ABSTRACT
A particularly di ffi cult task, in carrying out the analysis for a parti cular zone, was to decide on the nature of the Wk, k t-j, backcloth for the zone. This is at the heart of the task of describinS, through these mechanisms, the evolution of the structure of the whole system: in describing what happens in a particular zone, we have to build in what is happening in a 11 other zones. Another i flllortant poi nt can usefully be noted at this stage: any ad hoc development - say the building of a large shopping centre by some particular entrepreneur - can obviously have a major effect on the development of the system as a whole. In this sense, the predictive modeller is faced with the task of building in 'historical accidents'. This also
demonstrates the truism that there are many possible particular paths of evo 1uti on.
