Optimal Foraging Models Research Articles

Time's crooked arrow: optimal foraging and rate-biased time perception

Time perception is critical to animal behaviours requiring anticipation of future events based on present information about the environment. Most models of animal foraging assume that animals are capable of measuring absolute time despite evidence that animals measure time with predictable biases in mean and variance. We incorporate the evidence for a rate-biased subjective time perception into a classic model of optimal foraging, the marginal value theorem. If acceleration of the clock rate is proportional to food intake rate and time is perceived similarly when in transit between patches as it is while waiting in a patch following eating, organisms are predicted to follow the predictions of the marginal value theorem exactly. However, a nonlinear relationship between clock rate and food intake rate, unequal wait and transit time perception, or any lag in the clock predicts characteristic suboptimal behaviour. We discuss how this mechanism for suboptimal behaviour compares with others in the literature and how it can be recognized in experiments. Copyright 2002 The Association for the Study of Animal Behaviour. Published by Elsevier Science Ltd. All rights reserved.

Why do Foraging Stream Salmonids Move During Summer?

We hypothesize that foraging stream salmonids move during summer because (1) they monitor habitat conditions at a reach scale (100s of m), and (2) dominant fish move when conditions in their present foraging location become sub-optimal relative to conditions at other locations in the reach. To test these ideas, we quantified temporal variation in foraging habitat quality between late spring and early fall in a reach of a small Rocky Mountain brook charr, Salvelinus fontinalis, stream, predicted optimal-foraging fish distributions within the reach, and experimentally manipulated access to foraging sites and measured fish responses. Our results show that high-quality foraging sites were located at certain places in the reach during one period, but at different places during others, consistent with the hypothesis that fish movement is required if dominant fish are to occupy high-quality foraging sites throughout summer. The optimal foraging model was able to predict foraging locations within study pools, but not the exact location of individual fish within the pools or the reach. However, empirical evidence suggests that fish were distributed in order to maximize energy intake at the reach scale. Finally, dominant fish excluded from their preferred foraging location either left the pools (three of six cases), or began to occupy focal points of the next largest fish which, in turn, exited the pool (two of six cases). If habitat selection was occurring only within habitat units, then large fish, when excluded from their preferred locations, would select the next best locations within the pool. Taken together, these results suggest that charr use summertime movements to both monitor habitat conditions at a large spatial scale, and to gain access to optimal foraging locations even as conditions change temporally.

Prey spatial structure and behavior affect archaeological tests of optimal foraging models: Examples from the Emeryville Shellmound vertebrate fauna

Foraging theory is increasingly used as a framework to analyze prehistoric resource depression , or declines in prey capture rates that stem from the activities of foragers. Resource depression, with the resulting declines in foraging efficiency, appears to be an important variable underlying many behavioral transitions in human prehistory. And, while two of the primary lines of evidence used to document resource depression archaeologically are temporal declines in the relative abundances of 'high-return' species and the mean age of harvested prey, this phenomenon may also be reflected by increases in both high-ranked prey abundances and mean age. These divergent signals are linked to variation in the behavior and spatial distributions of different prey species and are illustrated with the Emeryville Shellmound faunal sequence. More detailed attention should be given to these variables in applications of foraging models to the archaeological record.

A new optimal foraging model predicts habitat use by drift‐feeding stream minnows

Abstract – There is substantial need for models that accurately predict habitat selection by fishes for purposes ranging from the elaboration of ecological theory to the preservation of biodiversity. We have developed a new and highly tractable optimal foraging model for drift‐feeding fishes that is based on the profitability of occupying varying focal‐point velocities in a stream. The basic model can be written as: Ix = (Ex * Px) = {(D * A * V) * [1/(1 + e(b + cV))]} − Sx, where: (1) Ix is the net energy intake at velocity x; (2) E is prey encounter rate; (3) P is prey capture success rate which can be modelled as 1/(1 + e(b + cV)) where b and c are fitting constants from the prey capture success curve; (4) D is the energy content of prey (J/m3) in the drift; (5) A is the visual reactive area of the fish; (6) V is velocity (cm/s); and (7) S is the cost of maintaining position (J/s). Given that D, A and S can be considered constant over the range of velocities occupied by these fishes, the model reduces to e(b + cV) = 1/(cV − 1) which we solved iteratively to yield an optimal focal‐point velocity for species in each sample. We tested the model by comparing its predictions to the mean focal‐point velocities (i.e. microhabitats) occupied by four species of drift‐feeding minnows in two sites in a stream in North Carolina, USA. The model successfully predicted focal‐point velocities occupied by these species (11 out of 14 cases) in three seasonal samples collected over 2 years at two sites. The unsuccessful predictions still were within 2 cm/s of the 95% confidence intervals of mean velocities occupied by fishes, whereas the overall mean deviation between optimal velocities and mean fish velocities was small (range = 0.9 and 3.3 cm/s for the warpaint shiner and the Tennessee shiner, respectively). Available focal‐point velocities ranged from 0–76 to 0–128 cm/s depending on site and season. Our findings represent one of the more rigorous field tests of an optimal foraging/habitat selection model for aquatic organisms because they encompass multiple species and years, and for one species, multiple sites. Because of the ease of parameterization of our model, it should be readily testable in a range of lotic habitats. If validated in other systems, the model should provide critical habitat information that will aid in the management of riverine systems and improve the performance of a variety of currently used management models (e.g. instream flow incremental methodology (IFIM) and total maximum daily load calculations (TMDL)).

