Fig. 6: Collection rates.

Point plots show the ICRs for the seven last reward captures averaged for the long-GUT (a) and short-GUT humans (b). Error bars equal ±standard error. The gray dashed line marks the averaged overall collection rate given by the number of total rewards divided by the total search time in seconds (the shaded area indicates the 95% confidence interval). The estimated capture rate at the time of leaving was defined as the inverse of the participant’s GUT with short GUTs leading to higher estimated ICRs at the time of leaving compared to long GUTs. Point plots show ICRs for the long-GUT (c)) and short-GUT (d)) gerbils of the last three reward captures. Note that only the last three reward captures are plotted here because this was the average number of rewards obtained. In contrast to the MVT prediction, ICRs in gerbils were well above the MCR (red dashed line) at the time of leaving (i.e., the ICR at which the last reward was captured). This estimate of collection rate in the moment of patch-leaving (=1/GUT) is independent of actual reward encounters, relying solely on the participants’ propensity to continue foraging unsuccessfully. A lower GUT implies a higher estimated collection rate at patch-leaving, potentially even leading to an increase of the rate compared to the collection rate at which the preceding rewards had been captured, as seen in Fig. 6b, where the estimated CR at patch-leaving was higher than the CR of the last and second last reward (i.e., ‘−1’ on the x-axis). It is essential to note that this estimate does not originate from an actual reward encounter and, thus, does not violate the assumption of a depleting patch.