Studies of simple and emerging systems have been undertaken to un derstand the processes by which a developing system unfolds, and to understand more completely the basis of the complexity of the fully formed structures. The nervous system has long been particularly in triguing for such studies, because of the early recognition of a multitude of distinctly differentiated states exhibited by nerve cells with different morphologies. Anatomical studies suggest that one liver cell may be very like another, but indicate that neurons come in a remarkable di versity of forms. This diversity at the anatomical level has parallels at the physiological and biochemical levels. It is becoming increasingly easy to characterize the different cellular phenotypes of neurons. The repeatability with which these phenotypes are expressed may account in part for the specificity and reliability with which neurons form con nections, and it has allowed precise description of the first appearance and further development of the differentiated characteristics of individ ual neurons from relatively undifferentiated precursor cells. This rep resents a major advance over our knowledge of development at the level of tissues, and makes it feasible to define and address questions about the underlying molecular mechanisms involved. Central to these advances has been the clear recognition that there is no single best preparation for the study of neuronal development. Furthermore, it has become evident that no single technique can tell us all we want to know.