Infanticide, related to a stallion’s aggression toward a foal sired by another stallion, and feticide related to a new stallion’s aggression and/or pheromonal influence (the Bruce effect) inducing loss of a fetus sired by another stallion, a female’s counteraction to infanticide, have been proposed for domestic horses (Equus caballus) in human-managed conditions. The aim of the present study was, in conditions close to natural, to investigate the influence of the natural succession of a harem stallion on the mares’ subsequent reproductive performance. In a population of semi-feral Konik polski horses observed for 31 years (reproductive seasons) in 8 bands, harem stallion changed 10 times. These changes involved 26 out of 48 mares and 60 out of 609 observed mare-seasons (MS, a year in which a mare experienced a reproductive event). Binary distribution and log link function were assumed. The marginal model included the classification variable (SCH) and the continuous variables (age of the mare and calendar year of reproductive event (birth of a live foal, abortion, foals lost or barrenness) in a given MS was analyzed with generalized linear mixed model. The reproductive fitness of mares and their reproductive success (foal surviving ≥ 1 year), did not differ between MS with and without SCH. Older females were more likely to stay barren, with chances increasing by 21% with each successive year; and less likely to give birth to a foal (13% decrease of chance), and rear a foal to one year of age (12% decrease of chance). The age did not affect the probability of abortions. Of the 26 MS when mares were pregnant when the stallion had changed, there were 25 healthy foals born. For the entire 31 years of monitoring, no aggression toward any foal was observed and all foals that were born in the harem of a new, succeeding stallion successfully reached adulthood. Due to the lack of incidents of infanticide and the lack of evidence suggesting that the presence of a new harem stallion leads to the termination of pregnancies sired by another stallion, the Bruce effect was not confirmed as a biological strategy to reduce investment in pregnancy and potential infanticide in studied population of semi-feral horses.Copyright © 2020 Elsevier Inc. All rights reserved.