Methods overview: Experiments were conducted in Palau, Micronesia (07o 30’ N, 134o 30’ E) in the northwest Pacific. We collected three regionally common branching coral species, Acropora digitifera, Acropora gemmifera and Anacropora spinosa, from the inner reef flat of south Airai fringing reef between two and five metres water depth. We had two adult treatments (adult absent and present) and four larval stocking densities (10, 50, 100 and 200). We established three replicate aquaria per adult treatment, and three replicate containers per stocking density (n = 9) per aquaria. Larvae of Anacropora spinosa were insufficient for full replication, so replicates were reduced for stocking densities to 6 replicates for 10-100, and 3 replicates for 200. Replicate containers had one conditioned settlement tile. Settlement was scored (number of successful settlers versus number unsettled) four days after larvae were added to the replicate containers. Settlers were counted visually on all tile orientations (top, bottom and sides) using a dissection microscope. Following the settlement census, the tiles were placed back into their respective aquaria on elevated plastic crates. Survival was then further censused weekly for two-time points (number survived versus number from prior time point): 7 days and 14 days post-settlement; settlement is represented as day zero. Settler spatial patterns were analysed considering the whole three-dimensional surface of the tile. These data are provided in the Supplementary as Tables. Whether settlers were touching other settlers (in direct contact) or not was assessed as probability of forming aggregates (1 = settler touching, 0 = settler not touching) against the number of total settlers at each time point (7 and 14 days). A settler was considered touching another settler if, at least, tissue margins were touching. All analyses were conducted in R. All analyses run and R packages used provided in manuscript.

Carrie Sims was supported during this research by an Australian Government Research Training Program Scholarship and a Westpac Future Leaders Scholarship, which funded the research trip. Funding for this research was provided by an ARC Centre of Excellence for Coral Reef Studies grant to John Pandolfi and others (CE140100020). Margaret Mayfield was supported by an ARC Future Fellowship (FT140100498).