Gallid herpesvirus 2, also known as Marek’s disease virus, is the causative agent of Marek’s disease in chickens that can cause up to 100 % mortality in unvaccinated hosts. Vaccination against MDV is one of the most successful vaccination campaigns in the history of veterinary medicine, reducing disease incidence by more than 99%. Despite this success, MDV is still prevalent in chicken flocks worldwide and has shown a remarkable increase in virulence over the past decades. A major reason for the persistence of MDV could be the fact that vaccination against MD is not inducing sterilizing immunity but is permissive for (reduced) viral replication and shedding. It is argued that the imperfection of vaccination drives viral evolution towards higher virulence by selecting for viral phenotypes that maintain lytic replication and thereby the ability to be shed and transmitted in the presence of vaccine-induced immune response. The phenotypes selected in this way could ultimately benefit from vaccination, as vaccinated chickens which survive the infection shed the most replication competent viruses for a prolonged time, and thus contribute to the spread and evolution of particularly virulent virus strains. As a result, the development of MDV vaccines is caught in a vicious circle – vaccination drives selection of rapidly replicating escape mutants, which requires development of new vaccines based on viral strains that can replicate in the vaccinated host. This scenario has indeed been observed with vaccines of the first and second generation. In the light of these possibilities, the increase in virulence observed during the last decades is undoubtedly alarming. In the context of selection for higher virulence, genetic variation of MDV in vaccinated hosts could provide a selective advantage similar to what is known for some RNA viruses, which have evolved error-prone genome replication and form highly diverse quasispecies. As large DNA virus, MDV is believed to be genetically relatively stable, employing a proofreading DNA polymerase for genome replication. There is, however, evidence for remarkable genetic variation among several large DNA viruses, including herpes viruses such as HCMV and HSV-1. The objectives of this study were 1) to develop a NGS sequencing strategy for this highly cell associated virus 2) to determine, if genetic variation in MDV is a function of the fidelity of its DNA polymerase and 3) to examine replicative fitness and pathogenicity of proofreading-deficient viruses in vivo. Following the development of a tiling array for highly specific capture of viral sequences from infected chicken cell extracts, we were able to sequence whole viral genomes from a variety of samples ranging from infected chicken embryonic cells to dust collected from chicken farms. Next, we constructed MDV mutants with point mutations, in the exonuclease and finger domain of Pol (UL30), that could enhance or reduce replication fidelity. The observed level of residual exonuclease activity correlated with the capacity of mutated viruses to replicate in cell culture. Viruses that encoded a DNA Pol which lacked the majority of its inherent exonuclease activity proved to be suicidal in cell culture, losing their replication fitness within a few passages after reconstitution from BAC DNA. Sequencing of clonal genomes obtained from virus propagated in chicken cells revealed that Pol mutants indeed exhibited higher mutation rates than wild type virus. Following in vitro characterization, three Pol mutants – a hypermutator (Y567F, mutation rate approximately 80-fold higher than WT), a weak mutator (Y547S, mutation rate approximately 3-fold higher than WT) and a putative hypomutator (L755F, mutation rate possibly slightly lower than WT) were examined in vivo. The survival of chickens indicates that a hypermutator phenotype (Y567F) is detrimental for viral pathogenicity while no significant difference between Y547S, L755F and WT was observed. Sequencing of MDV DNA enriched from different chicken tissues showed that this difference in virulence correlates with a higher mutation rate in the Y567F virus. Increasing the mutation rate through reducing MDV Pol fidelity seems to be deleterious for the replicative capacity and fitness of MDV in vitro as well as in vivo through generation of an excessive number of mutations. Nevertheless, the potential of escaping this “error catastrophe” by partial repair of exonuclease function and establishment of a highly diverse viral population with WT like fitness was observed in cell culture for one of the hypermutator mutants (Y567F). The formation of functional and hyperdiverse populations by Pol mutant herpesviruses should be further investigated with special respect to potential quasispecies population dynamics.