Development of a mare-foal attachment index and its predictive value for subsequent mare response to artificial weaning Abstract Artificial weaning - before and more abruptly than under natural conditions - is common in the equine industry yet research on factors influencing mare welfare at weaning is limited. Mare-foal attachment is key for foal development, but its importance for mare welfare has been overlooked. Therefore, we investigated whether post-natal and pre-weaning mare-foal attachment influenced mares' welfare at weaning. We recorded foal-directed behaviours (e.g. licking, sniffing, distance to foal) of 11 mares post-partum (video) and pre-weaning (live); then mare behavioural responses (eg locomotion) at weaning (live). We also recorded mare age and foal sex to assess their impact on mare behaviour. Principal Component (PC) Analysis of post-foaling and pre-weaning observations identified three attachment indices explaining 73% of the variation: mare-foal proximity/protection (PC1), post-foaling nursing/olfacto-tactile behaviour (PC2), and pre-weaning nursing/agonistic interactions (PC3). None were influenced by mare age or foal sex, and only PC3 was negatively correlated with mare locomotion at weaning. Our preliminary study shows mare-foal attachment affects mare responses at weaning, however further investigation with a greater study population is needed to replicate these results, which may allow to develop weaning methods mitigating the welfare impact of abrupt and early separation of foals on their mares.