Foraging and community consequences of seed size for coexisting Negev Desert granivores

We examined the effects of seed size on patch use and diet selection for three co-existing Negev Desert granivores: Allenby's gerbil ( Gerbillus allenbyi ), greater Egyptian sand gerbil ( Gerbillus pyramidum ), and crested lark ( Galerida cristata ). We manipulated size and spatial distribution of seeds in experimental food patches and quantified foraging behavior by measuring giving-up densities (GUDs: the amount of food remaining in a resource patch following exploitation by a forager). In one experiment, we presented small (<1.4 mm in diameter cracked wheat), medium (2.0–3.3 mm), and large (>3.4 mm) seeds in separate trays; in a second, we presented small and medium seeds separately and mixed together. Gerbils had a higher handling time efficiency on smaller seeds, but a much higher encounter probability on larger seeds (20 times higher on large than medium seeds, and 2–5 times higher on medium than small seeds). This led gerbils to have significantly lower GUDs on larger seeds than smaller seeds and to harvest a higher proportion of the larger seeds. When presented with rich and poor patches, G. allenbyi tended to equalize GUDs in both patches, indicating a quitting harvest rate rule for patch exploitation. In contrast, larks appeared to use a fixed time rule for patch exploitation. For larks, seed size did not influence encounter probabilities, and they showed no seed-size selectivity. Still, larks had higher handling efficiencies on smaller than larger seeds, and consequently had a significantly lower GUD on small than medium seeds. Despite large differences between the gerbils and larks in their foraging, our results do not support species coexistence via seed-size partitioning: the larks had much higher GUDs than the gerbils on all seed sizes. Nonetheless, seed size, seed abundance, seed distribution and the animal's patch use behavior all played major roles in determining gerbils' and larks' diet selectivities and GUDs.     

The effects of habitat manipulation on population distribution and foraging behavior in meadow voles

Individuals, free to choose between different habitat patches, should settle among them such that fitness is equalized. Alternatives to this ideal free distribution result into fitness differences among the patches. The concordance between fitnesses and foraging costs among inhabitants of different quality patches, demonstrated in recent studies, suggests that the mode of habitat selection and the resulting fitness patterns may have important implications to the resource use of a forager and to the survival of its prey. We studied how coarse scale selection between habitat patches of different quality and quitting harvest rate in these patches are related to each other and to fine scale patch use in meadow voles (Microtus pennsylvanicus). To demonstrate these relationships, we manipulated habitat patches within large field enclosures by mowing vegetative cover and adding supplemental food according to a 2×2 factorial design. We tracked vole population densities, collected giving‐up densities (GUDs, a measure of patch quitting harvest rate), and monitored the removal of seeds from lattice grids with 1.5 m intervals (an index of fine‐scale space use) in the manipulated habitat patches. Changes in habitat quality induced changes in habitat use at different spatial scales. In preferred habitats with intact cover, voles were despotic and GUDs were low, but increased with the addition of food. In contrast, voles in less‐preferred mowed habitats settled into an ideal free distribution, GUDs were high and uninfluenced by the addition of food. Seed removal was enhanced by the presence of cover but inhibited by supplemental food. Across all treatments, vole densities and GUDs were strongly correlated making it impossible to separate their effects on seed removal rates. However, this relationship broke down in unmowed habitats, where GUDs rather than vole density primarily influenced seed removal by voles. GUDs and seed removal correlated with predation on tree seedlings formerly planted into the enclosures, demonstrating the mechanisms between coarse‐scale habitat manipulations and community level consequences on a forager's prey.     

Defense of Food Supply by Eusocial Colonies

Overdispersion of colonies exists in many eusocial insects.Overdispersion can be generated by direct attack on coloniesor founders, by defense of space, by defense of food resourcesbeing harvested, or by exploitative competition. When directcompetitive interactions lead to colony overdispersion, territorialityis said to occur. Whereas solitary territory holders typicallydefend space, most eusocial colonies defend resource patchesrather than space per se. Also unlike solitary territory holders,colonies with forager communication can simultaneously defendseveral spatially separated food patches. A model explores optimalnumbers of scouts (discoverers of patches) and recruits (followers)needed to maximize net rate of energy intake by the colony.Territorial costs are added to the model by requiring a higherinvestment of foragers per unit resource collected. Accordingto the model, optimal colony size and percentage scouts aremore sensitive to changes in patch size than in patch density.If patch defense is required for resource control, a declineoccurs in optimal percentage of scouts; the decline is greatestfor small colonies. Colonies that must defend patches in orderto harvest from them suffer a loss in net energy intake; theloss is greatest for small colonies. It is predicted that amongeusocial insects, those with territorial defense of resourcesshould preferentially visit large patches and have comparativelylarge colony sizes and relatively few scouts. Ways of testingthese predictions are discussed.     

