This dataset describes the size distribution of Pocillopora colonies used in the experiment.

We examined the effects of habitat complexity and timing of arrival in the context of the survival of focal Thalassoma quinquevittatum settlers. The study was conducted in the northern lagoon of Moorea between April and June 2008. We used an array of 30 isolated live-coral patch reefs separated by *10 m in water 2–4 m deep. Reefs were located within a sand-flat, separated from each other, and from nearby natural reefs, by a minimum of 15 m. We constructed reefs to minimize habitat variation by standardizing size, rugosity, and water depth. Each reef consisted of a base of live Porites lobata coral with an average area of 2.23 m2 (SD = 0.56), and a mean height of 0.59 m (SD = 0.10). We controlled habitat complexity by manipulating the availability of the branching coral Pocillopora verrucosa. This was achieved by drilling holes into the upper surface of patch reefs. Into these holes, we inserted stainless steel pins attached to P. verrucosa colonies with Z-Spar Splash Zone Compound (Kopcoat, Pittsburgh, PA, USA). Mean colony surface area was 0.2 m² (SD = 0.07). We crossed the availability of P. verrucosa (two levels: two, or four colonies) with the presence of three tagged T. quinquevittatum competitors (three levels: absent, introduced simultaneously with (0 days), or 5 days earlier than the focal individuals). To each reef, we simulated settlement by introducing three tagged T. quinquevittatum focal individuals. Thus, our design had six treatments: (1) focal individuals without competitors, with two P. verrucosa colonies; (2) focal individuals and competitors introduced simultaneously, with two P. verrucosa colonies; (3) focal individuals with competitors introduced 5 days previously, with two P. verrucosa colonies; (4) focal individuals without competitors, with four P. verrucosa colonies; (5) focal individuals and competitors introduced simultaneously, with four P. verrucosa colonies; and (6) focal individuals with competitors introduced 5 days previously, with four P. verrucosa colonies. We ran the experiment in two temporal blocks (17–23 April and 1–7 May 2008), yielding ten replicates (five in each temporal block) for each of the six treatments. We surveyed reefs twice daily (approximately 0800 and 1600 hours) for 5 days after the introduction of focal individuals.