When carrying resources from a collecting point to the nest, one would assume that animals would attempt to carry as much as possible to maximize their foraging efforts. However, among social insects that is not always the best strategy. Foragers carrying large loads might overwhelm the individuals processing the resources in the nest, causing a bottleneck. Additionally, a heavy load slows the carrier down. That may not be a significant cost for a solitary forager, but it may reduce the gains for a colony as a whole. Carrying loads well below maximum carrying performance actually reduces the burden on the resource processors and speeds up the forager, allowing for more foraging trips per unit time. But there is yet another factor whereby lighter loads can increase colony efficiency. Walking foragers, such as ants, often forage along well-defined trails and, depending on the number of foragers, this can potentially cause traffic problems. In that context, when one ant is slowed down by a heavy load it also slows down those ants following behind, regardless of load size. This phenomenon, called the ‘truck-driver’ effect – where a heavily laden truck slows down normally faster cars – describes another situation where carrying too much negatively impacts whole colony foraging efficiency.

Alejandro Farji-Brener and colleagues, from CRUB-Universidad Nacional del Comahue e INIBIOMA-CONICET in Argentina, investigated and quantified the effects of heavily laden ants on foraging ant traffic in the leaf-cutting species Atta cephalotes. Working in Costa Rica, they first characterized ants and load sizes by collecting individual ants and their loads and determining leaf-cutting and ant dorsal surface areas. ‘Highly laden’ ants were 25% bigger and carried 100% larger loads than ‘ordinary laden’ ants. Then they performed two field manipulation experiments on foragers of several A. cephalotes colonies. In the first they identified highly laden ants and measured walking speeds for the highly laden ant, the ordinary laden ant following directly behind (treated ant) and another nearby ant as control. Then they removed the highly laden ant and mimicked the removal hand-motion over the control ant. After 5 s they then measured the treated and the control ant's speeds. In the second experiment they ‘created‘ highly laden ants. Initially, the speeds of two ordinary laden non-consecutive ants were measured. Then the load mass of an ant directly ahead of one of the two measured ants was artificially increased by adding a 10 mg piece of aluminium foil to her leaf cutting. After 10 s the speeds of the treated ant (following the ‘created’ highly laden ant) and the control were measured again. They also investigated the relationship between numbers of highly laden ants and ant traffic flow.

Removal of highly laden ants increased ordinary laden ants' walking speeds from 1.9 to 2.9 cm s–1. Similarly, the ‘creation’ of highly laden ants reduced their followers' speeds from 2.4 to 1.6 cm s–1. These results clearly show how ant load can impact the overall colony foraging rate and, by extension, the intra-colony fitness. Furthermore, Farji-Brener's team also found a negative relationship between the number of highly laden ants and ant density (traffic) on foraging trails. Thus, at low traffic flow ants can cut and carry larger leaf fragments without the concern of slowing other ants, but when traffic flow increases they refrain from cutting larger fragments. It appears that individual ant foragers leaving the nest can estimate the outbound traffic flow and use this information to estimate the future flow of returning laden ants, thereby modulating the sizes of leaf cuttings made in order to avoid delays in overall colony foraging rates. This study shows remarkable flexibility in foraging behaviour and supports the idea that leaf-cutting ants make choices not only as individuals but also collectively.

A. G.
F. A.
Sanchez Cuervo
A. M.
The ‘truck-driver’ effect in leaf-cutting ants: how individual load influences the walking speed of nest-mates
Physiol. Entomol.