As more and more polymorphisms are discovered in the genes encoding cytokines, a crucial question is whether this polymorphism has any functional effect. One of the most widely studied cytokine genes in this respect is the gene encoding human TNF-alpha. Much of the literature investigating the issue of whether TNF-alpha promoter polymorphisms have any functional effect on TNF-alpha transcription or influence disease susceptibility appears to report negative results, giving the appearance and leading some authors to conclude that polymorphism at this locus is functionally silent and exists only because of linkage disequilibrium with selectable HLA alleles. This review presents a new analysis of the available data which suggests that polymorphism in the TNF-alpha promoter is not randomly distributed and therefore that it most likely does have some functional and selectable effect. Further, a comparison of available data suggests that there is more consensus in the literature than may at first appear to be the case.