Depletion of seed patches by Merriam’s kangaroo rats: are GUD assumptions met?

Depletion of experimental seed patches by granivorous animals often is used as a qualitative assay of foraging activity. An optimal foraging model suggests that seed amounts remaining when foragers leave patches ("giving-up-density", GUD) also provide quantitative measures of foraging economics, diet strategies and foraging abilities. Such quantitative uses of GUDs rest on several largely untested assumptions. We tested two of these with Merriam's kangaroo rats: that gain curves are smoothly decelerating, and that foragers leave patches at a constant harvest rate. Harvest rates indeed declined with patch residence time, but in the piecewise linear fashion expected of systematic search. Animals also revisited areas within patches less frequently than expected with random search. In the field, they depleted patches in multiple visits and did not use a constant-rate leaving rule. These deviations from model assumptions cast doubt on inferences about foraging ecology that have been based on quantitative GUD theory.     

Seasonal variation in patch use in a tropical African environment

We used giving-up densities (GUD) to study patch use decisions of small granivorous passerines throughout the year. We measured GUDs continuously in four sites for a period of 9–10 months per year during 2004 and 2005 in a savannah area in Jos, central Nigeria. The study thus covered a period from the middle of the dry season, through the wet season to the beginning of the next dry season in each year. We placed experimental food patches in both open areas and in cover to investigate possible effects of predation risk and thermal hazard on the foraging behavior of the birds. We found a difference in GUDs between the microhabitats, with a consistently lower GUD in cover throughout the year and for the two years. During both years GUDs followed a pattern coinciding with the seasonal change in local seed availability. An initial decline in GUDs late in the dry season was followed by a steady increase during and after the rains. A similar trend in GUDs observed for both years supports the conclusion that GUDs measure the feeding birds' assessment of environmental quality, possibly in combination with other factors changing predictably during the year. We conclude that food abundance may act with other environmental and ecological factors to affect foraging decisions throughout the year.     

Spatial and temporal scaling in habitat utilization by klipspringers (Oreotragus oreotragus) determined using giving‐up densities

An animal's pattern of habitat use can reveal how different parts of its environment vary in quality based on the costs (such as predation risk) and benefits (such as food intake) of using each habitat. We studied klipspringer habitat use in Augrabies Falls National Park, South Africa using giving‐up densities (GUDs; the amount of food remaining in a resource patch following exploitation) in experimental food patches. We tested hypotheses related to how salient habitat variables might influence klipspringers' perceptions of foraging costs. At small spatial scales (3–4 m), klipspringer GUDs did not vary with cover and open microhabitats, or with the four cardinal aspects (shading) around shrubs. Adding water adjacent to food patches did not influence GUDs, showing that water is not a limiting complementary resource to food. Generally, klipspringers do not appear to be physiologically constrained. There was no difference in GUDs between four daily time periods, or between summer and winter; however, a significant interaction effect of time‐of‐day with season resulted from GUDs during the midday time period in winter being significantly higher (perceived value lower) than during the same time period in summer. At moderate spatial scales (10–60 m), klipspringer GUDs increased with distance from rocks because of increased predation risk. Based on GUDs collected at the largest scale (two 4.41‐ha grids), klipspringers preferred foraging at greater distances from drainage lines and on pebble and cobble substrates. Overall, this study has shown the efficacy of measuring GUDs to determine klipspringers' habitat utilization while foraging.     

