[Truncated abstract] The design of effective immunotherapies, using tumour antigens to stimulate a functional effector cytotoxic T cell (CTL) response in a tumour bearing host, requires an understanding of the 'real time' in vivo relationship between the host immune system and antigens expressed by the developing tumour. However, effector function of endogenous anti-tumour CTLs generated during tumour progression has largely been assessed by indirect ex vivo assays and often focused on a single antigen. Therefore, studies in this thesis evaluated the endogenous in vivo CTL response to multiple tumour antigenic epitopes in murine tumour models using Lewis lung carcinoma cells transfected with ovalbumin (an antigen that contains several intra-molecular MHC class I epitopes with a defined hierarchy) or a polyepitope (that contains a string of immunodominant MHC class I epitopes). Potent effector CTLs were generated to multiple dominant tumour antigenic epioptes early in tumour progression. However, in general, these CTL effectors only transiently retarded tumour growth, and at the later time points of tumour growth they were no longer generated in tumour draining lymph nodes. This coincided with diminished tumour antigen presentation in the same nodes which was found to be due to antigen loss. In both models antigen loss was the result of two processes; immuno-editing of the tumour by the host immune response and genetic instability resulting in antigen loss variants that could evade immune surveillance. A third model was generated that maintained low level tumour antigen expression throughout tumour progression. ... The impact of pre-existing endogenous dominant-epitope specific CTLs on tumour expressing the same epitope was also assessed, and resulted in a reduced tumour incidence and a CTL response restricted to a single antigen of the same MHC allele. Finally, the effects of two different immunotherapy regimens were examined. Intratumoural IL-2 treatment enhanced pre-existing CTL responses to the dominant epitopes leading to tumour regression. In addition, use of a multiple peptide vaccination regimen that avoided T cells competing for peptide-MHC complexes on APC was far more likely to be effective than one that did not. These results demonstrate that immunotherapies targeting tumours that express several dominant neo antigenic epitopes can be effective. The caveat for this approach is that it will only be effective in tumours that have generated an endogenous CTL response and must be used before antigen loss variants emerge.