Colonies of ants allocate exploratory individuals to where they are ecologically needed

Abstract

Individual differences in behavior have large consequences for the way in which ecology impacts fitness. Individuals differ in how they explore their environment and how exploratory behavior benefits them. In group-living animals, behavioral heterogeneity can be beneficial because different individuals perform different tasks. For example, exploratory individuals may discover new food sources and recruit group members to exploit the food, while less exploratory individuals forgo the risks of exploration. Here we ask how individual variation in exploratory behavior affects the ability of Argentine ant Linepithema humile colonies to (1) locate novel food sources, (2) exploit known food resources, and (3) respond to disruptions while foraging. To address these questions, we conducted field experiments on L. humile foraging trails in which we manipulated food availability near and at the foraging trails and disrupted the foraging trails. We sampled individuals based on their response to the perturbations in the field and tested their exploratory behavior in the lab. We found that exploratory individuals benefit the colony by locating novel foods and increasing resource exploitation, but they do not play an important role in the recovery of a foraging trail after disruption. Thus, the benefits of behavioral heterogeneity to the group, specifically in exploratory behavior, differ across ecological contexts.

Keywords: exploration; foraging; individual variation; novel resources; recruitment; resilience.

Figure 1.

Experimental procedure for testing the three different hypotheses. The three boxes on the left detail the sampling procedure for each experiment in the field. On the right is a photo of the 8-armed radial maze in which sampled ants were tested for exploratory behavior in the lab.

Figure 2.

Exploratory behavior in relation to deviating from the foraging trail. Exploratory behavior of individual ants (number of visits to arms in an 8-armed maze in the lab) was greater (A) for ants that deviated from the foraging trail (off trail, orange) for all bait distances (N = 160 ants, N = 6 trails), than for ants that were collected from the foraging trail (on trail, purple) (N = 160 ants, N = 6 trails). However, (B) exploratory behavior did not relate to the distance of the bait from the foraging trail (7 cm [N = 120 ants, N = 6 trails], 14 cm [N = 100 ants, N = 5 trails], and 21 cm [N = 100 ants, N = 5 trails]). Here and in all following box plots, horizontal lines are the medians, boxes extend to 25 and 75 percentile, whiskers extend to 1.5 times the interquartile range, and dots are outliers. Asterisk indicates statistical significance (P < 0.0002) with a GLM.

Figure 3.

Trail rate and exploratory behavior after food supplementation. (A) Trail rate decreased over time (measured every 15 min) after food supplementation (N = 7 trails). (B) Exploratory behavior of ants collected after food supplementation did not change (N = 400 ants). Points are slightly jittered along the x- and y-axis to enhance visibility, the lines are linear fits to the data with 95% confidence interval (shaded area around the lines) plotted using the geom_smooth() function in the ‘ggplot2’ package of R, with the default Loess method. Solid line (A) indicates a statistically significant relationship and dashed line (B) a non-significant relationship from a GLM.

Figure 4.

Exploratory behavior in response to disruption on the foraging trail. Ants collected before the trail was disrupted are referred to as “pre-disruption” (purple) (N = 70 ants, N = 5 trails), ants that were the first to cross the midpoint of the disrupted area are referred to as “first to cross” (pink) (N = 70 ants, N = 5 trails), and ants that turned around when they reached the disrupted portion of the trail are referred to as “avoided disruption” (blue) (N = 70 ants, N = 5 trails).