Data from: Ectoparasites modify escape behaviour, but not performance, in a coral reef fish

Parasitized (N = 28) and unparasitized (N = 12) Scolopsis bilineata were collected in February and March 2012 using monofilament barrier nets (10 mm stretched mesh) and silicone hand nets from sites around Lizard Island, Northern Great Barrier Reef, Australia (14° 40’ S; 145° 28’ E). Fish were transported in aerated 20 l Handy Pail buckets (maximum 4 fish per bucket) to the Lizard Island Research Station within 1 h of capture, and transferred with silicone nets to individual holding aquaria (40.0W × 29.0L × 18.0H cm) with a flow-through water system transported directly from the reef. Tanks were kept under a natural light and temperature regime (28 ± 1 °C). All fish were provided with a round PVC shelter for refuge. Fish were fed to satiation once a day with raw prawn and fasted for 24h prior to the experiments to ensure fish were in a standardized (post-absorptive) state that maximizes energy availability for swimming. Holding tanks were siphoned out daily to maintain high water quality. All animals were kept in aquaria for a minimum of three days before performing swim trials to ensure all fish were healthy. Fast-start experiments We tested three groups of fish with different infection statuses: unparasitized (fish total body length BL; 13.2 ± 0.8 cm; mean ± SD, N = 12), parasitized (BL; 11.8 ± 1.1 cm; mean ± SD, N = 16) and parasite-removed (BL; 12.7 ± 1.3 cm; mean ± SD, N = 12). Parasites were removed using forceps 24h before the start of the fast-start experiments by holding the fish in a shallow water bath and gently unhooking the isopod. This procedure took approximately 45s. Fish resumed normal behaviour within 2h of being returned to their holding tanks. Experiments were conducted in rectangular acrylic aquaria (70.0 L × 60W × 35H cm) supported by a wooden stand. Water height was maintained at 12 cm for all trials. This height limited movement of the fish in the vertical plane (Langerhans, 2009). A mirror was inclined at a 45 ° angle below the aquarium to allow filming of the escape responses in 3 dimensions within the same camera frame. Videos where vertical movement occurred were excluded from the analyses (< 2 % of videos). The tank was illuminated by three 150 W spotlights positioned 70 cm above the water. A continuous flow of seawater maintained the water temperature at a constant 28 °C. Prior to escape response trials, fish were transferred from their holding tanks to the experimental arena using silicone nets to prevent mucus loss and fin damage, and left undisturbed for 30 min to acclimate to the arena. Numerous aerial predators, including raptors and pelagic seabirds, are commonly seen feeding on fishes around Lizard Island. Thus we used a mechano-acoustic stimulus, which simulates an aerial attack, to induce escape responses in S. bilineata. We attached a 50 ml PVC container filled with lead weights with a string to a platform 30 cm above the water surface. To ensure that the escape response was not initiated prior to the stimulus’ contact with the water, the stimulus fell inside an opaque 15 cm wide PVC tube positioned 1 cm above the water surface. The time of contact of the stimulus with the water surface was clearly visible in the mirror from below. We filmed the responses at 240 Hz with a high speed digital camera (Exilim EX-FH100, Casio, USA) mounted on a tripod directly facing the aquarium and the mirror. Individuals were stimulated up to three times at 30 minute intervals. The stimulus was dropped when the fish was facing the stimulus at a distance of no more than 10 cm from the bottom of the PVC tube. After the trials, individuals were returned to their holding tanks. In total, we filmed 127 escape responses in 40 fish. Two fish developed a bacterial infection several days following experimentation. These individuals were anesthetized with an overdose of Aqui-S solution and then euthanized in an ice-slurry. No other individuals showed any signs of sickness, and care was taken to prevent the spread of disease by rinsing nets in a freshwater bath prior to handling different individuals. All healthy fish (38 individuals) were released back to their site of collection within 1 week of trials. We used the MtrackJ plugin in the ImageJ v. 1.43 software to analyse the escape sequences. We tracked the two-dimensional coordinates of the fish’s centre of mass (CoM) every 4.2 ms starting 21 ms (5 frames) before