For cementless femoral stems, there is debate as to whether a collar enhances primary stability and load transfer compared to collarless designs. Finite Element (FE) analysis has the potential to compare stem designs within the same cohort, allowing for subtle performance differences to be identified, if present. Subject-specific FE models of intact and implanted femora were run for a diverse cohort (21 males, 20 females; BMI 16.4-41.2 kg/m , age 50-80 yrs). Collared and collarless versions of Corail (DePuy Synthes, Warsaw, IN) were sized and positioned using an automated algorithm that aligns the femoral/stem axes, preserves the head-center location, and maximizes metaphyseal fit. Joint contact and muscle forces simulating peak forces in level gait and stair climbing and were scaled to the body mass and applied to each subject. Three failure scenarios were assessed: Potential for peri-prosthetic fibrous tissue formation (stem micromotion), potential for peri-prosthetic bone damage (equivalent strains), and calcar bone remodeling (changes in strain-energy density). Comparisons were performed using paired t-tests. Only subtle differences were found (mean 90th percentile micromotion: Collared = 86 µm, collarless = 92.5 µm, mean 90th percentile interface strains: Collared = 733 µϵ, collarless = 767 µϵ, and similar remodeling stimuli were predicted). The slight differences observed were small in comparison with the inter-patient variability. Statement of clinical significance: Our results suggest that the presence/absence of a collar is unlikely to substantially alter the bone-implant biomechanics nor the initial mechanical environment. Hence, a collar is likely to have minimal clinical impact. Analysis using different femoral stem designs is recommended before generalising these findings. © 2017 Orthopaedic Research Society. Published by Wiley Periodicals, Inc. J Orthop Res 36:1185-1195, 2018.