Alleles conferring higher mutation rates (mutators) can fix in asexual populations through a process called ‘mutator hitchhiking’. Theory predicts that repeated mutator hitchhiking can occur in an adapting asexual mutator population. I tested this prediction in two settings: a mutL- population under lethal selection and a mutS- population under soft selection. In both experiments, the starting mutation rate was 100-fold higher than wild-type. In the lethal selection experiment, two replicate populations were exposed to a sequence of three different antibiotics. In both replicates, all survivors sampled after the final antibiotic exposure had undergone further genomic mutation rate increases. Whole-genome and Sanger sequencing revealed that an identical spontaneous 1-bp insertion in mutT (a known mutator gene) rose to probable fixation in both populations. Complementation tests demonstrated that the mutT- defect was responsible for the increased mutation rates. In the soft selection experiment, 30 isogenic populations were propagated in limited glucose media. After 900 generations, five clones were isolated from each population. Relative to the ancestor, 9% of the clones had increased mutation rates and 68% had unchanged mutation rates; surprisingly, 23% had decreased mutation rates. Most populations (21 of 30) had at least one clone whose mutation rate either increased or decreased. One population exhibited apparent fixation for a mutator and one other population exhibited apparent fixation for an antimutator. Some of the sequenced clones with altered mutation rates had mutations in known (anti)mutator loci. I conclude that the mutators likely arose by hitchhiking and that the antimutators likely confer pleiotropic direct fitness benefits. Competitions between the evolved clones and the ancestor demonstrated that all clones and populations had increased in fitness since generation 0. No relationship was detected, however, between mutation rate and relative fitness. These experiments provide evidence supporting the prediction of repeated mutator hitchhiking. More broadly, the work described in this dissertation reveals multiple ways in which the mutation rates of asexual populations may be evolutionarily unstable, with potential implications for evolving asexual systems, including infectious agents and cancer.