Organisms determine environmental quality using their senses and personal experience (personal information), but can also use by-products of other individuals’ activities, i.e. public information. The ability to use public information originating from both con- and heterospecifics gives an advantage over individuals relying only on personal information or conspecific cues. The role of public information in invasion ecology is of high concern, as any differences in this aspect between alien and native species may determine the success of the former. Here we used two pairs of sympatric invasive and native demersal fish species (racer goby Babka gymnotrachelus / European bullhead Cottus gobio; monkey goby Neogobius fluviatilis / gudgeon Gobio gobio) facing two types of public cues (associated with frightened and foraging individuals) as a model to check if the invaders are more effective in public information use than the natives. Both invaders and the native gudgeon used danger cues from con- and heterospecifics, while the native bullhead failed to recognize heterospecific danger cues. The monkey goby and both native species appear to be attracted to foraging cues from donors less likely to exert competitive pressure on the observer (i.e. native species rather than potentially more aggressive invaders), while the racer goby appeared unable to correctly recognize heterospecific cues. Our results showed that public cues can enable invaders to read threats from a wide range of individuals and to find optimal food patches, which may contribute to their invasion success.

Uploaded data present the behavior of fish during the experiment assessing effectiveness in social information use. Test fish behavior was recorded, and, to minimize observer bias, we used the Ethovision XT® 10.1 program to analyze it from the videos recorded.

Experimental setup:

We used LCD computer monitors to present the visual cues to the fish tested. The cues were prepared by recording particular behavior of demonstrators' individuals (antipredator behavior or foraging) and presenting them to the tested fish individual. The experimental setup consisted of a glass experimental tank (30 x 30 x 30 cm; length x width x height), with two computer monitors (BenQ GW2280-B, Taipei) placed on two opposite sides of the tank to display the stimulus videos (one monitor was to display visual cue, second one acted as a control). An IP camera (Samsung SNB-6004P, Changwon, South Korea) was placed 50 cm above the water surface to record the fish behavior.

Behavior analysis:

For Experiment 1 (antipredatory behavior), we set two zones on the experimental tank bottom, covering the whole tank width: (1) the “stimulus zone”, adjacent to the “stimulus” monitor (showing demonstrators), and (2) the “control zone”, adjacent to the monitor displaying the control sequence. We measured the time spent by the test fish in each zone (expressed as % of the total period time) and its general activity, expressed as the time spent on the move (% of the total period time).

To analyze Experiment 2 (foraging), we set two zones on the experimental tank bottom: (1) the “foraging zone”, set up around the feeder on the monitor displaying foraging demonstrators, and (2) the “non-foraging zone” adjacent to the wall facing the monitor displaying the “control” sequence. We measured the time spent by the test fish in each zone (expressed as % of the total period time).

For both experiments, we calculated zone preference indices using times spent by fish in particular zones in both experiments, according to the formula:

((time spent in the zone of interest) - (time spent in the other zone)) / ((time spent in the zone of interest) + (time spent in the other zone))

The value of “0” meant that the fish showed no preference for or avoidance of any zone. The positive and negative values meant that the fish showed a preference for or avoidance of the zone of interest, respectively.

Statistics:

To test the effects of observer species, demonstrator species, and period on zone preference index (Experiments 1 and 2) and general fish activity (in Experiment 1), we performed Linear Mixed Models (LMM), separate for each set of species and experiment. The observer species and demonstrator species were set as between-subject fixed effects, the period (pre-stimulus/flight response/inactive demonstrators in Experiment 1, pre-stimulus/stimulus in Experiment 2) as a within-subject fixed effect, and the test fish ID as a random effect. In both experiments, we interpreted the main effect of period and its interactions with other factors as indicators of fish reaction to different types of cues presented. The parametric test assumptions were not violated based on the visual inspection of residual plots.