This datasets records the background fish community at the start of each experimental run.

We constructed an array of 30 live-coral patch reefs and used these to conduct a field experiment that examined competitive asymmetry between bird wrasse, fivestripe wrasse and sixbar wrasse. On our constructed reefs, we aimed to minimize habitat variation by standardizing the reefs’ size, rugosity and water depth. To achieve this, we selected natural reefs (based upon a set of morphological attributes that included a base of live Porites lobata coral with a surface area [mean ± SD] of 2.23 ± 0.56 m2, and a height of 0.59 ± 0.10 m) from a nearby location and transplanted them to our study site (17°29.010’ S, 149° 50.346’W), an open sand flat 2 to 4 m deep. Each reef was separated from its nearest neighbor and other non-experimental reefs by a minimum of 10 m. To each reef we attached 3 similar-sized colonies (colony surface area = 0.2 ± 0.07 m2) of the branching coral Pocillopora verrucosa using Z-Spar Splash Zone Compound (Kopcoat).

Prior to starting the experiment, we removed all bird wrasse, fivestripe wrasse and sixbar wrasse from the reefs and manipulated the relative abundances of other resident fish species via selective removals and additions so that the relative abundance of all species was similar among the 30 reefs. To each reef, we randomly assigned 1 of 6 treatments: (1) 6 bird wrasses; (2) 6 fivestripe wrasses; (3) 6 sixbar wrasses; (4) 3 bird wrasses and 3 fivestripe wrasses; (5) 3 bird wrasses and 3 sixbar wrasses; or (6) 3 fivestripe wrasses and 3 sixbar wrasses.

We ran the experiment in 2 temporal blocks (21 to 25 May 2008 and 2 to 6 June 2008), yielding 10 replicates for each of the 6 treatments, with treatments randomly assigned in each temporal block. We surveyed reefs twice daily (approximately 08:00 and 16:00 h) for 5 d after the introduction of fishes. During surveys, we searched neighboring non-experimental reefs for tagged immigrants. We found no immigrants or emigrants.

Surveys were conducted the day immediately before deployment of fishes in each temporal block, with 30 reefs surveyed on 20 May 2008 and 24 reefs surveyed on 1 June 2008.