Experimental system and nest architecture.

(A) We used large nest frames 80 ∗ 60 ∗ 0.8 cm3 to allow for long-term excavation. Nest frames are filled with homogenous fine sand and wet regularly for structural cohesiveness. Water infiltrates through the water column and keeps the sand moist and cohesive. (B and C) A typical nest at the start and termination of an experiment.

Excavation and population dynamics during colony maturation.

(A) & (B) Temporal change of the number of ants (A), and excavated area (B) starting from a single queen. Solid red, and grey circles show the mean number of ants and excavated area across the 22 colony maturation experiment. Error bars depicted in black represent the standard error (here, and in all other figures). In (A), we fit a population growth model (white dotted line) using the number of ants as averaged over all colonies (Materials, and methods). (C) The average excavated area is plotted against the corresponding population size. The solid red line depicts a linear fit y = 10.71x, R2 = 0.97. The color bar represents the time in days. (D) The normalised cross-correlation coefficient between the digging rate, and population growth rate is shown. The digging rates are correlated with earlier (negative lags) population change with maxima at lag ‘−10’, rather than with the current (zero lag) or future (positive lags). Solid squares correspond to the mean cross-correlation coefficient. The vertical green line highlights the lag time with the maximum mean cross-correlation coefficient.

Details of colony maturation experiments.

Details of fixed demographics experiments.

Excavation dynamics for similar-age groups.

(A & B) The area excavated by the fixed demographics young and old colonies with a colony size of 5,10, and 15 ants. Solid stars and circles indicate the mean excavated area on the day of collapse. Circular crosses indicate the mean area recovered after the manual collapse. (C) The mean area excavated by colonies of different group sizes from the colony maturation and fixed demographics experiments are shown. Squares (colony maturation), stars (fixed demographics young), and circles (fixed demographics old) correspond to the mean area excavated by colonies of a particular group size. Shaded error bars and error bars correspond to the SEM.

Age dependent nest excavation.

(A) The target area per ant is represented as a function of the age of individual ants (dotted line). Target areas decrease linearly with ant age (y = −0.032x + 11.2, R2 = 0.96). The fixed target area (11.6cm2ant−1) calculated from the colony maturation experiment is shown with a solid blue line. (B) The mean excavated area as predicted by the age-dependent digging model is shown alongside the excavated area from the colony maturation experiments. Error bars in (A,& B) represent the SEM. (C) Total excavated area during normal colony growth as broken down by young ants, old ants, and the queen as predicted by the age-dependent model. The age threshold for young ants was set to be 56 days. Solid circles represent the area excavated by ants across 22 experiments. Boxes and median lines represent inter-quartile range and median values, and whiskers represent minimum and maximum values of data within 1.5-fold of the inter-quartile range (D) The area excavated by each worker ant during normal colony growth as estimated by the age-dependent model.

Colony demographics from nest population.

Structural classification of nests.

(A, B, & C). Panels represent the nests excavated from the colony maturation (A), fixed demographics young (B), and old (C) nests in terms of the tunnel, wide tunnel, and chambers for a colony size 15. (D, E, & F) The contribution of tunnels and chambers to the total excavated area for a fixed group size across all the experimental conditions. Error bars represent the SEM. (G, H,&I) Angular orientation of tunnels and chambers across conditions.

Individual experiments - Area, Population.

The change of area and population across individual colony maturation experiments (n = 22) are provided. The left y-axis shows the excavated area and the right y-axis shows the population.

Temporal change of the area per ant.

The temporal change of the area per ant from the colony maturation experiments. The area per ant is calculated by dividing the excavated area by the number of ants in the experiment. In Fig. 2 A, day ‘0’ corresponds to the start of the experiment, whereas in 2 B, day ‘0’ corresponds to the day on which the first workers were born. The area per ant stabilizes at 11.1(±1)for Fig. 2 A, and 11.6(±1.15) for Fig. 2 B. Shaded error bar in both figures shows the SEM.

Step-like relationship between excavated area and population growth.

A. The excavated area is plotted against their corresponding population sizes across specific colony maturation experiments where the step pattern was most evident. We can see a “step” like the relationship between the variables.

(B, C, D, E). Shows the change in area and population across experimental duration across the four specific nests. Peaks correspond to the increase in area or population.

Age distribution from the young and old colonies.

The age distribution of workers used for the fixed demographics young and old experiments is shown as a histogram. The ants in the fixed demographics young experiments had an average age of 40 ± 16 days, and old experiments had an average age of 171.56 ± 20 days. Errors represent the standard deviation.

Maximum excavated area from the fixed demographics excluding the area excavated by the queen.

A. The area excavated by the queen from the colony maturation experiments was excluded, and compared to the resultant area from the fixed demographics young, and old experiments. We find that the area excavated from the colony maturation experiments is similar to the fixed demographics young experiments.

Maximal digging rate before and after collapse.

The maximum digging per ant before (A) and after (B) the manual collapse events from the fixed demographics experiments. The maximum digging rate per ant is calculated by normalizing the maximum digging rate (change in area per day) by the number of ants. We find that there is no significant difference before and after collapse events (Kruskal-Wallis test, χ2 (5,68) = 6.01, p = 0.30)

Excavated area and digging rate.

A.The excavated area across group sizes from the fixed demographics young, and old experiments. We performed a linear regression on the excavated area across different group sizes for the fixed demographics young (y = 6.35x + 0.73, R2 = 0.43), and old (y = 4.57x − 0.73, R2 = 0.57). The positive slope indicates that the excavated area increases with group size for both the young, and old colonies. (B). The digging rate per ant for different age categories is shown. We find that the digging rate per ant is age-independent. Error bars represent the SEM.

Density model.

The progression of the excavated area based on the constant density regulation model, and its comparison to the colony maturation experiments. The constant density was fixed to be at 11.6cm2ant−1. Error bars represent the SEM.

Nest segmentation algorithm.

(A) Nest skeletonization - The medial-axis transform (MAT) image of the excavated area of a sample nest from the colony maturation experiments. The warmer color of the skeleton represents a larger distance to the edge of the excavated structure, shown in black. (B) The decision tree used for the structural segmentation of the nest images into three classes: namely, (1) Tunnels, (2) Wide tunnels, and (3) Chambers.

Nests excavated by colonies from the fixed demographics treatments.

The nest structures excavated by the young and old colonies from the fixed demographics experiments are shown. The first row corresponds to the nest images from the young colonies, and the second corresponds to the old.