PREDATOR FUNCTIONAL RESPONSES: DISCRIMINATING BETWEEN HANDLING AND DIGESTING PREY

We present a handy mechanistic functional response model that realistically incorporates handling (i.e., attacking and eating) and digesting prey. We briefly review current functional response theory and thereby demonstrate that such a model has been lacking so far. In our model, we treat digestion as a background process that does not prevent further foraging activities (i.e., searching and handling). Instead, we let the hunger level determine the probability that the predator searches for new prey. Additionally, our model takes into account time wasted through unsuccessful attacks. Since a main assumption of our model is that the predator's hunger is in a steady state, we term it the steady state satiation (SSS) equation. The SSS equation yields a new formula for the asymptotic maximum predation rate (i.e., asymptotic maximum number of prey eaten per unit time, for prey density approaching infinity). According to this formula, maximum predation rate is determined not by the sum of the time spent for handling and digesting prey, but solely by the larger of these two terms. As a consequence, predators can be categorized into two types: handling-limited predators (where maximum predation rate is limited by handling time) and digestion-limited predators (where maximum predation rate is limited by digestion time). We give examples of both predator types. Based on available data, we suggest that most predators are digestion limited. The SSS equation is a conceptual mechanistic model. Two possible applications of this model are that (1) it can be used to calculate the effects of changing predator or prey characteristics (e.g., defenses) on predation rate and (2) optimal foraging models based on the SSS equation are testable alternatives to other approaches. This may improve optimal foraging theory, since one of its major problems has been the lack of alternative models.

Size‐dependent attack rate and handling capacity: inter‐cohort competition in a zooplanktivorous fish

Two size‐dependent processes, metabolic requirements and foraging capacity, heavily influence the competitive ability of organisms. We studied the size‐dependent competitive ability of roach (Rutilus rutilus) in a laboratory experiment to determine the attack rate of roach as a function of roach and zooplankton sizes. The estimated size‐dependent attack rates, size‐dependent metabolic demands and handling capacities were subsequently used to interpret the outcome of a competition experiment between two size classes of roach. Furthermore, size‐dependent attack rates were implemented in an optimal foraging model to predict consumption rates and zooplankton selection to reveal the mechanisms behind competition. The attack rate first increased with roach size to a maximum around 160 mm to thereafter decrease. Based on this result, we predicted that, small (150 mm) roach had a double advantage in competition for zooplankton in the pond experiment due to their higher attack rate and their lower metabolism compared to large (230 mm) roach. As expected, small roach depressed total zooplankton biomass to a higher extent than large roach, included more zooplankton in their diet and consumed smaller zooplankton. Predicted smallest size class of zooplankton in the diet was close to the observed. Also as expected, large roach fed more on the benthic resource and depressed the benthic resource to a larger extent than small roach. Large roach affected large zooplankton to a higher extent during the first part of the experiment, which could be related to their overall higher handling capacity. The higher impact of large roach on large zooplankton during the first part of the experiment, in turn, resulted in a lower estimated mass intake of zooplankton by small roach in the mixed treatment compared to small roach only treatments. Both small and large roach had a lower growth rate in the mixed treatment compared to single size class treatments. We relate the lower growth rate of small roach in the mixed treatment to large roach's higher efficiency on benthic resources and on large zooplankton during the first part of the experiment. In correspondence with diet data, large roach preferred the shore area of the pond with more benthic invertebrates and were also found more often close to the bottom. Although our results are explainable by exploitative interactions, the fact that the presence of large roach affected the feeding position of small roach points to that social interactions were also involved. Overall, our study implies that a mechanistic understanding is crucial for the interpretation of competition experiments, especially in systems with size‐structured interactions.

