Progress has been made in recent years in defining the spectrum of molecular changes seen in human pancreatic cancer. A range of animal models of this disease exists, and the nitrosamine induced tumours in the hamster show "clinical", morphological, immunophenotypic and, importantly, molecular similarities with human disease. This may prove to be an important model for testing therapeutic stratagems. In vitro systems for assessing pancreatic carcinogenesis and neoplasia are much less well developed. Although many human pancreatic primary carcinoma cell lines exist, it is difficult to culture normal human pancreatic cells effectively. Future studies will probably need to concentrate on the development of rodent systems (especially hamster), which can be more easily maintained. Studies in transgenic mouse systems have given new insights into the genesis of pancreatic neoplasia, and taken with those studies of acinar to ductal transdifferentiation, they cast new light on the possible origin of the tumour. This important issue remains unresolved, and its further understanding will require the generation of new data on the kinetic and spatial organization of the pancreas where key questions remain unanswered--whether there is a stem cell compartment in the organ not least among them. Progress in the cellular and molecular aspects of the disease are essential; so too are some important morphogenetic issues.