Plasmodium falciparum malaria is one of the leading causes of child mortality worldwide, and currently there is no vaccine which confers sterilising immunity against infection. One approach to reducing the morbidity and mortality associated with P. falciparum malaria would be the development of a vaccine which targets P. falciparum virulence phenotypes, thereby protecting against the severest forms of the disease. Rosetting, where erythrocytes parasitised with P. falciparum bind to uninfected erythrocytes, is a well-established virulence phenotype that could serve as such a target. Although the major family of parasite proteins responsible for rosetting, Plasmodium falciparum erythrocyte membrane protein 1 (PfEMP1), displays a staggering degree of antigenic polymorphism, there is some evidence to suggest that functionally related PfEMP1 subsets are relatively conserved. Therefore, it is possible that antigenic conservation within the rosette-mediating subset of this family exists. In this thesis, I aim to determine whether rosette-mediating PfEMP1 variants are sufficiently conserved to be considered viable vaccine targets. Firstly, I used rosetting Kenyan culture-adapted P. falciparum lines to characterise four novel rosette-mediating PfEMP1 variants via var gene expression profiling, and tested the ability of recombinant protein domains from these variants to bind to uninfected erythrocytes. I then examined the relationship of the var gene sequences encoding these new variants to those of known rosetting variants. Of the four newly characterised rosetting variants, three were not characteristic of rosetting variants. One was from a PfEMP1 subset which would be predicted to bind to endothelial protein C receptor, and two had a previously uncharacterised binding phenotype. This surprising discovery suggests that multiple different subsets of PfEMP1 can mediate rosetting, and some potentially have dual binding phenotypes strongly linked to severe malaria. Secondly, I investigated the existence of naturally acquired, broadly strain-transcending antibodies to rosetting parasitised erythrocytes in individuals living in malaria endemic countries. To do this, I used a method for the elution of intact IgG from human plasma from the surface of parasitised erythrocytes and transfer to heterologous parasite lines. I found that individuals do raise strain-transcending antibodies to rosetting parasite lines, suggesting the presence of conserved epitopes. Importantly, this strain-transcending recognition was seen with allopatric plasma-parasite pairs, implying that the conserved epitopes are present in geographically diverse regions. Lastly, I set out to determine whether rosetting clinical P. falciparum isolates from Uganda are recognised by polyclonal IgG raised in rabbits to rosetting PfEMP1 variants expressed by laboratory lines derived from Southeast Asia, Central America and Kenya. To do this, I undertook a study to collect P. falciparum isolates from children with malaria attending Mbale District Hospital in eastern Uganda, and then tested these isolates for recognition with the PfEMP1 antibodies by flow cytometry. I found that some of the IgG do cross-react with certain clinical isolates, suggesting the presence of conserved epitopes, but these were not universally present in all rosetting isolates. In summary I have found that, whilst rosette-mediating PfEMP1 variants may be diverse, individuals do raise strain-transcending antibodies to rosetting PEs, and conserved epitopes can be detected on PfEMP1 variants displayed by rosetting clinical isolates. These findings suggest that several distinct PfEMP1 subsets can mediate rosetting, and that rosetting PfEMP1 variant types linked to severe malaria merit further investigation as vaccine candidates.