Slashdot Mirror
Ants Use Pedometers to Find Home
Ant writes "New Scientist (a short video clip included) reports that desert ants have an internal pedometer that keeps track of how many steps they take, according to a new study. The insects seem to rely on this system to find their way back to the nest after foraging. Other insects may also possess this pedometer-like system. Some types of ants appear to use visual cues or leave scent trails to find their way home. But desert ants have a remarkable ability to retrace their steps from their nesting site even though they travel on flat terrain that is devoid of landmarks, and any odors quickly fade in the hot temperatures."
So if you let an ant crawl onto your hand, then place it down somewhere else.. it will get lost? :P
They should have constructed a mini treadmill (complete with moving walls) --seriously-- to see if the ants with normal legs still walk the same distance for a reward. That would really drive there point home.
I sure wish there were more information. If this is true; it's somewhat interesting. But with so little to go on, it could just as easily not be true.
As the most obvious example to spring to mind; they tried ants with legs (we're left to assume) 50% longer that went 50% further than home and legs 50% shorter that only got halfway home. They then say this is because he counts steps -- obviously each step takes the one ant 50% further and the other 50% shorter.
So what if the ant goes by the amount of time it's been traveling; nothing to do with counting steps at all --?
You'd expect exactly the same results.
I hope it's just the awful article -- if the study is so poor they've really learned nothing.
It's been a while since I worked on this, but these idea have been propagated through networking protocols for years. When I was in University at Dalhousie I spent quite a bit of time on a directed study of somethink called the 'AntNet Routing Protocol'.
:-)
The idea was based on the pheramone trails left behind whne ants seek food. You see, one ant leaves behind a trail, not a big one, but a small scent to be picked-up by other ants. When it finds food, it will retrace it's steps backwards and double the intesity of the pheramone trail. If another ant happens upon a trail, it will follow the trail to the food and increase the trail's intensity again. If the trail ever ends without a prize, ants look around to try and pick-up the trail again. Simple concept, right?
Adapting this behaviour from ants to packets on a network was easy. You had ants that walk forward and ants that walk backwards. Forward ants would collect hostnames, IP address and time stamps as they passed through any PC and kept going to their host. Backward ants updated the routing table when they retraced their steps. If any route had a lower cost (latency) then the entry already in the routing table, then an updated entry was posted. There was also a hidden advantage to all this - if, for any reason, a node went down or dropped off the network it was easily and quickly detected. Furthermore if a link went down, alternate routes were already in place if you kept double-layered routing table... quick, easy and fast network response times were the result. Consider time stamps like a tick on a pedometer...
In case you're wondering, all computers on the network ran NTP to sync the time and give us one less hassle to worry about (this could be easily incorporated if need-be).
My main area of research was to figure-out where and when the Ants started to impeed the network instead of help it. I found it to be a function of the number of discovery ants versus time and nodes on the network... some pretty rough math ensued from what I remember, but the time delta between discovery ants was paramount in any effective benefit to the network.
Food for thought... or to the trail with the most ants.
Jonathan Connell* built a mobile robot @ MIT which used a not too dissimilar system for navigation. It counted the number of doors that it passed through, and the number of turns to the left. This robot, Herbert, had the goal of collecting soda cans and would wander about the lab autonomously collecting these cans and returning home but making an appropriate number of entries through doorways and turns to the right using a magnetic compass as a rough guide. There was no internal map, no master plan, to 3D model of the world, no GPS yet this robot was able to navigate very complex, real-world spaces effectively. It's interesting to see that there's a biological model here that validates many of these assumptions.
** I hope I'm correct on the details... I'm going from memory from a reading of Connell's Master's Degree disseration I read probably ten years back... I believe the title was "Minimilist Mobile Robotics" but I'm certain it was published through Academic Press. This was one of the early MIT Mobile Robot Lab robots to use Subsumption Architecture.
What I find most interesting is how important it is to an ant colony to have inhabitants which wander and can navigate back to the nest. It seems so important, they've devised two different methods; one which depends on odor and this pedometer method. This specific need is so fundamental to their existence, it appears to drive their evolution.
I don't have a subscription to read the original article, but the glossy schtick pointed to in the original post was pretty weak: "we mutilated ants and they couldn't find their way home, and if I buy fish it won't rain on monday, so therefore they have a pedometer hidden inside their gasters!". Hopefully the original has more actual science.
Silver ants (they look more like they are chrome-plated than silver) also live in the Sahara. They come out at the hottest time of day, when all predators are hiding, and they are extremely reflective. They have a special gait that allows them to keep half their feet off the sand in the shadow of their bodies, and they keep switching off so their feet don't cook. They move about in a fairly normal search pattern, but when they find something they run directly back to the nest without retracing their original route! Although they are believed to have good vision, their environment contains almost no visual cues - one sand dune's pretty much like another - and they will pass through territory they haven't seen on the way back to the nest.
Silver ants are also very hive-oriented or "altruistic". Individual foragers will go past their survival distance looking for food, but they turn around and come back so that their dead bodies are within the survival distance and can be recovered by other foragers. That way, if there is a food/water source that is further out than an ant could travel without such resources, they will still find it and use it.
All this is from memory and the wiki article is lame. If anybody has some good links for silver ants please post!
In fact I used to perform such "experiments" with ants. The ants at my place used scent trails. If I rub-off the scent trail left behind, the ants coming behind get dis-oriented for a while.
And when I transport an ant manually to an unknown territory, it raises its head and looks around for familiar landmarks.
Not all ants use scent trails. I found that the larger ones use the direction of a light source (or their shadow) to navigate to a place.
I have a truly marvelous proof of the Riemann hypothesis which this sig is too short to contain...
Yo may also take a look to my own simulator of Ant's food-gathering behaviour:
r m
http://www.geocities.com/chamonate/hormigas/antfa
It tries to emulate the usual ants, that find the food and the way back using pheromone traces.
That would seem to be a very likely possibility, but it seems that you and others on here are missing one more important detail. The BBC's report says that the researchers found the ants with modified legs had no trouble returning to their home if they made BOTH the outward AND the homeward journeys with modified legs. This implies that modifying their legs might not be disorienting them too much after all. Instead, it reinforces the hypothesis that something else is involved, namely the counting of their steps. Here's the BBC's coverage of the same story: http://news.bbc.co.uk/2/hi/science/nature/5128604. stm