This study investigates the effects of fitness changes on hippocampal microstructure and hippocampal volume. Fifty-two healthy participants aged 59-74 years with a sedentary lifestyle were randomly assigned to either of two levels of exercise intensity. Training lasted for six months. Physical fitness, hippocampal volumes, and hippocampal microstructure were measured before and after training. Hippocampal microstructure was assessed by mean diffusivity, which inversely reflects tissue density; hence, mean diffusivity is lower for more densely packed tissue. Mean changes in fitness did not differ reliably across intensity levels of training, so data were collapsed across groups. Multivariate modeling of pretest-posttest differences using structural equation modeling (SEM) revealed that individual differences in latent change were reliable for all three constructs. More positive changes in fitness were associated with more positive changes in tissue density (i.e., more negative changes in mean diffusivity), and more positive changes in tissue density were associated with more positive changes in volume. We conclude that fitness-related changes in hippocampal volume may be brought about by changes in tissue density. The relative contributions of angiogenesis, gliogenesis, and/or neurogenesis to changes in tissue density remain to be identified.