The picture of the school as an ‘egg crate’, composed of segregated classrooms, co-ordinated by an invisible administrative apparatus, is perhaps in keeping with the classical bureaucratic model. For today's schools it is rather misleading at least. Not only have schools evolved multiple layers of interdependence and communication, they have also increased in size — to the extent that a comprehensive high school of over 1,000 pupils is no longer a rarity. At the same time schools have diversified into a range of specialised units that further compartmentalise and fragment their environment into more manageable sections. Attempts to discover the genesis of local school bureaucracies to some extent may therefore miss one of the more immediate problems of school organisation, the features of the task environment which lead to structural changes. If we were to ignore for the moment the larger problems of state educational systems and were to concentrate on the more concrete, contingent and practical aspects of school life, we may develop deeper insights into organisational processes. This reversal of perspective was encountered earlier in Dreeben's argument that Parsons's ‘values-goals-functions’ scheme cannot handle technology and does not help in the empirical study of organisations. Dreeben asks: ‘Why should one organization differ from another (universities and schools) and why should the structural characteristics of any organization cluster as they do?’ (1976, p. 869). Dreeben also claims that ‘Viewing structural variation in technological terms not only allows an understanding of why organisations with the same goals differ, it also does justice to their historical development’ (p. 869).