2.3.2. Philippines

Unfed juvenile gnathiids were given one day to acclimatize in an aquarium after collection, before host fish, *Dascyllus trimaculatus* (Pomacentridae) and various species of Labridae, were placed in the aquarium overnight to allow them to feed. The gnathiids did not feed again for the duration of the experiment. The following day, fed, mobile and healthy-looking individuals were selected. However, as the gnathiids' species and therefore the consequent size range for each stage was not known, they could not be separated by juvenile stage as in the GBR study. Instead they were sorted into two size classes (<2 mm and >2 mm, to account for any effect of size of the gnathiid on its molting rate and survival.

Gnathiids were then placed in 270 mL plastic containers, 5–10 gnathiids per container, with plankton mesh (55 μm) secured on the top, and the containers were submerged in one of five 27 L aerated aquaria, each with a different set temperature. Each container was labelled with the size class, treatment, trial and replicate. The first trial consisted of 5 temperature treatments, ambient (28 ◦C), 30 ◦C, 32 ◦C, 34 ◦C and 36 ◦C. A second trial was also conducted to obtain a finer resolution of the temperature effect, with treatments of 30 ◦C, 32 ◦C, 33 ◦C, 34 ◦C and 35 ◦C. Aquaria were individually heated gradually with 100W and 200W aquarium electric heaters (Venusaaqua®) over a 10 h period to their desired temperature. Temperature readings were taken daily with an aquarium-mounted digital thermometer (Doutop®) to ensure the desired water temperatures were maintained and to calculate the average temperature for each treatment per trial (Table S1a,b). Containers from each treatment were inspected daily for evidence of changes in gnathiid development and mortality. Dead gnathiids were removed from the containers and molted adult males were preserved in ethanol for future species identification. The experiment was concluded for each treatment when all gnathiids were dead or when 20 d had passed. One third of the water in each aquarium was removed daily and replaced with ambient temperature, filtered, fresh seawater. Over the 20 d duration of each replicate the aquaria used for each temperature treatment was alternated every week (once per replicate) to ensure there were no confounding effects associated with individual aquaria. Three replicates of each temperature treatment were run for each trial. Trials 1 (*n* = 369 gnathiids) and 2 (*n* = 318 gnathiids) had a range of 60–76 and 59–68 gnathiids per temperature treatment respectively (Table S2a,b).

The baseline ambient temperatures for Trials 1 (28 ◦C) and 2 (30 ◦C) were similar to the SST in the Bohol Sea, Philippines, which fluctuated by about 3 ◦C (about 27–30 ◦C) during 2017; the SST during the experiment (July–October 2017) averaged at 30 ◦C ± 0.04 SE. [109].
