Experimental animals
Three behavioural experiments were conducted using juvenile common rudd (S. erythrophthalmus) as the model organism. A substantial body of prior research has been conducted on this species, examining its physiological and behavioural responses to temperature changes19,41. A total of 180 individuals were utilised in each experiment. Prior to the commencement of the experiments, the fish were weighed while anaesthetised with buffered MS-222 (tricaine methanesulfonate) at a concentration of 100 mg × L− 1. The mean fresh weight of the fish was 2.01 ± 0.42 g (Exp I), 1.82 ± 0.30 g (Exp II) and 2.32 ± 0.68 g (Exp III).
The red and yellow elastomer tags (Visible Implant Elastomer (VIE) Tags, Northwest Marine Technology, Inc., USA) were prepared in accordance with the manufacturer’s recommended procedure. Subsequently, the tags were injected into the anaesthetised fish as a thin strip just under the skin using insulin syringes, in close proximity to the dorsal fin on both sides. This enabled the differentiation between groups of experienced and naive fish when they were combined in an experimental system. The fish that had not been tagged with the elastomer were administered a physiological salt injection (0.06% sodium chloride dissolved in distilled water) in the same body part. The unmarked fish were employed in order to facilitate the replacement of half the fish in the second stage of the experiment with naive fish, while maintaining the same overall number of fish in both stages of the experiment. The fish marked with yellow elastomer served as naive fish during the first stage of the experiment and as experienced fish during the subsequent stage. Conversely, the fish marked with red elastomer performed the function of naive fish during the latter stage of the experiment. The selected unmarked and marked fish were acclimated to a specified temperature for a minimum of seven days. Subsequently, the fish were provided with a limited quantity of frozen Chironomidae sp. larvae and Artemia salina nauplii, administered in equal proportions across the entire system. The prey utilized during the foraging sessions was A. salina nauplii, hatched daily from dried brine shrimp cysts (Sanders Brine Shrimp Company, Inc.), sourced from the fishless Great Salt Lake in Utah, USA. The two-day-old nauplii remained viable and motile in the freshwater system for a period of 24 h. The established experimental temperatures of 16 and 26 °C were selected to encompass the optimal temperature range for this species, with the lower temperature falling within this range and the higher one slightly above it. This approach is supported by previous research42,43,44.
Experimental system
The experimental system consisted of two identical sections. Each section was constituted of ten tanks, with each tank containing 200 L of tap water. The tanks were rectangular in shape and connected to one another via circular openings with a diameter of 8 cm. The openings were furnished with rotatory blinds, thereby enabling the openings to be opened and closed at any desired point in time. The system maintained: (1) a uniform light intensity of 0.8 ± 0.2 µmol × m⁻² × s⁻¹ at a depth of 0.5 m, and (2) a constant water temperature, with one section at 16 °C and another at 26 °C. The movements of fish within the final tank, which contained a food reward, were recorded using high-resolution infrared cameras (P.P.H. Matar KT-370/540, Poland) connected to a computer and positioned 0.5 cm below the water surface. A detailed description of the system can be found in our previous study19. The number of tanks employed in the experiments differed, with four tanks used in Experiment 1, six in Experiment II, and eight in Experiment III. This resulted in a variation in the distance between the initial tank and the final tank, which contained the food reward (a high density of A. salina nauplii). The distance was 1.3 m in Experiment I (three obstacles to overcome), 1.6 m in Experiment II (five obstacles to overcome) and 2.8 m in Experiment III (seven obstacles to overcome). During the experiments, the outer blinds were closed in the start and last tank, thereby enabling the fish to leave the tanks only in one direction.
Experimental procedure
The experiments were conducted in two distinct phases. In the initial phase of the experiment, 30 fish marked with yellow elastomer and 30 unmarked fish were placed in each section and permitted to learn to locate the food patch over the course of 16 daily foraging sessions, with the temperature set at either 16–26 °C. In the second stage of the experiment, 30 marked fish with red elastomer were randomly substituted for the 30 unmarked fish. Over the following eight days, the recently introduced fish were permitted to follow the cues in the presence of 30 individuals with prior experience.