and ending 84 ms (20 frames) after the onset of the fish’s first movement. The CoM was visually estimated at a proportional distance from the tip of the head corresponding to approximately 29% of an individual’s total length. The following escape performance metrics were measured: response latency (time between stimulus onset and fish response in ms), size- adjusted cumulative distance travelled (D in fish total body lengths; BL), size-adjusted maximum escape speed (Umax in BL s-1) and maximum acceleration (Amax in m s-2). Distance–time variables (D, Umax, Amax) were evaluated within a fixed time period of 58 ms (14 frames), corresponding to the approximate mean duration of stages 1 and 2 of the escape response across all treatments. A five-point moving quadratic polynomial regression was used to obtain smoothed values of speed and acceleration, the first and second derivatives of distance. For each fish, the best value (e.g., highest Umax or shortest latency) of each escape performance variable across the three stimulus presentations were chosen for analysis (see Domenici, 2011; Marras et al., 2011). Flight initiation distance We estimated FID in S. bilineata from the lagoon and adjacent reefs in front of the Lizard Island Research Station on calm weather days from July-August 2013. Water depth varied between 2 and 3 m at these sites with a visibility of approximately 15 m. We used snorkelers as the stimulus for flight initiation. Many studies of FID in terrestrial and aquatic systems have used humans as a stimulus to elicit flight. Recent studies found FID estimates to be relatively robust to variation among observers regardless of experience or, in aquatic systems, whether observations were made on snorkel or SCUBA. Two snorkelers swam around the reef in search of S. bilineata. Only solitary, adult individuals that were foraging or moving slowly over the reef in an open area where they could be approached directly were targeted. Individuals less than 1.5 m from branching corals or other shelter were not approached to avoid the confounding effects of distance to a refuge on FID. Similarly, trials were abandoned if individuals began swimming in any direction at a consistent speed before the observer initiated their approach. Once a suitable individual was spotted, we recorded fish infection status (parasitized or unparasitized, hereafter referred to as treatment) and total length (± 1 cm; actual error). Before data collection, all observers practiced estimating fish length underwater using model fish and objects of various sizes until they reached a precision of ± 1 cm. One snorkeler positioned themselves in direct line of view of the fish at a distance of approximately 5 m. The other snorkeler positioned themselves off to the side to avoid obstructing the trial. The first snorkeler duck-dived under the water until they were close to the substrate (within 1 m), and visually relocated the individual. The snorkeler then approached the focal fish at a quick but steady swimming speed, holding two weights marked with flagging tape beside their head, which was assumed to be the onset of the stimulus. When the fish began to flee (i.e. first began to turn away from the approaching snorkeler), the snorkeler dropped one weight where they were, and took a visual landmark of where the fish had been, which was marked with the second weight. The two snorkelers then measured the horizontal distance between the two landmarks with a tape measure to the nearest 1 cm (FID). The observers also scored the strength of an individual’s reaction, or flight intensity, on a scale from 0 – 4 as follows: 0: no response (i.e. fish did not move in response to the snorkeler), 1: fish ceased previous activities (i.e. foraging) and moved a short distance away, but did not leave the immediate area, 2: fish changed directions and began a slow displacement away from the area, 3: fish changed directions and fled the area at a fast, but constant speed, 4: fish initiated an escape response characterized by a “C-start” unsteady burst behaviour. Increasing intensity was assumed to be related to more energetically costly forms of locomotion, and therefore provided an estimate of the costs of flight. Recent studies suggest that S. bilineata are strongly site attached and rarely travel far from their small territories during the day. Therefore, to avoid pseudoreplication, we never sampled two similarly sized fish with the same infection status within 25 m.