The Effect of Resource Aggregation at Different Scales: Optimal Foraging Behavior ofCotesia rubecula

Resources can be aggregated both within and between patches. In this article, we examine how aggregation at these different scales influences the behavior and performance of foragers. We developed an optimal foraging model of the foraging behavior of the parasitoid wasp Cotesia rubecula parasitizing the larvae of the cabbage butterfly Pieris rapae. The optimal behavior was found using stochastic dynamic programming. The most interesting and novel result is that the effect of resource aggregation within and between patches depends on the degree of aggregation both within and between patches as well as on the local host density in the occupied patch, but lifetime reproductive success depends only on aggregation within patches. Our findings have profound implications for the way in which we measure heterogeneity at different scales and model the response of organisms to spatial heterogeneity.

Foraging patterns of juvenile walleye (Stizostedion vitreum) in a system consisting of a single predator and two prey species: testing model predictions

We tested the hypothesis that the presence of a large number of energetically inferior prey (Daphnia pulex) in the environment with energetically superior prey (larval carp, Cyprinus carpio) interferes with foraging efficiency of juvenile walleye (Stizostedion vitreum). We monitored functional responses of juvenile walleye feeding on larval carp alone or in combination with D. pulex. When walleye were offered larval carp at increasing densities (10, 20, 50, and 100 individuals/30-L aquarium), they responded in accordance with a type-II functional response in 10-min trials. Walleye captured the maximum number of larval carp when offered a carp density of 20 individuals/aquarium without the daphnids. Further increase in density of larval carp had no effect on walleye foraging rate. The presence of D. pulex (900 individuals/aquarium) suppressed the efficiency of foraging walleye. The predator consistently captured fewer larval carp in all treatments when they were offered together with the daphnids even though walleye continued to select the more profitable larval carp. The negative effect of daphnids on the feeding rate is an example of environmental constraints interfering with optimal foraging models. In this study, walleye appeared to experience a confusion effect caused by the large number of inferior prey. Consequently, the predator could not gain as much energy per unit time when confronted with two prey types as it was when foraging on larval carp alone.

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.

Optimal patch‐leaving behaviour: a case study using the parasitoid Cotesia rubecula

Summary Parasitoids are predicted to spend longer in patches with more hosts, but previous work on Cotesia rubecula (Marshall) has not upheld this prediction. Tests of theoretical predictions may be affected by the definition of patch leaving behaviour, which is often ambiguous. In this study whole plants were considered as patches and assumed that wasps move within patches by means of walking or flying. Within‐patch and between‐patch flights were distinguished based on flight distance. The quality of this classification was tested statistically by examination of log‐survivor curves of flight times. Wasps remained longer in patches with higher host densities, which is consistent with predictions of the marginal value theorem (Charnov 1976). Under the assumption that each flight indicates a patch departure, there is no relationship between host density and leaving tendency. Oviposition influences the patch leaving behaviour of wasps in a count down fashion (Driessen et al. 1995), as predicted by an optimal foraging model (Tenhumberg, Keller & Possingham 2001). Wasps spend significantly longer in the first patch encountered following release, resulting in an increased rate of superparasitism.

