Fig. 3.

(Top) Percentage change in mass (mean ± SE) for the corals Acropora millepora (A) and Porites cylindrica (B) over two months (January–March 2014) of contact by differing densities of Sargassum polycystum. (Middle) Percentage change in mass (mean ± SE) during March–September 2014 for Acropora (C) and Porites (D) previously exposed to different densities of surrounding Sargassum, but with no Sargassum present during this period of growth assessment. (Bottom) Total mass change (g) (mean ± SE) during January–September 2014 for Acropora (E) and Porites (F) initially exposed to different densities of Sargassum for two months (January–March 2014), but with Sargassum then removed and absent for the next 7 months (March–September 2014). For all graphs, data for each species were analyzed by sequentially testing nested GLS models via likelihood ratio tests to obtain the optimal fixed structure for each model. P-values from these analyses are presented above each figure. D = inclusion of Sargassum density as a fixed term; A = inclusion of area (coral- or macroalgal-dominated area) as a fixed term. Numbers within bars indicate sample size.