Wildlife fencing is the most important mitigation measure for reducing road mortality. While more effective measures exist (e.g., placing roads in tunnels underground or raising roads on pillars), they are often unrealistic. Less expensive measures (e.g., wildlife warning signs and reflectors) have been shown to be ineffective, while wildlife fences are highly effective at reducing roadkill. They are often used in combination with wildlife passages. In contrast, wildlife passages alone (i.e., without fencing) do not reduce road mortality (Rytwinski et al. 2016). Mortality-reduction graphs serve to prioritize road sections for fencing, following an adaptive fence-implementation plan (Spanowicz et al. 2020). If fences were 100% effective, fencing many short road sections would require less total length of fencing to achieve the same predicted reduction in road mortality than fencing a few long road sections. However, in reality, animals frequently move around the fence ends. Therefore, such a “fence-end effect” makes short fences less effective than the use of a few long fences, because animals moving along a long fence are more likely to change course before arriving at the fence end. This trade-off between the total number of fence sections and the length of the fence sections has been called the FLOMS (Few-Long-Or-Many-Short) fences trade-off (Spanowicz et al. 2020). When considering the fence-end effect, how long is long enough for a fence that can be expected to be effective? To address this question, we present a novel analytical model for predicting the fence-end effect as a function of fence length (L). We consider four variations of the model and compare the predictions with empirical data. In the simplest form of the models, effective fence-length is Leff = L – R in Model A, Leff = L – 0.5 R in Model B, Leff = L – 0.4521 R in Model C, and Leff = L – 0.226 R in Model D, where R is the radius of the home range of the target species. Accordingly, the probability of fence success is PFS = 1 – R/L in Model A, PFS = 1 – 0.5 R/L in Model B, PFS = 1 – 0.4521 R/L in Model C, and PFS = 1 – 0.226 R/L in Model D. We use these four models to predict the minimum length of wildlife fencing that can be expected to be effective for various species. We also present modifications to these models for fences that are not properly maintained. The models are included in the mortality-reduction graphs to predict the effectiveness of any amount of fencing at reducing wildlife mortality and to help planners design effective and efficient configurations of fencing. – Cited references: Rytwinski et al. (2016): How effective is road mitigation at reducing road-kill? A meta-analysis. PLoS ONE 11(11): e0166941. doi: 10.1371/journal.pone.0166941; Spanowicz et al. (2020): An adaptive plan for prioritizing road sections for fencing to reduce animal mortality. Conservation Biology 34(5): 1210-1220. doi.org/10.1111/cobi.13502.