About half the world lives in cities and the proportion is rising. Societies across the world look to engineers to provide a sustainable urban existence by cleaning or managing the quality of our soils water and air; cleaning up the pollutants we produce today and the pollution created by earlier generations. For over a century now engineers have turned to biological treatment systems to help them in this task. These systems can be used to clean domestic and industrial wastewater, contaminated groundwater, sources of smell, municipal refuse and contaminated land. The genius of engineers and the paradox of engineering is that systems can be successfully manipulated even though we don't fully understand how they work. This is particularly true for engineered biological systems where engineers must operate with only a sketchy of how even the most basic concepts of the nature abundance and activities microbes that treat the waste. Much of what we know has been learnt by empirical (trial and error) research and theoretical foundation we exploit is at best often based on science from the 1940s. The history of engineering tells us that improving the theory behind design can help revolutionise engineering practise (compare the bridges of the late 18th century with the bridges of the late 19th century). We believe that, if environmental engineers can improve the theoretical basis of design, we have similarly radical effect on the design of biological treatment systems. The study of the relationship between an organism and its environment is called ecology. We therefore believe that theoretical ecology is the science from which we can learn the most. We have therefore sought to use ecological theory to improve our ability to predict three aspects of a system: the nature of the species present, the presence (or otherwise) of chaotic dynamics, the relationship between pollutant consumption and resource manipulation by the engineer. We believe that success and failure in a biological treatment system often relates to one of these three interlinked aspects of biological treatment. Using ideas originally developed to predict the composition of tropical forests and tropical islands, we has successfully developed a model that allow us to predict the number and possible identity of bacteria in biological treatment systems. This model appears to work in a very wide variety of situations and may well be universal. We would like to formally corroborate this model, and calibrate it for a wide range of engineered systems and see if it will tell us something about the fate of endocrine disrupting compounds in wastewater. The out come of this model can be fed directly into another model that we are developing with colleagues all over Europe to predict how microscale growth of these species translates into the macroscale behaviour in nitrifying biofilms and microbial fuel cells. We have developed and tentatively employed tests of chaotic dynamics in biological treatment systems. We need to complete those tests and back them up with mathematical models of possible causes of chaos to see if we can relate the basic biology to the stability of the system. We have also developed models to relate resource availability to the probability of failure in a system. At present this can be used to determine the probability of for a given aeration rate however it could be used much more widely. The major problem is the calibration of the models. Calibration means determining the abundance of individual species. This is possible at present but slow. Faster simpler methods are needed and the search for such methods is emerging as a priority for us. What we propose is difficult and requires large amounts of data and specialist skill. We have accelerated our progress and enhance our funding by cooperating internationally and we have disseminated our skill through workshops. We will continue to both cooperate and train as we work towards our goal.