could not be compared statistically as they occurred over completely different ranges of sociality. Experiment 2: patch choice Downy woodpeckers clearly preferred thin and medium patches over thick patches (Fig. 5; sign test, P<0.05). The apparent preference for medium over thin patches was not significant. There were no indications of any preference for a particular side of the experimental apparatus, nor were there any apparent sexual differences in behavior. Experiment 3: vigilance and the perception of risk Stationary scanning rates increased significantly during high-risk sessions (Fig. 6a; Table 2). Mobile vigilance was affected similarly by risk (Fig. 6b; Table 2), but the

6 287 0 LL 15 Table 2. ANOVA results for experiment 3: Table entires are P values for the significance of the factor in question Behavioral variable Source n Indivi- Risk Soc. env. R x S duals (R) (S) LU 5 0. CHOICE Fig. 5. Patch choice results from experiment 2: the number of times solitary woodpeckers chose a given patch during the pairwise test indicated along the x-axis. Stationary scans Rate <0.001 < Angle < < Duration Mobile scans Rate < Angle' Duration Data derive only from those bouts in which the rate of mobile vigilance >0 HI)I WHH pck? d. a. m be Z02 z L HIGH OW HIGH LOW ^ 45-7z , a 0.8 i', the 50 native results from experi- 5 r0.85 0< v< <z sc a c z >_ LEVEL OF RISK LEVEL OF RIS Z 0.3 o 20 ~ 15l C ~ I I~ -I 0_ 0.0 HIGH LOW HIGH LOW LEVEL OF RISK 0.2 LEVEL OF RISK Fig. 6a-d. Aspects of downy woodpecker body movements during side-to-side vigilance in experiment 3: averages (SE) for each level of predation risk for solitary woodpeckers (solid bars) and woodpeckers feeding in the presence of other birds (open bars). a Rate of stationary scanning. b Rate of mobile scanning. c Angle of stationary scanning. d Duration (elapsed time) of stationary scans overall effect was more marked than seen in stationary vigilance; the effect of sociality was not significant (P> 0.005). Since the two modes of side-to-side and mobile vigilance are mutually exclusive, an increase in one might lead to a decrease in the other. This, however, was not the case relative to the thick patch results from experiment 1 (Fig. 2a, b); stationary scanning increased ca. 20% and mobile scanning increased ca. 250% during experiment 3. There were no significant differences in individual behavior for either of these two scanning rates, nor for any of the other behavioral measures in Fig. 6 (Table 2). There were also no significant patch diameter - social environment interactions for any behavioral measure. The angle of stationary vigilance increased significantly during high-risk sessions (Fig. 6c; Table 2). A similar pattern is evident for the duration of stationary scans (Fig. 6d), but neither the risk nor sociality effects were significant (Table 2). Neither the angle nor the duration of mobile scans were influenced significantly by risk or sociality (Table 2); pooling over patch diameters, the average angle and scan duration for mobile vigilance were 53? and 1.5 s, respectively. Discussion A unique aspect of this "non-standard", arboreal environment is that the very substrate upon which animals feed is potentially a refuge from attack that simultaneously obstructs predator detection. The behavioral responses of downy woodpeckers to changes in their foraging substrate indicates that these birds perceive the nonstandard nature of their environment. In particular, the woodpeckers' progressively greater investment in antipredatory vigilance with increasing trunk diameter suggests strongly that they perceive their refuge/foraging substrate as an obstruction to predator detection. This conclusion regarding predator detection rests strongly on the assumption that vigilance is anti-predatory in nature, which is clearly supported not only by the results of experiment 3 (Fig. 6), but also by Sullivan's (1984a) observation that vigilance in downy woodpeckers responds positively to a recent appearance of a model hawk. Furthermore, the influence of social feeding on vigilance (Fig. 4) also fits the "classic" form predicted by theories of anti-predatory vigilance (see Lima 1990b) and seen earlier in downy woodpeckers responding to the presence of chickadees (Sullivan 1984a, b). Nevertheless, some studies suggest that much vigilance may be directed toward detecting aggression from dominant individuals rather than predators (e.g. Knight and Knight 1986; Waite 1987; see also Elgar 1989). However, Fig. 4 shows clearly that vigilance decreased markedly even in the presence of dominant woodpecker species. Perhaps a sensitivity to (intraspecific) male-male aggression is the reason for the greater vigilance in male vs. female downy woodpeckers, but this effect was observed only during experiment 1. Overall, while the anti-aggres-

