An extensive investigation into and charactaisation of factors influencing transgene expression following introduction of the transgoie into tobacco via Agrobacterium- mediated transformation was carried out. Characterisation of material supplied at the outset of this project revealed that this material was unacceptable for further analysis. It was thus deemed necessary to obtain large populations of transgenic tobacco heterogenous for levels of transgene expression. Characterisation of these populations (CaMV-lecA and ssRubisco-lecA plants) showed that all plants fell into one of four segregation classes based on segregation of the kanamycin-resistance selectable marker. Results showed that the majority of regenerants contained multiple nptII-containing inserts, while the presence of one or two such inserts was also found, albeit at a much lower frequency. Segregation analysis based on detection of the lecA transgene agreed, in the majority of cases, with these results. However, in a few cases it was found that data obtained from both segregation analyses did not agree, with the presence of a single lecA-containing transgene being detected in plants shown to contain two copies of the nptII-contaning transgene. This result indicates the occurrence of T-DNA rearrangement either within the tobacco genome or during T-DNA transfer and integration. Southern blot analyses allowed a detailed characterisation of T-DNA structure, copy number and number of integration sites to be undertaken. Results from these analyses revealed a higher frequency of T-DNA rearrangement within plants containing multiple inserts. However, such rearrangements did not correlate with a significant reduction in levels of transgene expression since all detected rearrangements were found to occur at or towards the left hand border of the T-DNA, that border distant to the lecA transgene. Plants containing more than one T-DNA were also frequently found to contain these T-DNAs arranged as an inverted repeat at a single locus although no significant relationship between copy number and the presence of such structures was found. Correlating transgene expression levels, as determined by radioimmunoassay-based quantitation of lectin protein in tissues of transgenic plants, with T-DNA copy number, organisation and structure revealed no significant relationship. It is thus feasible to conclude that the major contributory factor influencing levels of transgene expression is the location of T-DNA integration within the plant genome. Subsequent work concerned with investigating the nature of those integration site-specific factors i.e. 'position effect' indicated a possible role for methylation-induced modulation of gene expression. Results presented in this thesis provide an insight into the fate of transgenes following introduction into the plant genome and clearly demonstrate the importance of further exploring the molecular mechanisms underlying transgene expression variability.