Seven days prior to the commencement of the experiments, the fish were introduced to the starter tank, with one group allocated to the section with a low temperature and the second group to the section with a high temperature, in order to facilitate their acclimatisation to the experimental temperature. Prior to the commencement of each experiment, the water flow through the tanks was temporarily halted. The fish that had been labelled with elastomer for utilisation in the second stage of the experiment were captured and transferred to one of the tanks that had not been utilised during the experiment.
The same experimental procedure employed by Gliwicz and Maszczyk (2016)19 was utilised to provide the fish with patchily distributed Artemia prey. Just before the start of each foraging session, 91% of the total number of Artemia was introduced into the high-density tank, while 1% of the total number of Artemia was added to each of the other tanks. The Artemia in the other tanks were introduced to encourage the fish to forage and search for a food patch. The experiment commenced with the opening of the connection between the starting tank and the subsequent experimental tank. The timer was activated, and camera recording began, allowing the fish to occupy the experimental tank unaccompanied.
Each foraging session lasted 60 min. The fish in the tank with the food patch were recorded at one-minute intervals throughout each foraging session to determine the time taken to complete the task and to measure the swimming speed of the fish. At the conclusion of each session, video recording was stopped, and water flow through the tanks was restored. The fish were then transferred back to the starting tank, and the connection between the starting tank and the adjacent tank was closed.
The experimental procedure was approved by the First Warsaw Local Ethical Committee for Animal Experiments (permission protocol no. 042/2016). All methods were performed in accordance with the relevant guidelines and regulations. The study is reported in accordance with ARRIVE guidelines.
Data analysis
The time to complete the task and the swimming speed of the fish (m × s⁻¹) were determined by analyzing archived video footage. It was determined that the task was completed during each foraging session when 20% of the fish marked in yellow during the first phase of the experiment and 20% of the fish marked in red during the second phase of the experiment were present in the tank with a food patch. The 20% threshold was adopted based on the experimental procedure used in our previous study19. This threshold was sufficient to reliably estimating the accumulation rate of the fish, while also ensuring that the foraging session ended before there was a significant reduction in food (Artemia) density in the high-density tank.
The fish swimming speed was measured by comparing two successive images, with the distance traversed by an individual fish in the lower 20 cm of the tank quantified using a scale marked at the tank’s base. The distance was then divided by the elapsed time to obtain the swimming speed.
The impact of temperature on fish performance, encompassing both task performance and speed, was quantified as a temperature coefficient (Q10), calculated in accordance with the formula proposed by Schmidt-Nielsen (1979)14. The temperature coefficient (Q10) was calculated using the following formula: Q10 = (R2/R1)(10/T2 − T1), where R2 is the measured reaction rate at temperature T2 (where T2 > T1) and R1 is the measured reaction rate at temperature T1.
The statistical analysis was conducted using the R software, version 4.3.2 (R Core Team, 2023). A 5% statistical significance level was set. The data set under analysis encompasses the initial six days following the introduction of the fish into the experimental system. The effects of temperature, fish group, distance to the patch, and their interactions on task performance were analyzed using generalized linear models (GLM)45, performed with the Template Model Builder (the glmmTMB package v.1.1.3)46. Temperatures (16 and 26 °C), fish group (fish without and with experienced individuals), and distance to the patch (short, medium, and long distances) were set as fixed effects. Task performance, measured as the time required to accumulate 20% of the fish in the high prey density tank, was designated as the response variable.
The same type of model was also used to test the significance of fixed effects: fish group (fish without and with experienced individuals) and distance to the patch (short, medium and long distances) on the Q10 coefficients for fish swimming speed, time to accumulate 20% of fish in the high prey density tank, and the ratio of these two measures.
All count data were modeled with a gamma distribution with log link function. The models fitting to the dataset were diagnosed using DHARMa scaled residual plots (the DHARMa package v.0.4.5)47. Significance of interactions between the factors was assessed using analysis of deviance and the Wald type II chi-squared difference test (χ2) (the car package v.3.0–12)48. Post-hoc multiple comparisons were based on the planned contrasts for estimated marginal means (EMMs; the emmeans package (v.1.7.2.)49. The Holm p-value adjustment was applied to control for type I error inflation due to multiple testing.
The Q10 for the daily rate of change in task performance during the first foraging session, the first three of the foraging sessions, and the first seven foraging sessions, for each fish group (fish with and without experienced individuals) and across different distances to the patch (short, medium, and long), were presented using descriptive statistics.