2012年2月至3月，研究人员使用拉伸网目为10 mm的单丝拦网与硅胶手抄网，在澳大利亚大堡礁北部蜥蜴岛周边海域（14°40′S；145°28′E）采集了感染寄生虫（N=28）与未感染寄生虫（N=12）的双线眶棘鲈（Scolopsis bilineata）。实验鱼通过充氧的20 L Handy Pail水桶（每桶最多容纳4尾鱼）在捕获后1小时内转运至蜥蜴岛研究站，再以硅胶手抄网将其转移至配有直接取自礁体的流水系统的单个饲养水族箱（40.0W × 29.0L × 18.0H cm）中。水族箱维持自然光照与温度条件（28±1 ℃），并为每尾鱼提供圆形PVC躲避屋。每日投喂生虾肉至鱼饱食，实验前24小时禁食，以确保所有实验鱼处于标准化的吸收后（post-absorptive）状态，最大化游泳所需的能量供给。每日通过虹吸清理饲养水族箱以维持优良水质。所有实验鱼在开展游泳性能测试前至少在水族箱中驯养3天，以确认个体健康状态良好。

快速逃逸启动实验
我们设置了3组不同感染状态的实验鱼：未感染组（鱼体全长（total body length, BL）：13.2±0.8 cm；平均值±标准差，N=12）、感染组（BL：11.8±1.1 cm；平均值±标准差，N=16）与寄生虫移除组（BL：12.7±1.3 cm；平均值±标准差，N=12）。在快速逃逸启动实验开始前24小时，研究人员将鱼置于浅水环境中，使用镊子小心移除等足类寄生虫，该操作耗时约45秒。鱼被放回饲养水族箱后可在2小时内恢复正常行为。
实验在配有木质支架的长方形亚克力水族箱（70.0 L × 60W × 35H cm）中开展，所有测试均维持水深12 cm，该水深可限制鱼体在垂直平面内的活动（Langerhans, 2009）。在水族箱下方以45°角倾斜放置一面镜子，以便在同一相机画幅内拍摄三维视角的逃逸反应视频，存在垂直运动的视频将被排除在分析之外（占总视频量的<2%）。通过位于水面上方70 cm处的三盏150 W聚光灯为水族箱提供照明，持续流动的海水将水温维持在恒定的28 ℃。
在开展逃逸反应测试前，以硅胶手抄网将实验鱼从饲养水族箱转移至实验水槽，以避免损伤黏液层与鱼鳍，并将鱼静置30分钟以适应实验环境。蜥蜴岛周边海域常见猛禽与远洋海鸟等空中捕食者捕食鱼类，因此研究人员使用机械声学刺激模拟空中攻击，以诱导双线眶棘鲈产生逃逸反应。将一个装有铅坠的50 ml PVC容器通过细绳连接至水面上方30 cm处的平台。为确保逃逸反应不会在刺激物接触水面前触发，将刺激物置于一个宽15 cm的不透明PVC管内，该管固定于水面上方1 cm处。刺激物接触水面的时刻可通过下方的镜子清晰观察到。使用一台架设在三脚架上、正对水族箱与镜子的高速数码相机（Exilim EX-FH100，卡西欧Casio，美国）以240 Hz的帧率录制逃逸反应视频。每尾实验鱼接受最多3次刺激，刺激间隔为30分钟。当实验鱼正对刺激源、且与PVC管底端的距离不超过10 cm时，释放刺激物。测试结束后，将实验鱼放回饲养水族箱。本研究共为40尾鱼录制了127次逃逸反应视频。
有2尾鱼在实验结束数天后出现细菌感染，研究人员使用过量的Aqui-S溶液对其进行麻醉，随后通过冰浆进行安乐死。其余实验鱼未出现任何患病迹象，为避免疾病传播，我们在处理不同实验鱼前均使用淡水浴清洗手抄网。所有健康实验鱼（共38尾）均在测试结束后1周内放归至其原采集海域。
