Skip to main content

Data from: A stabilizing eco-evolutionary feedback loop in the wild

Cite this dataset

Zamorano, Laura et al. (2023). Data from: A stabilizing eco-evolutionary feedback loop in the wild [Dataset]. Dryad.


Eco-evolutionary dynamics are predicted to involve feedback loops, where ecological changes driven by rapid evolution feed back to shape further evolution. However, direct experimental evidence for such loops in the wild is lacking. Here, we show that eco-evolutionary dynamics in a plant-feeding arthropod community involve a negative feedback loop. Specifically, adaptation in cryptic coloration in an abundant stick-insect species mediates bird predation, with local maladaptation increasing predation. In turn, the abundance of arthropods is reduced by predation. Here, we experimentally manipulate arthropod abundance to show that these community-level changes feedback to affect stick-insect evolution. Specifically, low arthropod abundance increases the strength of selection on crypsis, increasing local adaptation of stick insects in a negative feedback loop. Our results suggest that eco-evolutionary feedbacks are able to stabilize complex systems by preventing consistent directional change and therefore increasing resilience.


We conducted an experiment to test whether variation in the arthropod community affects selection on Timema cristinae cryptic morphs. In this context, we manipulated the abundance of arthropods into low- and high-abundance treatments on individual host bushes of Adenostoma fasciculatum in nature. We then transplanted T. cristinae to the same bushes in a 50% ratio of striped and unstriped individuals. We allowed the experiment to run for three days in natural conditions after which we resampled each plant to determine the total number of T. cristinae individuals recaptured and changes in the proportion of each morph.

We used a replicated, randomized block design with two treatment levels: low-abundance (200 arthropod individuals) and high-abundance (600 arthropod individuals). Each of the two treatments was replicated five times within the study site. Experimental bushes were chosen to be approximately the same size and have a similar gross light exposure. The area below and surrounding the experimental bush was cleared out of other vegetation in order to minimize dispersal of arthropods to or from adjacent plants. Thus, we created experimental islandssuch that any change in density or morph frequency of T. cristinae was assumed to be from mortality. Our experimental design implements a complete randomized block design which allowed us to account for possible sources of error intrinsic to field experimentation (e.g., differences in light exposure, micro-climate etc). Such variations might occur but should be random with respect to treatment such that they cannot readily explain treatment effects.

On March 15, 2022, we transplanted 40 T. cristinae individuals to each treatment in each block. Because we were interested in whether arthropod abundance has an effect on color-pattern morph frequency in T. cristinae, we used a 50% ratio of striped and unstriped individuals, thus maximizing the potential for evolution and selection. In parallel, the arthropods corresponding to each treatment were transplanted to the experimental bushes. To ensure that the arthropods were transferred to the bushes, we placed the open plastic containers in the middle of the plant and let the arthropods move from the plastic container to the plant undisturbed.

We let the experiment run for three days after which we recaptured the surviving individuals using sweep nets and examining each experimental bush. Each branch of the experimental bush was systematically sampled for T. cristinae individuals. Sampling ceased when no new individuals were collected after at least several minutes of extensive searching. In addition, the total number of T. cristinae recaptured on each experimental bush and the color-pattern of each individual were recorded.

Usage notes

The data are in a plain text file.


European Research Council, Award: 770826 EE-Dynamics