Abstract: Abstract. The relation between memory for travel time and foraging decisions was studied experimentally. The temporal properties of two environments with patchily distributed food were simulated in the laboratory using pigeons, Columba livia, as subjects. The two environments differed in mean travel time, while the coefficient of variation of travel time and the decelerated function relating cumulative food gain to time in the patch were held constant within and between environments. Each environment contained a uniform mixture of five travel times experienced in a random order. Two of the five travel times were common in both environments. Effects of travel time were studied by comparing prey collected per patch visit (PPV) after various travel times within each environment, and by comparing patch exploitation after equal travel times between environments. Within the environment with long mean travel time (LMT) PPV was positively correlated with the last and the penultimate travel times but not with travel times before that. The increase in PPV per second of last travel time was six times greater than the increase per second of penultimate travel time, implying very steep memory discounting. In the environment with short mean travel time (SMT), there was no correlation between PPV and previous travel times. However, comparisons between environments of visits following travel times common to both environments (thus removing the effect of the last travel time) showed that substantially more prey were taken after equal travel times in the LMT than in the SMT environment. This difference cannot be accounted for by the within-environment effect of penultimate travel time, implying that there is a different, less steeply devalued, effect of the mixture of travel times. A model of information processing based on combining Scalar Expectancy Theory with the predictions of rate maximization under the Marginal Value Theorem is presented. The model can approximate the results obtained in this and previous experiments and provides a framework for further analysis of memory mechanisms of foraging behaviour.

Abstract: Laboratory and field experiments have shown that, as predicted by the marginal value model, starlings, Sturnus vulgaris, stay longer in a food patch when the average travel time between patches is long. A laboratory analogue of a patchy environment was used to investigate how starlings respond to rapidly fluctuating changes in travel time in order to find out the length of experience over which information is integrated. When there was a progressive increase in the amount of work required to obtain successive food items in a patch (experiment 1), birds consistently took more prey after long than after short travel times; travel experience before the most recent had no effect on the number of prey taken. Such behaviour does not maximize the rate of energy intake in this environment. The possibility that this is the result of a simple constraint on crop capacity is rejected as, when successive prey were equally easy to obtain up until a stepwise depletion of the patch (experiment 2), birds took equal numbers of prey per visit after long and short travel times: the rate-maximizing behaviour. A series of models are developed to suggest the possible constraints on optimal behaviour that affect starlings in the type of environment mimicked by experiment 1.

Abstract: Recent studies have used labels such as `work ethics', `sunk costs' and `state-dependent preferences' for apparent anomalies in animals' choices. They suggest that preference between options relates to the options' history, rather than depending exclusively on the expected payoffs. For instance, European starlings, Sturnus vulgaris, trained to obtain identical food rewards from two sources while in two levels of hunger preferred the food source previously associated with higher hunger, regardless of the birds' state at the time of testing. We extended this experimentally and theoretically by studying starlings choosing between sources that differed not only in history but also in the objective properties (delay until reward) of the payoffs they delivered. Two options (PF and H) were initially presented in single-option sessions when subjects were, respectively, prefed or hungry. While option PF offered a delay until reward of 10 s in all treatments, option H delivered delays of 10, 12.5, 15 and 17.5 s in four treatments. When training was completed, we tested preference between the options. When delays in both options were equal (10 s), the birds strongly preferred H. When delay in H was 17.5 s, the birds were indifferent, with intermediate results for intermediate treatments. Preference was not mediated by disrupted knowledge of the delays. Thus, preferences were driven by past state-dependent gains, rather than by the joint effect of the birds' state at the time of choice and knowledge of the absolute properties of each alternative, as assumed in state-dependent, path-independent models of optimal choice.

Abstract: In nature foragers must exploit resources that vary randomly in both the energy acquired per item (reward) and the time required to pursue, capture and process an item (delay). Furthermore, rewards and delays associated with particular resources may often covary significantly. An analytical model asks how variance-covariance levels for rewards and delays could influence choice of resources when lack of information or cognitive limitation implies that a consumer attempts to maximize its short-term rate of energy gain. Both greater expected reward and reduced expected delay clearly should enhance preference for a resource. The model predicts that increased delay variance and reduced reward-delay covariance should increase a forager's preference for a resource. A forager should be risk-averse towards reward variance when the reward-delay covariance is positive, but should become risk-prone towards reward variance when the reward-delay covariance is negative.