BERKELEY — An amazing variety of arboreal animals have learned to glide through the forest, from flying squirrels to flying lizards and frogs, and even — frightening as it sounds — snakes.
Now, add ants to that list, say biologists at the University of California, Berkeley, and the University of Texas Medical Branch in Galveston (UTMB).
Gliding ants — the only wingless insects known to actively direct their fall — were first observed last year outside Iquitos, Peru, by insect ecologist Stephen P. Yanoviak of UTMB. While perched 100 feet up in the rainforest canopy waiting for mosquitoes to alight and feed on his blood, Yanoviak casually brushed off a few dozen ants that were attacking him and noticed their uncanny ability to land on the tree's trunk and climb back to the very spot from which they'd fallen.
"I brushed them off the branch with my hand, and I noticed that maybe 20 or 30 of them fell in unison and then made this nice little cascade back to the tree trunk," said Yanoviak via phone from Iquitos. "That's when I realized something was up with this behavior and it was worth checking into a bit more."
By painting the ants' rear legs with white nail polish, he was able to track their fall and establish that they come in backwards to the tree, hit and hang on, though they often tumble down the trunk a few feet and occasionally bounce off. They can actually make 180 degree turns in midair, however, so even when they fail the first time, they can execute a mid-air hairpin turn and glide in for another try.
"It's an amazing discovery," said Robert Dudley, a UC Berkeley expert on flying and gliding creatures ranging from hummingbirds and lizards to moths and bees. "Apparently, it's fairly common among a lot of tropical canopy ants."
Yanoviak, Dudley, a professor of integrative biology at UC Berkeley, and zoologist Michael E. Kaspari, an ant ecologist at the University of Oklahoma, Norman, report the discovery in the Feb. 10 issue of Nature.
After Yanoviak's initial observation, he conducted further drop studies and established that about 85 percent of the ants — a widespread tropical canopy species called Cephalotes atratus — are able to land on the tree trunk and climb back up, as compared to a mere 5 percent that would be expected to land on the trunk if the ants were parachuting randomly to the ground.
"I noticed this five years before in Panama, but it was at that moment in Peru 18 months ago that I realized why they do it: They're trying to get back to the tree, they're trying to not get lost in the understory of the forest," he said. Once they hit the forest floor, he added, they're unlikely to find a chemical trail back to the nest, and most likely will be eaten by predators.
Yanoviak said that perhaps more important from the perspective of survival, the forests are flooded as much as half the year, so an ant falling into the water will most likely end up as fish food.
"That's what I think is the major evolutionary driving mechanism behind the behavior," he said. "In Amazon forests, you really don't want to fall out of your tree and in the water, because then you're definitely dead."
After capturing some quick video snippets with his digital still camera, Yanoviak contacted UC Berkeley's Dudley to arrange a more scientific study of the ants' gliding behavior. They met up with Kaspari last summer in Panama for a week to videotape the same species of ant in freefall, with a white sheet standing in for the tree trunk.
Yanoviak's earlier observations combined with the videotaped tests in Panama established that the ants basically reorient their bodies so their hind legs and abdomen point toward the tree, and use their head-up fall through the air to take them feet-first toward the trunk. How they land is still a mystery, but evidently claws on their back legs act like grappling hooks to snag the trunk and hang on.
"When they drop, they often glide away from the trunk, then turn and come in backwards," Dudley said. "Their 180-degree turns are pretty dramatic in the absence of wings."
The team found that, unlike flying squirrels that can glide horizontally for long distances, the ants fall at a relatively high velocity, around 4 meters per second (12 feet per second), and approach the tree at a steep angle. This means they frequently bounce off the trunk, though, as Yanoviak initially noted, they just make a 180 degree turn and try again — usually with success.
What allows the ants to change direction so quickly is still a mystery. They have long, slightly flattened hind legs which, when combined with abdominal movements, might allow the ants to reorient in midair. They also have an unusual flattened head with flanges that could act as a rudder, Dudley said.
"My guess is that, by gliding backwards and using their legs and also their flat head with flanges, they could steer," he said, though more studies are needed before the question can be answered.
Yanoviak noted that worker ants are secondarily wingless, meaning they probably evolved the ability to glide after they lost their permanent wings, which were common in primitive ants.
"This discovery doesn't necessarily say something about the evolution of flight in terms of insect lineages, but it does give us good information about what kinds of morphologies are involved in the transition from just parachuting, or just falling uncontrolled, to being able to control where you land," Yanoviak said.
The gliding behavior would be a definite advantage for ants, like Cephalotes, that forage at the outer reaches of branches and are in greatest danger of being knocked off by wind gusts or passing monkeys. There is some evidence, Dudley said, that these ants sometimes purposely drop off the tree to avoid predators, evidently secure in the fact that they can glide back to the same tree.
"To me, one of the most intriguing elements of the story is that so many ants jump voluntarily," Kaspari said. "The last view of a lizard chasing a Cephalotes ant may be that ant dropping toward the ground and spiraling back to the tree, safe from both the litter below and the predator above."
All told, Yanoviak has dropped about 60 species of ants from the treetops, and has found some form of gliding — what he and his colleagues call directed aerial descent — in 25 species representing five separate genera. It is the norm in only two groups, however: the Cephalotini tribe, which includes Cephalotes atratus, and the arboreal Pseudomyrmecinae ants.
"It turns out that every species in the genus Cephalotes that I've dropped so far shows this gliding behavior, and they are by far the best gliders," Yanoviak said. But the long, cylindrical wasp-like Pseudomyrmecinae ants are also pretty good at it, as are several species of carpenter ants he recently dropped.
"There are three species of Cephalotes that also occur in the States - one in Arizona, one in Texas, and one in the Florida Keys," he noted. "I'd like to get up there and drop those to see what they do."
"Steve's first observation may be leading to a whole Pandora's Box of gliding arthropods," added Kaspari. "Steve, Robert and I right now are exploring how pervasive such directed descent might be — we are, in fact, finding it in a variety of wingless arthropods that glide in using a variety of techniques. Plummeting to the earth may be the exception, not the rule."
Yanoviak is in Iquitos to explore how deforestation affects the mosquito transmission of Venezuelan equine encephalitis. In hopes of establishing a mosquito colony in the laboratory, he goes in search of mosquitoes and larvae in the canopy where they live, mostly above private land around the city. Using ropes and cave-climbing equipment, he ascends trees and collects mosquito larvae from water-filled tree cavities and also collects adult females who've had a recent meal of blood — his own.
He and Dudley are applying for grants to do a more systematic survey to see how common gliding is among the hundreds of species of canopy ants as well as other insects that live in the trees, and to explore the biomechanics of their gliding.
Yanoviak also is affiliated with the University of Florida Medical Entomology Laboratory in Vero Beach. The work was supported in part by grants from the National Science Foundation and the National Institutes of Health, and with the assistance of the Smithsonian Tropical Research Institute in Panama.
Cite This Page: