31Ant guards can increase plant fitness by deterring herbivores but may also 32 reduce it by interfering with pollination, hence ant-plant interactions are ideal 33 systems in which to study costs and benefits of mutualisms. While ant impacts 34 on herbivory are well-studied, much less is known about impacts on 35 pollinators and associated consequences for plant mating systems and fitness. 36 We used field experiments to quantify the effect of ant guards on pollinator 37 community composition, frequency and duration of flower visits, and 38 cascading effects on plant mating system and plant fitness in Turnera velutina 39 (Passifloraceae). Although ant patrolling did not affect pollinator community 40 composition or visitation frequency, it decreased pollinator foraging time and 41 flower visit duration. Such behavioural changes resulted in reduced pollen 42 deposition on stigmas, decreasing male fitness whilst increasing outcrossing 43 rates. This study contributes to understanding how non-pollinators, such as 44 these defensive mutualists, can shape plant mating systems.45 46 87 tested. Alternatively, an increase in fruit set in ant-patrolled plants of 88 Psycothria limonensis was attributed to pollinator relocation, where ant 89 threats might have caused pollinators to spend less time per flower and visit 90 5 more flowers, promoting pollen transfer (Altshuler 1999). Again, this 91 mechanism was inferred, but not experimentally tested. Previous experiments 92 on Turnera velutina showed that ant corpses placed inside flowers reduce 93 pollinator visit duration (Villamil et al. 2018). However, such an experimental 94 setup may differ from natural circumstances as flower occupation by ants is a 95 rare event, and live ants (in contrast to dead ones) do not remain immobile in 96 the flowers for long periods. Overall, the presence of ants can promote 97 changes in pollinator community composition, visit frequency and duration. 98 This, in turn, could drive positive or negative impacts of ant-pollinator 99 interactions on plant reproduction (Altshuler 1999; Ness 2006). To date, no 100 study has quantified the impacts of ant patrolling on pollinator visitation 101 behaviour, plant mating systems, and fitness under natural conditions. 102 We estimated the ecological and potential evolutionary consequences 103 of myrmecophily on the pollination biology, mating system, and fitness of 104 Turnera velutina (Passifloraceae), a self-compatible ant-plant using an ant 105 exclusion field experiment. We addressed the following questions: (i) What is 106 the effect of ant patrolling on pollinator visitation? (pollinator community 107 composition, visitation frequency, duration, and behaviour) (ii) Does ant 108 patrolling affect the host plant mating system? (iii) Does ant patrolling affect 109 pollen transfer dynamics? (iv) Does ant patrolling affect plant male fitness? 110 First, because smaller or solitary pollinator taxa are expected to be more 111 vulnerable to predation risk than larger or social species (Dukas & ...