Twelve small colonies each of Montipora monasteriata, Acropora muricata, and Pocillopora damicornis were collected at 4-5m depth from the reefs at Heron Island, Queensland, Australia (23.4423°S, 151.9148°E) and placed in treatment tanks.
Daily temperature was recorded using Hobo temperature loggers every minute within the treatment tank. Chl a, total soluble lipid, soluble animal protein, and soluble animal carbohydrate concentrations, were each measured on a 1cm2 cored plugs of M. monasteriata and from 1cm long branch tips of A. muricata and P. damicornis from each ramet. Each measurement was made on whole coral samples (skeleton, animal tissue, and endosymbiotic algae) that were ground with a mortar and pestle and normalized to total ash-free dry tissue biomass of the organic fraction (animal tissue and endosymbiotic algae). Chl a was extracted using methods modified from Jeffrey and Humphrey (1975). Total soluble lipids were extracted using methods described in Rodrigues and Grottoli (2007), while soluble animal carbohydrate and protein concentrations were measured using the methods modified from Dubois et al. (1956) and Smith et al. (1985), respectively, as described in Levas et al. (2018). Biomass was measured according to methods outlined in McLachlan et al. (2020).
Coral fragments were airbrushed to remove all tissue from the skeleton. The host tissue and endosymbionts were separated by centrifugation and filtered onto prebaked GF/F filters. Animal host tissue and endosymbiotic algal fraction δ15N values (δ15Nh and δ15Ne, respectively) were reported relative to air (δ15N = per mil deviation of the ratio of stable nitrogen isotopes 15N:14N relative to air). Animal host tissue and endosymbiotic algal fraction δ13C values (δ13Ch and δ13Ce, respectively) were reported relative to Vienna Peedee Belemnite Limestone standard (δ13C = per mil deviation of the ratio of stable carbon isotopes 13C:12C relative to V-PDB). Repeated measurements of internal standards (n = 20) had a standard deviation of ± 0.14‰ for organic δ15N and ± 0.07‰ for organic δ13C. δ15N and δ13C values were determined using a Costech Elemental Analyzer where the resulting N2 and CO2 gases were analyzed for δ15N and δ13C with a ThermoFisher Delta IV stable isotope ratio mass spectrometer (IRMS) via a Conflo II interface in the Grottoli lab at the Ohio State University.