Longhorn crazy ants earn their name from their tendency to zip about erratically rather than moving in straight lines, but as a recent study shows, they are not as scatter-brained as they may appear. Biologists, physicists and computer scientists based in Israel and France have put their heads together to explain one curious phenomenon demonstrated by this common ant species: an impressive ability for an army of longhorns to pool forces to carry phenomenal loads back to home base, without getting waylaid by obstacles.

Using experimental observations, the researchers have detected a formerly unknown type of scent trail used by the species specifically in mass-transport situations: ‘locally-blazed trails’ laid down by non-carrying ants—which the carrying group will then either follow or ignore.1 Random though the technique may sound, the research team has demonstrated its startling efficiency, thus raising the possibility of new strategies for handling comparable scenarios in the human world.

Scientists have long established that in the superbly well-organized society of an ant colony, when a foraging ant locates a food source, it releases pheromones from a gland in its posterior abdomen to create a chemical-signal trail that flags the full route from food to nest. Fellow colony members that pick up on the scent with their antennae only have to follow the trail to bring the food home. Simple, right? 

Not always. The situation gets trickier when the food item is too heavy for a single ant and demands a gang of workers—a test of coordination as well as muscle power, as anyone who’s ever joined in on a group moving a massive furniture item can attest. Things get even more dicey if the scent trail laid down by the forager ant passes through tight spots. As indicated by CNRS mathematician and computer scientist Amos Korman of the IRIF,2 who led the French team investigating the algorithms behind the longhorn strategy, “individual ants have a limited capacity to perceive reality on how to best guide the horde of carriers.” In human terms, a comparison can be drawn between the individual ant’s trail and traffic road signs intended for drivers of small cars, which can sometimes be misleading for drivers of large trucks when the suggested itinerary includes narrow roads and low bridges inaccessible to oversized vehicles. At the same time, truck drivers—or working parties of insects—miss out on useful information if they ignore the hints altogether. So what, then, is the best way for a squad of ants to negotiate its way back to its nest?

This challenge is one familiar to longhorn crazy ants, a species with a tendency to build huge connected colonies that naturally drive a need for ‘truckloads’ of food. To study the way in which longhorns collectively transport hefty burdens, researchers in France and Israel joined forces, basing their study on ant experiments led by physicist Ofer Feinerman and ethologist Udi Fonio, both from the Weizmann Institute of Science in Rehovot, Israel. These observations were made possible by the scientists’ development of  “an unprecedented method for detecting where ants lay down scent,” explains Korman. “We noticed that just before depositing scent, ants move backwards slightly. So by filming the ants from above, we captured their backward movements and determined where exactly scent was left.”

The research team’s observations revealed, first of all, that when an ant locates a food item too big to carry alone, it returns to its nest to recruit help, all the while depositing undecane—a distinctive type of short-term trail pheromone. Next, when enough longhorn helpers are rounded up, many ants on hand don’t actually pitch in with the lifting but tag along for the walk home. The scientists noticed that some of these escorts individually lay down undecane scent marks along the way. But rather than tracing an unbroken, well-defined route between food and nest like a classic ant trail, these ‘locally-blazed trails’ follow what Korman describes as a “dynamic, punctuated, staccato-like rhythm.” Indeed, they take the moving load as their starting point and only mark a small leg of the trip, strikingly emerging as the first-ever identification of an ant trail traced during the food-retrieval process rather than prior to it.

What the carrying ants actually do with these trail scents is also noteworthy: they ignore them about 20 % of the time. Given that some of the route information only applies for a single 3 mm-long ant and not a whole throng, it is only logical that not every lead will be useful. Sure enough, there are times when longhorn carriers get caught out and need to backtrack, but hit and miss is all part of the strategy. “What’s surprising,” notes Korman, “is that the ants may follow arbitrary directions even when no obstacles lie in their path. The thing is, the ants don’t know whether they’re going to get stuck or not so their strategy simply integrates randomness.” And this strategy works: as indicated mathematically by a navigation model designed by the France-based side of the research team, an agent using this random strategy will reach its destination in near-optimal time. The efficiency predicted by the mathematical model was further validated empirically when the scientists measured the time taken for ants to skirt around obstacles. By using the undecane trails as a ballpark guide only, load-lifting ants succeed in heading in the right direction while generally avoiding the trap of overly cramped pathways. Having detected how longhorns respond to locally-blazed trails, the French-Israeli team now aims to “determine the specifications of this obstacle-bypassing mechanism,” namely “how and when individual ants decide where to leave scents.”

In light of the longhorns’ high-performing navigation tactic, Korman also anticipates that humans can follow the ants’ trail in various domains. For instance, navigation applications that define road itineraries for drivers display “routes which change all the time according to traffic conditions, but directions don’t always indicate the shortest path as it’s hard to anticipate what will happen in the future. Because these instructions are not 100 % reliable, a better approach may involve a bit more randomness,” suggests the CNRS researcher. More widely, he points out that the longhorn strategy of not following every single instruction “in fact illustrates a very basic principle. As a thought exercise, imagine a parent giving child advice: the parent can’t know everything about what the child needs, so while most of the advice may be correct, in some cases, it will be plain wrong. What the ants may teach us is that the child would be quite right to roll a virtual dice and follow each piece of advice according to a certain rate of probability only.”