The immune system of laboratory mice raised in an ultra-hygienic environment resembles that ofnewborn humans, but can be induced to resemble the immune system of adult humans or 'dirty' mice by co-housing with pet store-bought mice. The laboratory mouse is by far the dominant model organism for in vivo immunological research which — particularly in the light of disappointing results obtained with some recent transfers of disease treatments from laboratory to clinic — raises the question of how accurately the model reflects the human immune system. These authors compare the immune status of laboratory mice with that of feral mice and with mice bought commercially as pets. They find that the immune system of the ubiquitous laboratory 'specific pathogen free' mouse approximates that of human neonates, rather than human adults. Co-housing laboratory mice with 'pet store' mice leads to maturation of the immune system, making it more similar to that of the human adult, and resulting in increased resistance in several models of infection. The use of such 'dirty' mice could supplement current models to either increase translational potential to human disease or to better inform the efficacy of preclinical prophylactic and therapeutic modalities. Our current understanding of immunology was largely defined in laboratory mice, partly because they are inbred and genetically homogeneous, can be genetically manipulated, allow kinetic tissue analyses to be carried out from the onset of disease, and permit the use of tractable disease models. Comparably reductionist experiments are neither technically nor ethically possible in humans. However, there is growing concern that laboratory mice do not reflect relevant aspects of the human immune system, which may account for failures to translate disease treatments from bench to bedside1,2,3,4,5,6,7,8. Laboratory mice live in abnormally hygienic specific pathogen free (SPF) barrier facilities. Here we show that standard laboratory mouse husbandry has profound effects on the immune system and that environmental changes produce mice with immune systems closer to those of adult humans. Laboratory mice—like newborn, but not adult, humans—lack effector-differentiated and mucosally distributed memory T cells. These cell populations were present in free-living barn populations of feral mice and pet store mice with diverse microbial experience, and were induced in laboratory mice after co-housing with pet store mice, suggesting that the environment is involved in the induction of these cells. Altering the living conditions of mice profoundly affected the cellular composition of the innate and adaptive immune systems, resulted in global changes in blood cell gene expression to patterns that more closely reflected the immune signatures of adult humans rather than neonates, altered resistance to infection, and influenced T-cell differentiation in response to a de novo viral infection. These data highlight the effects of environment on the basal immune state and response to infection and suggest that restoring physiological microbial exposure in laboratory mice could provide a relevant tool for modelling immunological events in free-living organisms, including humans.