The effect of CpG islands on transgene expression was first tested in cultured cells. In transient transfection it was demonstrated that CpG islands do not influence the expression of a transgene when not integrated into the genome. Even when integrated into the genome of cultured cells, CpG islands are not able to confer position-independent, copy number-dependent transgene expression, as confirmed by the analysis of individual cell lines. However, the results from bulk analysis of primary clones suggest that CpG islands improve the level of expression in cultured cells, and increase the proportion of highly expressing clones. Transgenic mice were used to study the effect of CpG islands on the level and pattern of transgene expression in vivo. Unexpectedly, from the nine transgenic lines generated, transgene expression was detected in only one line. In the rest of the lines transgene expression was silenced, and in these cases the transgene was densely methylated. In half of the silenced lines transgenes were found to localise in the pericentromeric chromatin. The results suggest that full size CpG islands used as promoters do not necessarily overcome the negative effects of neighbouring chromatin to give ubiquitous transgene expression independent of the integration site. During the study of stable cell lines it was observed that cells within a clone do not have the same level of transgene expression, and this heterogeneity in expression was not reduced by the presence of the full size CpG island. In order to elucidate this phenomenon, the hypothesis that intraclonal heterogeneity is caused by individual cells switching transgene expression on and off over a period of time was tested. Clones of cells expressing Green Fluorescent Protein (GFP) were monitored at regular time intervals, and in some cells rapid extinction of fluorescence was observed. It was concluded from this result that transgene expression varies with time.