Stigmergy: how termites build their nests

Have you ever heard of stigmergy? No, me neither! The dictionary definition is, “the trace left in the environment by an individual action that stimulates the performance of a succeeding action by the same or different agent.” And, as he explains to Chris Smith, this is how Giulio Facchini, from CNRS in Paris, thinks termites, despite measuring only millimetres in length, can work together in their thousands to construct massive and intricate nests. He gave them coloured pellets of mud to work with to make it simple to follow their progress, and fashioned some initial curved shapes into the study arena. These shapes motivated the termites to assemble the coloured mud into very specific structures: in other words, by building a certain shape, other workers who encounter that shape are guided as where to put the next piece...

Giulio - Termites, like other insects that work collectively, interact according to a principle that is called stigmergy, which simply means that each individual doesn't know the global plan of their activity - that could be construction, that could be foraging whatever action - they don't know the full plan, but every single action of an individual leaves a trace for the following one to take on the work. And so we know this, but there is no common agreement on what should be this kind of trace. And there are a lot of different ideas in the literature that could be an odour, that could be a gas concentration, that could be some touching feeling. So we wanted to explore what this trace could be.

Chris - One example, for instance, these animals build big complicated nests that they've all worked together to construct that architecture. Mm-Hmm, <affirmative>. So how would they all know what to do and where to go in order to do that? Where are the cues coming from? I suppose that's part and parcel of this, isn't it?

Giulio - Yes. For example, ants can leave some odour within the small piece of mud that they add to a construction to like inform the following ants that they should add more mud at the same place. And that can lead to the formation of pillars. The same thing was proposed for termites, but actually we showed that odours in this case are not necessary and information can be transmitted just through the shape of the environment.

Chris - So talk us through how you reached that conclusion then. What did you actually do to disprove the pheromone or chemical marking idea and come up with the idea that it's all to do with the shape of what's already there?

Giulio - The idea was to prepare some topography instead of having, for example, a flat environment where termites move, you add some bumps, some details curvature. So we prepare this environment with mud that has been sterilised in the oven. So we are sure that there is no odour in that, there is no like chemical traces. And we observed that there will be some organisation in the building actions even without this odour.

Chris - So how did you track that? Was that sort of you video a bunch of termites interacting with that environment you've created to see how the shape of the environment changes the behaviour of the insect?

Giulio - Okay, so the idea was quite simple but quite tricky at the same time. The first thing is that if you leave termites with mud, at some point they will dig and prepare some small balls of mud that they use to build things. So the idea was to leave termites, prepare these pellets In some preliminary experiments for that we used some grey mud and then we collect these pellets, we put them in an oven, so we sterilise them and then we add that to a new experiment where the, all the environment that the one with shapes that I told you before is prepared with the same mud but of a different colour. So red. And so this way when we are filming the, the system, we can actually separate the, the pellets from the environment. And for example, when a pellet is taken out, we can do what is called the differences between images. And we will see a dark trace where a pellet has been taken out and the light trace when a pellet has been deposited, which give us information about the building actions.

Chris - And because the pellets have been baked in the oven, you reason this has driven off all the chemical cues. So the only thing left that could be guiding the deposition and the removal and the movement is the shape?

Giulio - Yes, exactly. And, and we, in order to be more sure about that, what we provide is a flat disc on which we add just two bumps, which are of the size a bit larger than than termites. And then we observed that the pellets that have been like placed were all over the disc were moved mainly to the top of the more curved region, which suggested that this shape can help, can guide termites in their building actions.

Chris - A couple of things obviously spring to mind off the back of that one is how do the termites gauge the shape and what sort of shape are they interested in? And second, how do you then rule out that having gauged the shape and begun construction, they don't then deposit some kind of chemical cue or reinforcement that then encourages other termites to, to carry on doing this?

Giulio - Two very good questions that we tried to address. So the first one, actually the answer came from a kind of surprise that we had in our experiment because I told you that we an increase of pellet deposition on the top of these pillars that are more curved. But we also observe the same kind of increase in deposition at the edge of the disc, at the flat disc, a region which is not very curved, which is pretty flat. This was a surprise for us because we had both a flat region and a curved region. But in the end what we understood is that what is in common between this tip of pillars and edge of the disc is the fact that the evaporation is very strong there. And it turns out that when, whenever you have a curved region, which is humid, evaporation is stronger there. So what actually happens is that both the edge of the disc and the tip of the pillars are regions where, thanks to this specific shape, evaporation is stronger. And we believe that the way termites sense this, this shape is precisely through evaporation flux. And this is quite well accepted in the community because termites tremendously need humidity to survive. So they must be able to sense where you switch from a dry to humid environment, which is exactly where evaporation is strong. And so concerning your second question, we cannot actually rule out the fact that once the building process is started, some pheromones, some specific order laid by termites could also play a role in the construction. So, and we, we say that in the paper, this mechanism could be like on the top of the shape guiding that could be pheromones guiding. But what we say is that at the very beginning there is no ferments. So there is no need for that to start, which is something which is something new because no one has suggested that before.