On the value of information: studying changes in patch assessment abilities through learning

Little is known about how animals acquire and use prior information, particularly for Bayesian patch assessment strategies. Because different patch assessment strategies rely upon distinct capabilities to obtain information, we analyzed whether foragers can alter their foraging strategy when they exploit predictable patches with periodic renewal. For this, we evaluated if learning contribute to increase foraging efficiency by improving patch assessment abilities in degus (Octodon degus), a diurnal caviomorph rodent from central Chile. Single degus exploited pairs of depleting patches that were renewed daily. During the initial two days of the experiment, degus exploited patches in agreement with a fixed‐time strategy, i.e. at the population level, giving‐up densities (GUD) were not distinguishable from density‐independence (i.e. consumption proportional to initial patch densities), and richer patches were under‐exploited. After day five, degus improved significantly their assessment strategy, showing agreement with Bayesian information updating. However, on day 15 and afterwards, degus foraged patches in agreement with a prescient strategy, because GUDs across patches indicated positive density‐dependence and equalization of GUDs. Although highly variable, the GUD ratio between rich and poor patches decreased significantly throughout time. Within‐subject data showed that as subjects learned patch qualities they showed a tendency toward GUD equalization and differentiation from density‐independence. By the end of the experiment, degus allocated more time to richer patches during the initial period of each trial, and allocated similar amounts of time by the end of trials. Further, the first visit of a session was significantly biased toward the rich patch by the final days of the experiment. The results suggest that assessment abilities can change when exploiting novel but predictable patches. When degus can incorporate adequate environmental information, prior and current information may become accurate enough to make animals exploit patches efficiently.     

Habitat and patch use by hyraxes: there’s no place like home?

Models of central place foraging predict that animals should forage more thoroughly in resource patches located closer to the central place. Travel time, cost of transporting food back to the central place, and exposure to predators should all act to increase foraging costs with increasing distance from the refuge. We examined habitat and patch use in rock hyraxes ( Procavia capensis ) inhabiting a group of kopjes in a semiarid savanna, Augrabies Falls National Park, South Africa. We tested the prediction of more intense patch use closer to the central place by measuring giving-up densities (GUDs) in experimental resource patches set at four different distances from the kopje and in two microhabitats differing in cover. Surprisingly, hyraxes had their lowest GUDs at intermediate distances from the kopje. These unexpected results suggest that the sentinel behaviour of hyraxes alters the probability of detection of predators for animals foraging away from the kopje.     

Foraging on spatially distributed resources with sub‐optimal movement,imperfect information,and travelling costs: departures from the ideal free distribution

Ideal free distribution (IFD) theory offers an important baseline for predicting the distribution of foragers across resource patches. Yet it is well known that IFD theory relies on several over‐simplifying assumptions that are unlikely to be met in reality. Here we relax three of the most critical assumptions: (1) optimal foraging moves among patches, (2) omniscience about the utility of resource patches, and (3) cost‐free travelling between patches. Based on these generalizations, we investigate the distributions of a constant number of foragers in models with explicit resource dynamics of logistic type. We find that, first, when foragers do not always move to the patch offering maximum intake rate (optimal foraging), but instead move probabilistically according to differences in resource intake rates between patches (sub‐optimal foraging), the distribution of foragers becomes less skewed than the IFD, so that high‐quality patches attract fewer foragers. Second, this homogenization is strengthened when foragers have less than perfect knowledge about the utility of resource patches. Third, and perhaps most surprisingly, the introduction of travelling costs causes departures in the opposite direction: the distribution of sub‐optimal foragers approaches the IFD as travelling costs increase. We demonstrate that these three findings are robust when considering patches that differ in the resource's carrying capacity or intrinsic growth rate, and when considering simple two‐patch and more complex multiple‐patch models. By overcoming three major over‐simplifications of IFD theory, our analyses contribute to the systematic investigation of ecological factors influencing the spatial distribution of foragers, and thus help in deriving new hypotheses that are testable in empirical systems. A confluence of theoretical and empirical studies that go beyond classical IFD theory is essential for improving insights into how animal distributions across resource patches are determined in nature.     

An evolutionarily stable joining policy for group foragers

For foragers that exploit patchily distributed resources thatare challenging to locate, detecting discoveries made by otherswith a view to joining them and sharing the patch may oftenbe an attractive tactic, and such behavior has been observedacross many taxa. If, as will commonly be true, the time takento join another individual on a patch increases with the distanceto that patch, then we would expect foragers to be selectivein accepting joining opportunities: preferentially joining nearbydiscoveries. If competition occurs on patches, then the profitabilityof joining (and of not joining) will be influenced by the strategiesadopted by others. Here we present a series of models designedto illuminate the evolutionarily stable joining strategy. Weconfirm rigorously the previous suggestion that there shouldbe a critical joining distance, with all joining opportunitieswithin that distance being accepted and all others being declined.Further, we predict that this distance should be unaffectedby the total availability of food in the environment, but shouldincrease with decreasing density of other foragers, increasingspeed of movement towards joining opportunities, increased difficultyin finding undiscovered food patches, and decreasing speed withwhich discovered patches can be harvested. We are further ableto make predictions as to how fully discovered patches shouldbe exploited before being abandoned as unprofitable, with discoveredpatches being more heavily exploited when patches are hard tofind: patches can be searched for remaining food more quickly,forager density is low, and foragers are relatively slow intraveling to discovered patches.     

