From Worker to Colony:
Understanding the Organisation of Insect Societies

David Sumpter

7th and 8th of December 2001

The Issac Newton Institute, Cambridge.

Problems of determining how individual ants or bees can work effectively together, without centralized control, to build complex structures and make collective decisions are central to understanding insect societies. Such problems also arise in the study of other areas of biology, such as the immune system and regulatory gene networks, where a high degree of co-operation between individuals produce complex but robust patterns. Addressing these problems requires both the development of new types of mathematical models of the 'individual-based' data which experiments provide, and new theoretical perspectives for understanding the processes behind the generation of complex patterns. The aim of the 'worker to colony' meeting was to report on recent experiments on the organization of insect societies and attempt to identify general processes underlying such organization.

The meeting was the largest social insect meeting in the UK in recent years with around 90 participants over two days, and proved a lively affair. There were five excellent plenary addresses by Nigel Franks (Bristol), Thomas Seeley (Cornell), Walter Tschinkel (Florida), Nick Britton (Bath) and Guy Theraulaz (Toulouse) and many good quality contributed talks. Full details can be found at

and abstracts are available on line at

Since a major aim of the meeting was to outline mathematical problems related to social insects, it is worth listing a few of these problems here. I encourage you to contact these peole with any ideas:

- Guy Theraulaz (Toulouse) presented experimental data on the formation of Turing-like patterns in ant cemeteries. He and his group have completed a lovely set of experiments measuring the individual behaviours of ants as they pick up, carry and drop other deceased ants. The researchers then used these observations to build both an individual-based model as well as a activator-inhibitor model of the phenomena. The living ants form regularly spaced piles of dead ants. The possibility of a Turing mechanism explanation of their experimental results was further supported by doubling the size of the arena in which the ants built their cemetery. These experiments are one of the strongest examples of a biological system really obeying the types of patterns that Turing and others predicted. A word of caution, however, because these ant cemeteries have not been observed in the field---and could be said to be a lab artifact. As one anonymous biologist commented "It is a bizarre state of affairs when we are meant to congratulate someone for making an experimental manipulation in order to demonstrate what is meant to be a general principle of biological organisation". Biologists, eh? I for one was still impressed.

- Walter Tschinkel (Florida) was less sure of the mechanisms underlying another aspect of construction by ants: nest digging. He had collected some fantastic field data on the structure of these nests by pouring plaster into the entrances. The resulting plaster casts, some up to 4 metres high, gave a clear picture of what these nests looked like. The nests spiral down through the (homogeneous) sand in one thin passageway with a number of regularly spaced chambers pushed out horizontally. Some of the chambers had a kind of fingered structure to them, which as Phillipe Rassee (Brussels) suggested in his talk, may be caused by bifurcations as ant densities become to high as they push forward through the sand. Another interesting fact is that 80% of the nest spirals go to the right. Are ants right-handed? Walter Tschinkel is very keen to get involved with mathematicians who might shed light on his structures. He certainly has good quality data.

- Nigel Franks (Bristol), Thomas Seeley (Cornell) and Stephen Pratt (Princeton) are all working on nest choice by ants and bees. Here the question is one of decision making. How do the insects make a collective decision between a number of different sites? and how do the achieve consensus? The mathematical questions posed here are less traditional. The speakers talked of differential equations, weighted average strategies and Buffons needle, but one approach which their observations of the actions of individually marked insects clearly supports is individual-based modelling of the nest migration processes.

- Iain Couzin (Leeds) presented individual-based models of how army ants organise their movement backwards and forwards along trails. Again, there appeared to be many aspects to this problem that was answerable only through study of interactions between individuals. It would be very interesting to see if a continuum model of ant trail traffic flow could be developed.

- Mary Myerscough (Sydney) was concerned with the role of individual differences in foraging by honey bees. This question of individual differences was also addressed in a poster by Han de Vries and Koos Biesmeijer (Utrecht). The approaches of these researchers was different, the former presenting a differential equation model, the latter an individual-based model. An important question raised here was how do you have an analytically tractable model of a population where individuals are either genetically different, or subject to different experiences?

- Marco Dorigo (IRIDIA, Brussels) and Martin Middendorf (Eichstat) both presented different aspects of ant algorithms. This is a fascinating area of research, where computer scientists have shown that ant-like behaviours can produce new, faster solutions to NP-hard and other computational problems. Fabrice Saffre (BT Exact technologies) is going to set loose some virtual spiders on the world wibe web in order to get our emails delivered faster.

There were also other interesting talks and posters presented at the meeting, and those listed above are only a subset. I would like to take this opportunity in thanking all the participants for making it an excellent two days in Cambridge. Thank you.


<< Move to UK Nonlinear News Issue 27 Index Page.
Last Updated: 5 February 2002.