model

Smelling collectively

You can see the steam rising above the cup in this coffee at Carbon Kopi. But you will have to imagine the aroma.

It is hard to choose the best thing about coffee, so many aspects combine to make a good cup. But one of the key things about drinking coffee, particularly if you have had a difficult meeting or have just come in from the cold, is the aroma that wafts up as you grind the beans, add water to bloom the coffee and then brew. In happier times, we may be walking down the street preoccupied about something that is going on and then suddenly get hit by a fantastic aroma that signals our proximity to a good cafe. We perhaps ‘follow our noses’ to the source of the smell and then breathe in the scents as we enter the cafe. Which brings us, in a round about way, to moths and a recent paper that appeared in Physical Review E.

It is not that moths have been shown to have a particular liking for the smell of coffee. That may be an area of future research for somebody. But they do need a very good sense of smell because they need to be able to ‘follow their noses’ in order to find the source of a smell that they are interested in (typically a pheromone released by a female moth). This female moth may be located 100s of metres away from the male and probably does not emit that much odour, so how do the male moths find her?

In a similar manner to our approach to the aromatic coffee shop, the moths first travel against the wind, aware in some sense that the smell is carried downstream. If they lose the scent, they then fly perpendicular to the wind flow in an attempt to sniff the aroma once more. This pattern of zig-zagging flight allows them to approach the source of the smell fairly quickly*.

Eggs of a large cabbage white butterfly. No real links with coffee and few with moths, but the adult pair may well have had to find each other using the sense of smell.

It’s a clever method that is perfect if the wind flows in one direction without any turbulence. But how many times have you watched as leaves have been swept up in the wind flow and danced a swirling vortex pattern before falling back to the ground? Or, as you approach the side of a tall building, you get hit by a gust of wind that seems to come in a number of directions all at once because of the way that it is being affected by the presence of the building wall? We can see a similar thing in babbling streams and in our coffee as the convection currents swirl in vortices. The real world is not so simple as a linear wind flow, in the real world the wind is turbulent.

And yet still the moths find their way to the source of the smell that they are seeking. How do they do it, and could we design a robot (or robots) to emulate the moths in order to find, for example, chemical leaks? It was these questions that were addressed by the recent paper in Physical Review E. In the study they used mathematical calculations to look, not at the behaviour of an individual moth, but at the behaviour of a swarm of moths, a group of moths all searching for a mate.

In the computer model, each individual moth could discern the wind speed and direction and also detect odour molecules. So, left to their own devices, the individuals in the model would follow the zig-zag pattern of individual moths observed in nature (this was a deliberate element of the model). But the model-moths were given another ‘sense’: the ability to see the behaviour of their fellow model-moths. Which direction were the others going in? How fast were they moving?

The model-moths were then provided with one final behaviour indicator, a parameter, β, which was called a ‘trust’ parameter. If β = 0, the model-moths did not trust what the others were doing at all and relied purely on their own senses to reach the prize. Conversely, if β = 1, the model-moths completely lacked confidence in their own ability to discern where the smell was coming from and followed the behaviour of their peers.

We find our way to a cafe via visual cues or perhaps the sounds of espresso being made. But can we also follow the aroma?

Running the model several times for different wind conditions including a turbulent flow, the authors of the study found that the moths reached the destination smell best if they balanced the information from their own senses with the behaviour of their peers. In fact, the best results were for a trust factor, β ~ 0.8-0.85 meaning that they trusted their peers 80-85% of the time and relied on their own decisions 10-15% of the time. If they did that, they reached the smell source in only just slightly longer than it would take a moth to fly directly to the source of the smell in a straight line. An astonishingly quick result. As the authors phrased it, the study indicated that you (or the moths) should “follow the advice of your neighbours but once every five to seven times ignore them and act based on your own sensations”.

Now it would be tempting to suggest that this study has no relevance for us individuals finding a coffee shop and minimal relevance to coffee. But that I think would be premature. For a start, a similar result was found when the question was not about moths but about the best way for a crowd of people to leave a smoke filled room. If everyone behaved individualistically, or conversely, if everyone behaved in a purely herd like manner, the crowd took longer to escape the room than if people balanced their individualistic needs with a collective behaviour. It is a push to suggest that the same thing may be relevant for us finding cafes, but who knows what may happen post-lockdown(s) as we collectively attempt to find a well made flat white to enjoy outside our homes. Maybe we too need to trust our own senses some of the time but be open to taking the advice of those around us too.

*You can read more details in the paper Durve et al., Phys Rev E, 102, 012402 (2020)