Damselfish face climate change: Impact of temperature and habitat structure on agonistic behavior

temperature by a thermostat. Fish were individually held in isolated tanks to avoid physical confrontation and damages. One system (12 tanks) was kept at 28˚C, the average water tem- perature on the Brazilian coastal reefs where animals were sampled [27], and the other (12 tanks) at 34˚C, the expected warming of tropical oceans projected to occur by the end of this century [24]. Since 34˚C is considered the long-term thermal limit for several species of reef fish [17], it represents the worst global warming scenario. To reach 34˚C, tanks at 28˚C were subjected to a 0.5˚C temperature increase every 2 hours for 24h. Tanks were also enriched or kept clean to provide a complex or barren habitat, respectively. The high-complexity habitat consisted of covering the walls and bottom of the tank with wallpaper simulating marine gravel substrate, and including a shelter (6 x 6 x 15 cm hollowed brick) and plastic plants in the tanks. The barren (low-complexity) habitat contained none of the aforementioned items and the tank was kept completely clean. Thus, four groups were formed: “complex habitat at 28˚C” (28C group–n = 12); “complex habitat at 34˚C” (34C group–n = 06); “barren habitat at 28˚C” (28B group–n = 9); and “barren habitat at 34˚C” (34B –n = 9). Fish were kept in these condi- tions for 1 month before the behavioral tests. They were fed twice a day ad libitum with frozen Artemia salina, shrimp paste and dried food pellets (algae-based tetra marine salt granules). When any type of disease/injury was observed during the 30- day period or when fish stopped

feeding for more than 5 days, fish were excluded from the behavioral test, resulting in groups with different sample sizes. A total of twelve fish were excluded from the tests. Following the research data collection, all animals used were euthanized using clove oil anesthetic. All animal procedures were authorized by the Animal Ethics Committee of the Federal University of Rio Grande do Norte (CEUA 100.12/2018). 2.2 Behavioral tests The experimental tanks (40 x 20 x 25cm, 15L) were filled with water under the same conditions and temperature as the stocking systems where each group was maintained, and an air stone provided constant aeration. The tanks were covered with white paper to prevent the fish from having any contact with the outside environment, but the right and the front walls remained uncovered for experimental purposes. A white partition was placed in front of the right wall, but could be removed when needed, allowing the fish to see a mirror positioned at 45˚ [28– 31]. In this position, one corner of the tank (Q1) was closer to the mirror than the other (Q2). On the opposite sides of the mirror (Q3 and Q4) were areas of less interactivity with the mirror image, where less responsive animals were expected to remain longer (Fig 1). The uncovered front wall allowed a camera (SONY1 DCR-SX45) to record fish behavior (camera positioned 50 cm from the tank), while another camera (SONY1 DCR-SX45) was placed 1m above the tank in order to record fish movements. Fish were not fed during behavioral tests.

Fish from the 4 stocking conditions (28C, 34C, 28B, 34B) were individually placed in the center of experimental tank, and after a 2- min acclimation period, behavior was recorded from the overhead and frontal cameras. During the first 5 min of recording, the white partition prevented the fish from seeing the mirror. The partition was then removed and fish behavior was recorded for another 5 min. The animals were then returned to the stock tanks. Behavior recorded by the frontal camera was visually tracked to identify aggressive displays directed toward the mirror. The number and type of displays were quantified. Attacks where considered when the fish approached the mirror from the front or the side quickly and/or suc- cessively, opening its mouth and trying to bite the side of the tank in contact with the mirror, and Threats when it erected its dorsal, pelvic and anal fins close to the mirror. Immobility was considered when the fish remained still for 2s or more, Vigilance when it moved through the tank fins down, and Substrate nibbling when it bit any possible particle at the bottom of the PLOS ONE Temperature and habitat structure on damselfish behavior PLOS ONE | https://doi.org/10.1371/journal.pone.0235389 June 30, 2020 3 / 13