Does group foraging promote efficient exploitation of resources?

Increased avoidance of food patches previously exploited by other companions has been proposed as one adaptive benefit of group foraging. However, does group foraging really represent the most efficient way to exploit non- or slowly-renewing resources? Here, I used simulations to explore the costs and benefits of exploiting non-renewing resources by foragers searching for food patches independently or in groups in habitats with different types of resource distribution. Group foragers exploited resources in a patch more quickly and therefore spent proportionately more time locating new patches. Reduced avoidance of areas already exploited by others failed to overcome the increased time cost of searching for new food patches and group foragers thus obtained food at a lower rate than solitary foragers. Group foraging provided one advantage in terms of a reduction in the variance of food intake rate. On its own, reduced avoidance of exploitation competition through group foraging appears unlikely to increase mean food intake rate when exploiting non-renewing patches but may provide a way to reduce the risk of an energy shortfall.     

Co-Occurrence and Occupancy of Mourning Doves and Eurasian Collared-Doves

Understanding how land cover and potential competition with invasive species shape patterns of occupancy, extirpation, and colonization of native species across a landscape can help target management for declining native populations. Mourning dove (Zenaida macroura) populations have declined throughout the United States from 1965–2015. The expansion of the Eurasian collared-dove (Streptopelia decaocto), an introduced species with similar food preferences, may further threaten mourning dove populations. We analyzed data from 2009–2016 from a large-scale monitoring program in the Western Great Plains of the United States in a 2-species occupancy model to assess the effects of collared-doves on mourning dove distributions, while accounting for imperfect detection and variation in land cover across the landscape. Mourning dove occupancy was stable or increasing across our study area, and despite overlap in resource use and co-occurrence between mourning doves and Eurasian collared-doves, we found no evidence that collared-doves are extirpating mourning doves from preferred habitat during the breeding season. © 2020 The Wildlife Society.     

Patch use behaviour of Elephantulus myurus and Micaelamys namaquensis: the role of diet,foraging substrates and escape substrates

We investigated the role of diet and substrate features in the coexistence and habitat affinities of the rock elephant shrew, Elephantulus myurus, and the Namaqua mouse, Micaelamys namaquensis. We measured giving‐up densities at experimental food patches that varied in foraging substrate, escape substrate surrounding the food patch and food type. In terms of food consumption, E. myurus favoured pebble (63% total harvest) over sand and sawdust, whereas M. namaquensis favoured foraging in sand (48% total harvest) over sawdust (29%) and pebbles (23%). Mealworms comprised most of E. myurus’s harvest, and M. namaquensis harvested seeds the most, followed by alfalfa and mealworms. In terms of escape substrates, M. namaquensis had significantly higher GUDs when the food patch was surrounded by tussocks of sedge (average 28.11 seeds/patch) than rock surfaces (17.41) or by bush/crevice (14.36). In conjunction with morphologic adaptations, E. myurus detects and recovers food using its snout and long tongue, and M. namaquensis digs and handles foods with its forepaws. The different foraging preferences of E. myurus and M. namaquensis suggest that the interaction of substrates with food types characterize their niches and promote coexistence. Elephantulus myurus travelled greater distances, whereas M. namaquensis was selective for microhabitats offering refuge or traction.     