HERBIVORE PHYSIOLOGICAL STATE AFFECTS FORAGING TRADE-OFF DECISIONS BETWEEN NUTRIENT INTAKE AND PARASITE AVOIDANCE

The trade-offs between nutrient and parasite intake for herbivores foraging in nutrient-poor systems are poorly understood. We tested whether a trade-off exists between the benefit of increased nutrient intake rate through grazing relatively tall swards and risks of parasitism in a grazing situation for sheep. The effect of level of feeding motivation and parasitic status on the grazing behavior of sheep faced with this trade-off was also investigated. Animals were presented with pairs of experimental swards (36 X 21 cm) that varied in height (12 cm = H+; 6 cm = H-) and level of contamination with feces from sheep infected with Ostertagia circumcincta (20 g feces per sward = FS; no feces = F-) and were allowed to graze for short periods. Experiment 1 presented four choices (H+F+ vs. H-F-; H+F+ vs. H-FS; H+F- vs. H-F-; H+F- vs. H-F+) repeated three times to 24 5-mo-old sheep divided into four animal treatment groups resulting from two levels of feeding motivation (high and moderate) and two parasitic states (parasitized and nonparasitized). Experiment 2 presented the above four choices three times each to 24 12-mo-old sheep in four animal treatment groups created from two parasitic states (parasitized and nonparasitized) and two immune states (immune and naive to O. circumcincta). All animals selected 12-cm swards over 6-cm swards in both experiments. In all choices except the trade-off choice, in both experiments, one award held a clear height or feces benefit (absence of feces) that was selected for by all treatments. When presented with the trade-off choice (H+F+ vs. H-F-), animals with a high feeding motivation selected the H+F+ sward more than moderately motivated animals did in Experiment 1. Immune animals selected the H+F+ sward more than naive animals did in Experiment 2. Subclinical parasitism in moderately feeding-motivated animals resulted in increased avoidance of feces and reduced grazing depths, thereby reducing further risk of parasitism in Experiment 1. This effect of parasitism was not repeated in the older animals of Experiment 2, where all animals maintained a high-parasite-risk grazing strategy by selecting the H+F+ award of the trade-off choice. However, parasitized animals in Experiment 2 reduced parasite intake and, therefore, the risks associated with grazing the H-I-FS sward by significantly reducing their grazing depth compared to all other treatment groups. Sheep grazing decisions involving forage intake and feces avoidance were affected by their feeding motivation, immune state, and parasitic state. Strong avoidance of feces-contaminated pasture can be overcome by the attraction of herbivores to tall swards. Trade-off frameworks are able to predict diet selection and grazing behavior, are useful when investigating host-parasite interactions, and may enhance the predictive powers of optimal foraging models.

Herbivore Physiological State Affects Foraging Trade-Off Decisions between Nutrient Intake and Parasite Avoidance

The trade-offs between nutrient and parasite intake for herbivores foraging in nutrient-poor systems are poorly understood. We tested whether a trade-off exists between the benefit of increased nutrient intake rate through grazing relatively tall swards and risks of parasitism in a grazing situation for sheep. The effect of level of feeding motivation and parasitic status on the grazing behavior of sheep faced with this trade-off was also investigated. Animals were presented with pairs of experimental swards (36 × 21 cm) that varied in height (12 cm = H+; 6 cm = H−) and level of contamination with feces from sheep infected with Ostertagia circumcincta (20 g feces per sward = F+; no feces = F−) and were allowed to graze for short periods. Experiment 1 presented four choices (H+F+ vs. H−F−; H+F+ vs. H−F+; H+F− vs. H−F−; H+F− vs. H−F+) repeated three times to 24 5-mo-old sheep divided into four animal treatment groups resulting from two levels of feeding motivation (high and moderate) and two parasitic states (parasitized and nonparasitized). Experiment 2 presented the above four choices three times each to 24 12-mo-old sheep in four animal treatment groups created from two parasitic states (parasitized and nonparasitized) and two immune states (immune and naive to O. circumcincta). All animals selected 12-cm swards over 6-cm swards in both experiments. In all choices except the trade-off choice, in both experiments, one sward held a clear height or feces benefit (absence of feces) that was selected for by all treatments. When presented with the trade-off choice (H+F+ vs. H−F−), animals with a high feeding motivation selected the H+F+ sward more than moderately motivated animals did in Experiment 1. Immune animals selected the H+F+ sward more than naive animals did in Experiment 2. Subclinical parasitism in moderately feeding-motivated animals resulted in increased avoidance of feces and reduced grazing depths, thereby reducing further risk of parasitism in Experiment 1. This effect of parasitism was not repeated in the older animals of Experiment 2, where all animals maintained a high-parasite-risk grazing strategy by selecting the H+F+ sward of the trade-off choice. However, parasitized animals in Experiment 2 reduced parasite intake and, therefore, the risks associated with grazing the H+F+ sward by significantly reducing their grazing depth compared to all other treatment groups. Sheep grazing decisions involving forage intake and feces avoidance were affected by their feeding motivation, immune state, and parasitic state. Strong avoidance of feces-contaminated pasture can be overcome by the attraction of herbivores to tall swards. Trade-off frameworks are able to predict diet selection and grazing behavior, are useful when investigating host–parasite interactions, and may enhance the predictive powers of optimal foraging models.

Group choice: the ideal free distribution of human social behavior.