我们使用ImageJ v.1.43软件中的MtrackJ插件分析逃逸序列。以每4.2 ms为间隔追踪鱼体质心（centre of mass, CoM）的二维坐标，追踪时段涵盖鱼体首次运动发生前21 ms（5帧）至运动发生后84 ms（20帧）。鱼体质心通过距鱼头尖端约为个体总长度29%的比例距离进行视觉估算。本研究测量了以下逃逸性能指标：反应潜伏期（刺激触发至鱼体产生反应的时间间隔，单位为ms）、经体长校正的累计游动距离（D，单位为鱼体全长BL）、经体长校正的最大逃逸速度（Umax，单位为BL·s⁻¹）与最大加速度（Amax，单位为m·s⁻²）。游动距离相关变量（D、Umax、Amax）的分析时段固定为58 ms（14帧），该时段对应所有处理组中逃逸反应第1、2阶段的平均持续时长。使用五点移动二次多项式回归对距离的一阶导数（速度）与二阶导数（加速度）进行平滑处理。针对每尾实验鱼，选取3次刺激测试中各逃逸性能变量的最优值（例如最高Umax或最短潜伏期）用于后续分析（参考Domenici, 2011; Marras et al., 2011）。

逃逸起始距离（Flight Initiation Distance, FID）实验
研究人员于2013年7月至8月的晴朗天气下，在蜥蜴岛研究站前方的泻湖与邻近礁坪海域估算双线眶棘鲈的逃逸起始距离（FID）。该海域水深介于2~3 m之间，能见度约为15 m。我们使用浮潜者作为逃逸触发刺激源。诸多针对陆地与水生生态系统逃逸起始距离的研究均使用人类作为刺激源以引发逃逸行为。近期研究表明，无论观察者经验与否，逃逸起始距离的估算结果均具有较好的稳健性；在水生系统中，无论观测者使用浮潜还是水肺潜水，估算结果均无显著差异。
两名浮潜者在礁坪上游弋搜寻双线眶棘鲈，仅选取单独活动的成年个体作为研究对象，这些个体需正在礁坪上觅食或以缓慢速度在开阔区域移动，且可被直接靠近。距离分枝珊瑚或其他躲避处不足1.5 m的个体不进行测试，以避免躲避处距离对逃逸起始距离产生混淆效应。若实验鱼在观察者开始靠近前即开始以恒定速度向任意方向游动，则放弃该次测试。
一旦发现合适的实验个体，研究人员将记录其感染状态（感染寄生虫或未感染寄生虫，以下简称处理组）与总长度（误差±1 cm，实际测量误差）。在数据收集前，所有观察者均使用模型鱼与不同尺寸的物体练习水下估算鱼体长度，直至达到±1 cm的精度要求。一名浮潜者在与鱼体正对的位置、距离约5 m处就位，另一名浮潜者则在侧方就位以避免干扰测试。第一名浮潜者潜水至靠近底质的位置（距离≤1 m），并重新定位目标实验鱼。随后该浮潜者以快速但平稳的游泳速度靠近目标鱼，同时将两根系有警示带的重物举在头部旁，该动作被视为刺激触发的起始时刻。当实验鱼开始逃逸（即首次转向远离靠近的浮潜者）时，浮潜者将其中一根重物掉落至当前位置，并以另一根重物标记实验鱼当时所处的位置。随后两名浮潜者使用卷尺测量两个标记点之间的水平距离，精确至1 cm，该距离即为逃逸起始距离（FID）。
观察者还将根据0~4的评分标准对实验鱼的反应强度（逃逸强度）进行评分：0：无反应（即实验鱼未对浮潜者产生移动响应）；1：停止此前的活动（例如觅食），并向远处移动一小段距离，但未离开当前区域；2：改变游动方向，并开始缓慢远离当前区域；3：改变游动方向，并以快速但恒定的速度逃离区域；4：启动以“C型启动（C-start）”为特征的非稳态爆发逃逸行为。评分越高代表运动的能量成本越高，因此可用于估算逃逸行为的成本。
近期研究表明，双线眶棘鲈具有较强的栖息地附着性，日间极少远离其小型领地活动。因此为避免伪重复（pseudoreplication），研究人员不会在25 m范围内采样两条体型相近且感染状态相同的实验个体。