Knowing your habitat: linking patch-encounter rate and patch exploitation in parasitoids

According to optimal foraging theory, animals should decidewhether or not to leave a resource patch by comparing the currentprofitability of the patch with the expected profitability ofsearching elsewhere in the habitat. Although there is abundantevidence in the literature that foragers in general are wellable to estimate the value of a single resource patch, theirdecision making has rarely been investigated with respect tohabitat quality. This is especially true for invertebrates.We have conducted experiments to test whether parasitic waspsadjust patch residence time and exploitation in relation tothe abundance of patches within the environment. We used thebraconid Asobara tabida, a parasitoid of Drosophila larvae,as our model species. Our experiments show that these waspsreduce both the residence time and the degree of patch exploitationwhen patches become abundant in their environment, as predictedby optimal foraging models. Based upon a detailed analysis ofwasp foraging behavior, we discuss proximate mechanisms thatmight lead to the observed response. We suggest that parasitoidsuse a mechanism of sensitization and desensitization to chemicalsassociated with hosts and patches, in order to respond adaptivelyto the abundance of patches within their environment.     

Reduced resource defence in an uncertain world: an experimental test using captive nutmeg mannikins

Recent models of economic defence in a group-foraging context predict that the frequency of aggressive interactions should decline as resource density increases, but empirical studies have provided only equivocal support for this prediction. We suggest that whether or not foragers have information concerning the location of patches will influence both the intensity of aggressive encounters and the effect that changes in food density will have on aggression. The intensity of aggression should be greatest when patch locations are known to all, making resources spatially predictable and the availability of alternatives more certain. When food is hidden, increasing the density of patches should have little effect on aggression levels, mostly as a result of the greater uncertainty about the availability of replacement food patches. To test these predictions, we investigated the effect of patch density on the use of aggressive behaviour in nutmeg mannikins, Lonchura punctulata, when food patches were either visible (signalled patch location) or hidden (unsignalled patch location) to all foragers. As predicted, we found that the intensity of aggressive encounters was higher when patch location was signalled than when it was not. Moreover, the effect of patch density on aggression depended on whether patch location was signalled or not. When patch location was unknown, the number of aggressive encounters was unaffected by changes in patch density, but when food location was signalled, increasing patch density resulted in the expected decline in the frequency of aggression.     

Opportunity costs influence food selection and giving‐up density of dabbling ducks

The interaction of animals with their food can yield insights into habitat characteristics, such as perceived predation risk and relative quality. We deployed experimental foraging patches in wetlands used by migrating dabbling ducks Anas spp. in the central Illinois River Valley to estimate variation in seed removal and giving‐up density (GUD; i.e. density of food remaining in patches following abandonment) with respect to seed density, seed size, seed depth in the substrate, substrate firmness, perceived predation risk, and an energetic profitability threshold (i.e. critical food density). Seed depth and the density of naturally‐occurring seeds outside of experimental plots affected seed removal and GUD in experimental patches more than perceived predation risk, seed density, seed size or substrate firmness. The greatest seed removal and lowest GUDs in experimental patches occurred when food resources in alternative foraging locations outside of plots (i.e. opportunity costs) appeared to be near or below a critical food density (i.e. 119–181 kg ha^–1). Giving‐up densities varied substantially from a critical food density across a range of food densities in alternative foraging locations suggesting that fixed GUDs should not be used as surrogates for critical food densities in energetic carrying capacity models. Foraging and resting rates in and near experimental foraging patches did not reflect patterns of seed removal and were poor predictors of GUD and foraging habitat quality. Our results demonstrated the usefulness of GUDs as indicators of habitat quality for subsurface, benthic foragers relative to other available foraging patches and suggested that food may be limited for dabbling ducks during spring migration in some years in the midwestern USA.     

Foraging costs and accessibility as determinants of giving-up densities in a swan-pondweed system

We measured the patch use behaviour of Bewick's swans (Cygnus columbianus bewickii) feeding on below ground tubers of fennel pondweed (Potamogeton pectinatus). We compared the swans’ attack rates, foraging costs and giving‐up densities (GUDs) in natural and experimental food patches that differed in water depth. Unlike most studies that attribute habitat‐specific differences in GUDs to predation risk, food quality or foraging substrate, we quantified the relative importance of energetic costs and accessibility. Accessibility is defined as the extent to which the animal's morphology restricts its harvest of all food items within a food patch. Patch use behaviours were measured at shallow (ca 0.4 m) and deep (ca 0.6 m) water depths on sandy sediments. In a laboratory foraging experiment, when harvesting food patches, the swan's attack rate (m³ s^?1) did not differ between depths. In deep water the energetic costs of surfacing, feeding and trampling were 1.13 to 1.21 times higher than in shallow water with a tendency to spend relatively more time trampling, the most expensive activity. Taking time allocation as measured in the field into account, foraging in deep water was 1.26 times as expensive as in shallow water. In the lake the GUD in shallow water was on average 12.9 g m^?2. If differences in energetic costs were the only factor determining differences in GUDs, then the deep water GUD should be 14.2 g m^?2. Instead, the mean GUD in deep water was 20.2 g m^?2, and therefore energetic costs explain just 18% of the difference in GUDs. At deep sites, 24% of tuber biomass was estimated to be out of reach, and we calculated a maximum accessible foraging depth of 0.86 m. This is close to the published 0.84 m based on body measurements. A laboratory experiment with food offered at a depth of 0.89 m confirmed that it was just out of reach. The agreement between calculated and observed maximum accessible foraging depths suggests that accessibility largely explains the remaining difference in GUDs with depth, and it confirms the existence of partial prey refuges in this system.     