Group choice refers to the distribution of group members between two choice alternatives over time. The ideal free distribution (IFD), an optimal foraging model from behavioral ecology, predicts that the ratio of foragers at two resource sites should equal the ratio of obtained resources, a prediction that is formally analogous to the matching law of individual choice, except that group choice is a social phenomenon. Two experiments investigated the usefulness of IFD analyses of human group choice and individual-based explanations that might account for the group-level events. Instead of nonhuman animals foraging at two sites for resources, a group of humans chose blue and red cards to receive points that could earn cash prizes. The groups chose blue and red cards in ratios in positive relation to the ratios of points associated with the cards. When group choice ratios and point ratios were plotted on logarithmic coordinates and fitted with regression lines, the slopes (i.e., sensitivity measures) approached 1.0 but tended to fall short of it (i.e., undermatching), with little bias and little unaccounted for variance. These experiments demonstrate that an IFD analysis of group choice is possible and useful, and suggest that group choice may be explained by the individual members' tendency to optimize reinforcement.

Foraging behaviour of red deer Cervus elaphus as a function of the relative availability of two tree species

Cervid populations are rapidly increasing in many part of Western Europe, where they cause damage to forest production. This necessitates a better understanding of what determines their dietary choices. In this experiment, we investigated the relationship between the relative availability of two tree species and diet selection of red deer Cervus elaphus. Three hypotheses were tested: (i) preference for rarity, frequently asserted by foresters, (ii) preference for the more abundant prof- itable species, as predicted by optimal foraging models, and (iii) frequency-independent selectiv- ity, as already observed in moose and roe deer. Six red deer hinds were therefore observed in short- duration tests when offered two plant species, willow Salix caprea and ash Fraxinus excelsior, in varying proportions. The relation between the relative consumption and availability of the two plant species revealed that red deer prefer ash to willow, and that their short-term selectivity is not a fre- quency-dependent process. Individual animals differed in the intensity of their preference, but all reacted similarly to variations in the relative availability of the two plant species. Our results advance our under- standing of the choices of red deer at the feeding site scale, but further research is needed to propose management practices, which would take advantage of their frequency-independent selectivity.

The Effect of Resource Aggregation at Different Scales: Optimal Foraging Behavior of Cotesia rubecula

Resources can be aggregated both within and between patches. In this article, we examine how aggregation at these different scales influences the behavior and performance of foragers. We developed an optimal foraging model of the foraging behavior of the parasitoid wasp Cotesia rubecula parasitizing the larvae of the cabbage butterfly Pieris rapae. The optimal behavior was found using stochastic dynamic programming. The most interesting and novel result is that the effect of resource aggregation within and between patches depends on the degree of aggregation both within and between patches as well as on the local host density in the occupied patch, but lifetime reproductive success depends only on aggregation within patches. Our findings have profound implications for the way in which we measure heterogeneity at different scales and model the response of organisms to spatial heterogeneity.

A Mechanistic Model for Partial Preferences

Classic prey optimal foraging model assumes that individual predators are globally omniscient; that is, they have exact knowledge of prey population densities in the environment. This study examines a spatially explicit individual-based model of a one-predator two-prey system where individual predators are assumed to be omniscient only locally, i.e., to know prey population densities only in the range of their perception. Due to local variations in prey numbers, the probability of acceptance of less profitable prey shifts from the zero–one rule to a gradually decreasing function, for which an explicit formula is derived, giving way to partial preferences. A corresponding predator functional response to more profitable prey is shown to have a sigmoid-like form.

Simulation of the Effect of Pollinator Movement on Alfalfa Seed Set

Using alfalfa grown for seed, we tested the hypothesis that seed set for a fixed number of pollinations is lower when the standing crop of open flowers is high than when it is low. This could occur if pollinators are more likely to move between flowers on the same plant, causing self-pollination, when flowers are abundant. The hypothesis has practical implications for agricultural production because self-pollinations produce fewer seeds per pod than do cross-pollinations in alfalfa. We simulated seed set in a model that included published movement patterns of the alfalfa leafcutting bee, Megachile rotundata (F.) (Hymenoptera: Megachilidae), and estimates of seed set when a series of flowers on the same plant are hand pollinated. Seed set from a fixed number of pollinations averaged only ≈;6–7% higher when standing crop was low than when standing crop was high. This small increase in seed set would be difficult to detect experimentally because of high variability between plants. M. rotundata appears to move with a 56% probability of leaving a given raceme and a given plant; this foraging behavior results in few flowers visited per plant. A pollinator with a higher probability of leaving racemes and plants could set slightly more seed than M. rotundata, whereas a pollinator with a lower probability of leaving racemes and plants would be more affected by flower standing crop than is M. rotundata. The distribution of flowers visited per raceme by bees in alfalfa can be predicted by superimposing the movement probability on the distribution of open flowers per raceme. This conceptualization of bee movement on flowers differs from (but is not mutually exclusive of) resource-driven optimal foraging models, and is useful for predicting geitonogamy.

