Emergent regulation of ant foraging frequency through a computationally inexpensive forager movement rule

  1. Lior Baltiansky
  2. Guy Frankel
  3. Ofer Feinerman  Is a corresponding author
  1. Department of Physics of Complex Systems, Weizmann Institute of Science, Israel
8 figures, 2 tables and 1 additional file

Figures

Example of a forager’s trajectory in the experimental nest.

A single frame from an experimental video shows the nest on the right and a foraging arena on the left, where a fluorescent food source was presented. The food source is marked as a red oval and the …

Figure 2 with 2 supplements
Empirical movement of foragers in the nest.

(A) All locations in the nest were binned according to distance from the entrance, with bin width of 1 typical ant length (as visualized by circular grid lines in panel C). At each interaction of a …

Figure 2—source data 1

Empirical data.

Data used for calculation of the foragers’ empirical crop-dependent bias. All foragers’ interactions are pooled from the 3 experiments presented in Greenwald et al., 2018. Each interaction entry includes information on its location in the nest, the direction of the next interaction of the forager, and the forager’s crop load.

https://cdn.elifesciences.org/articles/77659/elife-77659-fig2-data1-v1.csv
Figure 2—video 1
Forager 421’s 12th unloading bout, when the colony was 90% full.
Figure 2—video 2
Forager 421’s 4th unloading bout, when the colony was 20% full.
Examples of two unloading bouts of a forager in the 2D simulation.

The 89 nest-ants are depicted by colored circles at their positions at the beginning of the bout. The color represents the ant’s crop state (purple represents an empty crop, yellow represents a full …

Foraging frequency scales linearly with empty colony state.

Foraging frequency was calculated as the inverse of the duration of the forager’s unloading bout in the nest. Unloading bouts were binned into five equally spaced bins of colony state, and the mean …

Figure 4—source data 1

Data from 1D model.

Output data from 200 runs of the 1D agent-based model. The file contains 3 spreadsheets: (1) Forager data. Includes data on the forager’s crop load and position in the nest at every step of the simulation. (2) Trophallaxis data. Includes data on the forager’s and the receiver’s crop loads, and the amount of food transferred at every interaction. (3) Trip data. Aggregated data on each trip of the forager inside the nest, including trip length and forager’s crop load upon exiting.

https://cdn.elifesciences.org/articles/77659/elife-77659-fig4-data1-v1.xlsx
Figure 4—source data 2

Data from 2D model.

Output data from 200 runs of the 2D agent-based model. Data within the file is as described for the 1D model data. Python code for the agent-based model is available on GitHub (Frankel et al., 2022).

https://cdn.elifesciences.org/articles/77659/elife-77659-fig4-data2-v1.xlsx
Comparison between empirical and simulation data of forager unloading bout dynamics.

Unloading bouts were binned into equally-spaced bins of colony state. Means and SEMs of different measures were calculated for each bin. Plots of the empirical data were taken from Figures 3D and 2B …

Schematic detailing the process of a forager at every step of the simulation.

The forager first moves according to a crop-dependent movement rule. Then either feeds, if at the entrance, or attempts to interact with another agent.

Appendix 1—figure 1
Empirical distributions of the angles between the foragers’ direction of movement and the direction to the nest entrance (an angle of 0o represents a direct movement toward the entrance).

Black arrow represents the mean of the distribution. Data was sampled at the end of each interaction of a forager. The angles are presented in two distributions, one where the foragers’ crop load …

Appendix 1—figure 2
Distributions of the Kendall τ rank correlation coefficients for nest-ant movement in 3 experiments compared to those of fully random shuffling.

Tables

Table 1
Movement biases for agent-based model.

The probabilities of a simulated forager to step inward, outward or to stay in the same cell, for two cases: when her crop load is lower than or higher than a threshold (0.45). The values of the …

Crop loadP(inward)P(outward)P(stay)
≤ 0.450.160.530.31
> 0.450.460.320.22
Table 2
Parameter values for different groups of agents in both models.

Parameters given to all agents are described under the ’Ants’ sub-population.

Sub-populationParameterModelValue
Ants
Crop state capacityAll1
Movement speed2D0.2 ant-lengths second-1
Nest-ants
Initial crop stateAll0
Position
1D45 ants, one on every cell
2D89 ants randomly placed
Radius of interaction2D0.2 ant-lengths
Forager
Initial crop stateAll1
Threshold valueAll0.45
Initial positionAllEntrance of nest
Foraging timeAll0
Interaction proportionAllpExp(10.15)
Biases in state A1D{0.32,0.22,0.46}
Biases in state B1D{0.53,0.31,0.16}
Possible angles in state A2DAppendix 1—figure 1, above
Possible angles in state B2DAppendix 1—figure 1, below
Boarder reflection noise2D[–0.3 radians, 0.3 radians]

Additional files

Download links