The amyloid β (Aβ)-protein and microtubule-associated protein, tau, are the major components of the amyloid plaques and neurofibrillary tangles that typify Alzheimer's disease (AD) pathology. As such both Aβ and tau have long been proposed as therapeutic targets. Immunotherapy, particularly targeting Aβ, is currently the most advanced clinical strategy for treating AD. However, several Aβ-directed clinical trials have failed, and there is concern that targeting this protein may not be useful. In contrast, there is a growing optimism that tau immunotherapy may prove more efficacious. Here, for the first time, we studied the effects of chronic administration of an anti-tau monoclonal antibody (5E2) in amyloid precursor protein transgenic mice. For our animal model, we chose the J20 mouse line because prior studies had shown that the cognitive deficits in these mice require expression of tau. Despite the fact that 5E2 was present and active in the brains of immunized mice and that this antibody appeared to engage with extracellular tau, 5E2-treatment did not recover age-dependent spatial reference memory deficits. These results indicate that the memory impairment evident in J20 mice is unlikely to be mediated by a form of extracellular tau recognized by 5E2. In addition to the lack of positive effect of anti-tau immunotherapy, we also documented a significant increase in mortality among J20 mice that received 5E2. Because both the J20 mice used here and tau transgenic mice used in prior tau immunotherapy trials are imperfect models of AD our results recommend extensive preclinical testing of anti-tau antibody-based therapies using multiple mouse models and a variety of different anti-tau antibodies.