The concept of Length of Stay (LOS) in animal shelters is critical for managing shelter capacity, improving animal welfare, and assessing and improving shelter operations 1. In research studies, LOS is typically computed for pre-defined animal cohorts tracked from intake to outcome, but this approach is problematic for ongoing operational or temporal assessments of shelters. For a given calendar period, a shelter’s LOS should consider every animal that was resident at any point during that period. We can’t have both a fixed animal cohort and a sharply defined period. For any fixed cohort, by definition, the period we’re examining is diffuse as animals arrive and leave. For studying a sharply defined period, we need to instead allow the cohort to be diffuse, with LOS computations including some animals for only a portion (the beginning, middle, or end) of their stays. When possible, we should obtain the full LOS distribution, and not just an average (e.g., mean or median). The LOS distribution is important for estimating future LOS for already resident animals. In this study, we seek clear definition and rigorous computation of LOS distributions for fixed calendar periods. As an example, we study the LOS of dogs in Orange County Animal Care (OCAC) in California.