The Dolphin Hunters: A Specialized Prehistoric Maritime Adaptation in the Southern California Channel Islands and Baja California

Synthesis of faunal collections from several archaeological sites on the three southernmost California Channel Islands and one in the Cape Region of Baja California reveals a distinctive maritime adaptation more heavily reliant on the capture of pelagic dolphins than on near-shore pinnipeds. Previous reports from other Southern California coastal sites suggest that dolphin hunting may have occurred there but to a lesser extent. While these findings may represent localized adaptations to special conditions on these islands and the Cape Region, they call for reassessment of the conventionally held concept that pinnipeds were invariably the primary mammalian food resource for coastal peoples. Evidence of the intensive use of small cetaceans is antithetical to the accepted models of maritime optimal foraging which assume that shore-based or near-shore marine mammals (i.e., pinnipeds) would be the highest-ranked prey because they were readily encountered and captured. While methods of dolphin hunting remain archaeologically invisible, several island cultures in which dolphin were intensively exploited by people using primitive watercraft and little or no weaponry are presented as possible analogs to a prehistoric Southern California dolphin-hunting technique. These findings also indicate that dolphin hunting was probably a cooperative endeavor among various members of the prehistoric community.

A state‐dependent model of activity patterns in homing limpets: balancing energy returns and mortality risks under constraints on digestion

Summary 1. Different species of limpets, and different populations of the same species, show an impressive variation in temporal organization of foraging activity. Field studies also reveal different patterns of activity among individuals of the same population. 2. Many hypotheses have been proposed to explain the observed behavioural patterns, in terms of effects of several biotic and abiotic factors acting in different tidal phases, yet without taking into account possible interactions among environmental and internal factors. Moreover, the importance of energy reserves and food processing has rarely been considered in this context. 3. The role of different environmental factors and functional traits in shaping the organization of foraging of limpets at different time scales was analysed by using a state‐dependent optimal foraging model, based on stochastic dynamic programming. The model includes interaction among mortality risks and energy costs incurred during the different tidal phases, the effect of different levels of energy reserves, gut fullness and changing food availability. 4. The model is able to reproduce a variety of foraging patterns with reference to the tidal cycle when changing combinations of values assigned to the risk of mortality, energy costs and food availability. The model also shows that the level of energy reserves, gut volume and rate of food processing can be of major importance in determining the short‐term organization of activity within each foraging phase.

Foraging rate versus sociality in the starling Sturnus vulgaris.

It is well established that social conditions often modify foraging behaviour, but the theoretical interpretation of the changes produced is not straightforward. Changes may be due to alterations of the foraging currency (the mathematical expression that behaviour maximizes) and/or of the available resources. An example of the latter is when both solitary and social foragers maximize rates of gain over time, but competition alters the behaviour required to achieve this, as assumed by ideal free distribution models. Here we examine this problem using captive starlings Sturnus vulgaris. Subjects had access to two depleting patches that replenished whenever the alternative patch was visited. The theoretical rate-maximizing policy was the same across all treatments, and consisted of alternating between patches following a pattern that could be predicted using the marginal value theorem (MVT). There were three treatments that differed in the contents of an aviary adjacent to one of the two patches (called the 'social' patch). In the control treatment, the aviary was empty, in the social condition it contained a group of starlings, and in a non-specific stimulus control it contained a group of zebra finches. In the control condition both patches were used equally and behaviour was well predicted by the MVT. In the social condition, starlings foraged more slowly in the social than in the solitary patch. Further, foraging in the solitary patch was faster and in the social patch slower in the social condition than in the control condition. Although these changes are incompatible with overall rate maximization (gain rate decreased by about 24% by self-imposed changes), if the self-generated gain functions were used the MVT was a good predictor of patch exploitation under all conditions. We discuss the complexities of nesting optimal foraging models in more comprehensive theoretical accounts of behaviour integrating functional and mechanistic perspectives.