The invitation to open this session set me thinking about the development of numerical procedures of classification and ordination. Their technical development has been reviewed by various authors from various viewpoints, e.g. Cormack (1971), Orloci (1975, 1978), Dale (1975), Goodall (1970), Greig-Smith (1954, 1964, 1980), Whittaker (1967, 1973). I do not intend to discuss this in more than broad terms, but there is another aspect that has received less attention. This concerns the influences and constraints which have affected the development of numerical methods and their acceptance by phytosociologists. Their acceptance is particularly important; numerical methods are tools and unless they are used in the investigation of real ecological problems we are wasting our time in developing them.