The hypothesis, that gap junctional intercellular communication plays a key role in carcinogenesis and more generally in growth control was formulated nearly 40 years ago. From this time, data accumulated, showing that this type of communication is frequently decreased or absent in cells treated with tumor promoting agents, among transformed cells or between transformed/tumor cells and normal cells. This observation has been made on various cell types and whatever their tissue and species origins, by using in vitro and in vivo models. It led to the general assumption that the inhibition of gap junctional intercellular communication may play a role in carcinogenesis at two levels: (1) during tumor promotion by favoring the clonal expansion of initiated cells and (2) after the phenotypic transformation of cells by preventing the diffusion of putative "normalizing" factors between tumor cells and surrounding normal cells. During the past decade, the discovery that gap junction proteins, the connexins (Cx), may act as tumour suppressors, by reverting the phenotype of transformed cells confirmed the idea that their lack of function would be actively involved in carcinogenesis. However, we still do not know precisely what are the molecular processes that gap junctional intercellular communication may regulate and still do have very few data concerning the gap junction situation in human cancers. All these aspects are presented from an historical point of view and discussed below.