7 288 sion component of vigilance in downy woodpeckers may be real, it does not appear very strong in relation to the anti-predatory component. Concerning vigilance behavior in general, an interesting effect in this arboreal environment is that an increase in the rate of scanning does not necessarily lead to an increase in the probability of predator detection. For instance, the low scanning rates and degree of body movement by downy woodpeckers on thin patches very likely reflect the fact that very little body movement is needed to peer around these patches (Fig. 1 a); in fact, the small sideways head movements necessary for extracting food are probably acting simultaneously as scans. On thicker patches, even with an increase in the angle of body movement (Fig. 1 a), woodpeckers always have much of their view blocked. The observed increase in scanning rates on thicker patches might alleviate this problem to some extent, but woodpeckers on thin patches likely have a higher probability of predator detection despite their relative lack of overt scanning. Analogous concerns might apply in "standard" environments in which visual obstructions exist in the vicinity of feeding sites (Elgar et al. 1984; Metcalfe 1984), but increasing scanning in such environments will generally lead to increased detection (Lima 1987). Subtle influences on vigilance may be many (Lima 1987), particularly in this arboreal environment. One such influence may be an undetected correlation between attack rate and trunk diameter: raptors may preferentially attack woodpeckers on visually obstructive tree trunks. This possibility cannot be assessed given the lack of attacks at the study site (although Cooper's hawks (Accipiter cooperii) were sighted several times). If present, however, this effect would undoubtedly accentuate the role of visual obstructions as a determinant of vigilance (see Lima 1987). A second subtle effect of trunk diameter concerns the probability of escape (POE) from attack. Several studies demonstrate that the separation between food and refuge (i.e. POE) is an important determinant of anti-predator behavior (Lima and Dill 1990). While there is no such separation between food and refuge in the current environment, an analogy does exist. Since woodpeckers escape attack by jumping to the opposite side of a trunk, it seems likely that there exists an optimal trunk diameter for escape. In particular, very thick trunks may require too much time to reach the safety of the opposite side, whereas thin trunks may provide too little protection from attack; Sullivan's (1984a) observation that downy woodpeckers abandon thin branches (< 8 cm) in favor of thicker ones after a simulated raptor attack supports this idea. If the POE is maximal for trunks of intermediate diameter, then minimal vigilance might occur at that diameter. The data in Figs. 2 and 3, however, suggest no such effect. Finally, foragers on thick tree trunks might attempt to "hide" from unseen predators rather than attempt early predator detection. However, there is no suggestion that such hiding occurred in this study, and indeed it may not be a viable option for foragers such as woodpeckers that must move continually to locate and con- sume food. Interestingly, Scott et al. (1976) suggest that the "sit-and-wait" lizard Anolis frenatus does use tree trunks to hide from predators. These lizards do not actually engage in overt vigilance movements, nor do they actually feed on the tree trunks themselves. Instead, they remain motionless in an alert posture while scanning the ground for prey (see also Moermond 1986). Scott et al. (1976) posit that these lizards would actually prefer the thickest available perches but for the fact that such perches also hinder the detection of their terrestrial prey. What light do the vigilance results shed on foraging substrate choice by woodpeckers? Field observations indicate that downy woodpeckers prefer substrates thinner than those used in the present study (Peters and Grubb 1983), suggesting that these birds prefer visually unobstructive substrates. However, under controlled food availability, the present woodpeckers showed no overall preference for the thinnest patch diameter (Fig. 5). It seems likely that, in addition to vigilance, diameter-dependent escape and thermo-ecological considerations (e.g., blockage of wind by thicker patches, cf. Bakken 1990; Grubb 1977; Grubb and Greenwald 1982) may be important in substrate choice, both of which would favor preferences for thicker trunks. If so, then observatios that free-living downy woodpeckers prefer thin foraging substrates (Peters and Grubb 1983) suggest that these birds are putting themselves at risk for a greater energetic gain. Many of the above considerations should apply gen- erally to species foraging in the this non-standard arboreal environment. Other woodpecker species, in particular, should have perceptions very similar to those of downy woodpeckers; in this regard, the visual obstruction provided by a given trunk diameter, as well as its effectiveness as a refuge from attack, should depend strongly on body size. Tree-trunk foraging birds like creepers (Certhiidae), woodcreepers (Dendrocolaptidae), nuthatches (Sittidae), piculets (Picumninae), etc., might have woodpecker-like perceptions of their environments, but these birds differ morphologically from woodpeckers (Norberg 1981), and information as basic as their escape behavior appears to be lacking. However, trunk-feeding arboreal lizards may exhibit escape behavior similar to that of woodpeckers (pers. obs.), and perhaps they too perceive a non-standard foraging environment. Further work with an appreciation of the anti-predatory implications of the arboreal tree-trunk environment promises to shed light on both the behavior of, and ecological relationships among, species of tree-trunk foraging animals. Acknowledgements. Indiana State University provided financial assistance throughout this study. R.M. Lee, A. Sih, K.A. Sullivan, and anonymous reviewers made several useful comments on the manuscript. References Bakken GS (1990) Estimating the effect of wind on avian metabolic rate with standard operative temperature. Auk 107:

8 289 Barnard CJ (1980) Flock feeding and time budgets in the house sparrow (Passer domesticus L.). Anim Behav 28: Bowers MA (1988) Seed removal experiments on desert rodents: the microhabitat by moonlight effect. J Mammal 69: Brown JS, Kotler BP, Smith RJ, Wirtz EO (1988) The effects of owl predation on the foraging behavior of heteromyid rodents. Oecologia 76: Caraco T, Martindale S, Pulliam HR (1980) Avian time budgets and distance to cover. Auk 97: Dill LM, Houtman R (1989) The influence of distance to refuge on flight initiation distance in the gray squirrel (Sciurus carolinensis). Can J Zool 67: Ekman J (1987) Exposure and time use in willow tit flocks: the cost of subordination. Anim Behav 35: Elgar MA (1989) Predator vigilance and group size in mammals and birds: a critical review of the empirical evidence. Biol Rev 64:13-33 Elgar MA, Burren PJ, Posen M (1984) Vigilance and perception of flock size in foraging house sparrows (Passer domesticus L.). Behaviour 90: Grubb TC Jr (1977) Weather-dependent foraging behavior of some birds wintering in a deciduous woodland: horizontal adjustments. Condor 79: Grubb TC Jr, Greenwald L (1982) Sparrows and a brushpile: foraging responses to different combinations of predation risk and energy cost. Anim Behav 30: Helfman GS (1989) Predation-risk-sensitive predator avoidance in damselfish-trumpetfish interactions. Behav Ecol Sociobiol 24:47-58 Hogstad O (1988) Social rank and antipredator behavior of willow tits Parus montanus in winter flocks. Ibis 130:45-56 Holmes WG (1984) Predation risk and foraging behavior of the hoary marmot in Alaska. Behav Ecol Sociobiol 15: Kilham L (1983) Life history studies of woodpeckers of eastern North America. Nuttall Ornithological Club, Cambridge, Massachusetts Knight SK, Knight RL (1986) Vigilance patterns in bald eagles feeding in groups. Auk 103: Kotler BP, Brown JS (1988) Environmental heterogeneity and the coexistence of desert rodents. Annu Rev Ecol Syst 19: Lima SL (1987) Vigilance while feeding and its relation to the risk of predation. J Theor Biol 124: Lima SL (1990a) Protective cover and the use of space: different strategies in finches. Oikos 58: Lima SL (1990b) The influence of models on the interpretation of vigilance. In: Bekoff M, Jamieson D (eds) Interpretation and explanation in the study of animal behavior: vol. 2. Explanation, evolution and adaptation. Westview, Boulder, Colorado Lima SL, Dill LM (1990) Behavioral decisions made under the risk of predation: a review and prospectus. Can J Zool 68: Metcalfe NB (1984) The effects of habitat on the vigilance of shorebirds: is visibility important? Anim Behav 32: Moermond TC (1986) A mechanistic approach to the study of communities: Anolis lizards and birds. Am Zool 26:23-37 Newman JA, Caraco T (1987) Foraging, predation hazard and patch use in grey squirrels. Anim Behav 35: Newman JA, Recer GM, Zwicker SM, Caraco T (1988) Effects of predation hazard on foraging "constraints": patch use strategies in grey squirrels. Oikos 53:93-97 Norberg RA (1981) Why foraging birds in trees should climb and hop upwards rather than downwards. Ibis 123: Peters WD, Grubb TC Jr (1983) An experimental analysis of sexspecific foraging in the downy woodpecker, Picoides pubescens. Ecology 64: Piper WH 61990) Exposure to predators and access to food in wintering white-throated sparrows Zonotrichia albicollis. Behaviour 112: Schneider KJ (1984) Dominance, predation, and optimal foraging in white-throated sparrow flocks. Ecology 65: Scott NJ Jr, Wilson DE, Jones C, Andrews RM (1976) The choice of perch dimensions by lizards of the genus Anolis (Reptilia, Lacertilia, Iguanidae). J Herpetol 10:75-84 Sih A (1987) Predators and prey lifestyles: an evolutionary and ecological overview. In: Kerfoot WC, Sih A (eds) Predation: Direct and Indirect Impacts on Aquatic Communities. University Press of New England, Hanover New Hampshire, pp Sullivan KA (1984a) Information exploitation by downy woodpeckers in mixed-species flocks. Behaviour 91: Sullivan KA (1984b) The advantages of social foraging in downy woodpeckers. Anim Behav 32:16-22 Todd IA, Cowie RJ (1990) Measuring the risk of predation in an energy currency: field experiments with foraging blue tits, Parus caeruleus. Anim Behav 40: Waite TA (1987) Dominance-specific vigilance in the tufted titmouse: effects of social context. Condor 89: Wilkinson L (1988) SYSTAT: The system for statistics. SYSTAT Inc, Evanston, Illinois Ydenberg RC, Dill LM (1986) The economics of fleeing from predators. Adv Study Behav 16: