Gene transfer into living organisms has evolved as a powerful approach to study in vivo effects of specific genes and to devise animal models of hereditary disorders. We have been particularly interested in an approach to introducing transforming genes into the nervous system. Since specific promoter sequences for targeting the expression of a transgene to many cell types of the brain are not yet isolated, a suitable transgenic mouse model was not available for these experiments. This has prompted us to develop an alternative strategy for gene transfer into the brain. The rationale is to introduce foreign genes into fetal brain transplants using embryonic CNS as donor tissue and replication-defective retroviral vectors as genetic vehicles. This technique relies on the extraordinary organotypic differentiation capacity of neural grafts and the expression of retrovirally transmitted genes in different cell types of CNS transplants. In contrast to transgenic animals but analogous to sporadic tumour formation, target cells for the retroviral vector will develop in an environment of unmodified neural tissue. We have introduced a number of neurotropic oncogenes into fetal brain transplants to study potential effects of such genes on the brain. This review will summarize some of the findings which have emerged from this experimental study including the tropism of several genes for endothelial cells, attempts to identify cooperating combinations of transforming genes and an experimental model for primitive neuroectodermal tumours in neural grafts.