Top-down effects of foraging decisions on local,landscape and regional biodiversity of resources (DivGUD)

Foraging by consumers acts as a biotic filtering mechanism for biodiversity at the trophic level of resources. Variation in foraging behaviour has cascading effects on abundance, diversity, and functional trait composition of the community of resource species. Here we propose diversity at giving-up density (DivGUD), i.e. when foragers quit exploiting a patch, as a novel concept and simple measure quantifying cascading effects at multiple spatial scales. In experimental landscapes with an assemblage of plant seeds, patch residency of wild rodents decreased local α-DivGUD (via elevated mortality of species with large seeds) and regional γ-DivGUD, while dissimilarity among patches in a landscape (ß-DivGUD) increased. By linking theories of adaptive foraging behaviour with community ecology, DivGUD allows to investigate cascading indirect predation effects, e.g. the ecology-of-fear framework, feedbacks between functional trait composition of resource species and consumer communities, and effects of inter-individual differences among foragers on the biodiversity of resource communities.     

Eavesdropping foragers use level of collective commotion as public information to target high quality patches

Public information offers a valuable means for social foragers to determine the relative quality of foraging patches. Despite much evidence that foragers use public information based on others’ feeding behavior, no experiments have examined whether foragers might use public information based on others’ competitive behavior, particularly the collective commotion that can be generated by aggregations. Such commotion could potentially provide a rich source of public information: as foragers compete in a patch with an especially high value resource, their heightened competition intensity could enable eavesdropping foragers to target this superior patch, based simply on its higher level of collective commotion. To test the hypothesis that the level of collective commotion is used as public information by eavesdropping foragers I conducted field experiments on terrestrial hermit crabs Coenobita compressus. These animals engage in collective competitive interactions in foraging patches for food and shells, generating variable levels of commotion across different quality patches. By experimentally manipulating the level of collective commotion in sham aggregations in the wild I show that a higher level of commotion is exploited by eavesdropping foragers to differentially target more valuable patches. Broadly, these results highlight an underappreciated significance of competitive by‐products and higher‐ order collective pheno mena as forms of public information for foragers.     

Patch selection by juvenile black surfperch (Embiotocidae) under variable risk: Interactive influence of food quality and structural complexity

The influence of risk on the selection of foraging patches by young-of-year black surfperch, Embiotoca jacksoni Agassiz, was investigated by laboratory and field experiments. These foragers harvest crustacean prey from a variety of benthic algal substrata. In field environments, patch types vary in two ways. First, substrata differ in structural complexity and probably afford different degrees of protection from predators. Second, substratum types vary in prey richness. There was no correlation between structural complexity and prey richness, and either or both factors could be a component of foraging patch value. Each patch is small and individual foragers are simultaneously confronted with arrays of patches encompassing the full range of variation in structure and prey richness. Furthermore, a major predator of young-of-year black surfperch, the kelp bass, Paralabrax clathratus (Girard), is patchily distributed in space and time. Thus similar arrays of patch types can be characterized by different levels of overall risk. Risk to foragers is dependent on light level as well as the presence and density of predators.The interplay between food quality and shelter potential in influencing patch choice was examined under different regimes of risk. Both laboratory and field experiments indicated patch preference was based primarily on food quality. However, the physical structure of a patch did become a component of patch choice as risk increased. The relative value of physical structure under high risk was dependent on the prey richness of a patch; food-poor substrata with high shelter potential remained unfavored even in situations of high risk.