Abstract

Predation is pervasive in the natural world and shapes evolution of predators and prey. Numerous behavioral adaptations allow predators to manipulate prey behavior, resulting in higher success in predator-prey interactions. The lionfishes Pterois miles and Pterois volitans are highly effective predators and their invasiveness is attributed to their high hunting success. How lionfish attain such a high predation effectiveness remains enigmatic. Lionfish are stalking predators and show a peculiar behavior while approaching their prey fishes: they actively blow directed jets of water at them. This is the only known example of a predator blowing jets of water at mobile prey underwater. The causes and consequences of jet-blowing remain unknown. Here, we tested the hypothesis that lionfish use jets of water to re-orientate their prey and facilitate head-first capture. We conducted two separate experiments with two different groups of lionfish and prey species. We show that lionfish catch more prey head-first when blowing and that prey exposed to jets orientate towards an approaching lionfish, with smaller lionfish being more likely to blow. However, blowing was not determined by the initial orientation of prey and was largely driven by lionfish size. In addition, in one of the experiments, blowing lionfish struck from a further distance, possibly due to prey swimming towards them in response to blowing. Our results show that jet-blowing allows lionfish to manipulate the behavior of their prey by re-orientating them to facilitate swallowing, providing the first evidence that jet-blowing can be used to manipulate the behavior of mobile prey.

Corresponding author: Davide Bottacini (davide.bottacini@wur.nl)

Dataset DOI: 10.5061/dryad.ncjsxktbd

Description of the data and file structure

Details on how the data were collected can be found in the materials and methods section of the paper.

This is a comprehensive read-me file that describes the data sets that will be analysed for our publication on jet-blowing in lionfish. The structure is described separately for Data_analysis_experiment_1.xlsx (data collected by Marije Ridder in July 2024) and Data_analysis_experiment_2.xlsx (data collected by Emil Bathow in October/November 2024). Data were collected at the Hellenic Centre for Marine Research in Crete, Greece.

The two data files are very similar, although not identical. The differences are:

1) In experiment 1, there may be multiple approaches per trial (see below for exact definitions of approach and trial), while in experiment 2 every trial had only one approach. This is also why the data set of experiment 1 has a separate column showing the trial.

2) The data set of experiment 1 only shows trials in which the prey was not moving, since approaches in which the prey was moving were not analysed. In experiment 2, every approach was analysed, regardless of whether a prey was moving or not. This is indicated in an additional column. However, only trials in which the prey was not moving will be analysed for the paper.

3) We used D. labrax as prey in experiment 1 and S. aurata in experiment 2.

4) In experiment 2, the angle between lionfish and prey is split into variables called prey_angle and lionfish_angle. When added up, these two variables give the equivalent of the angles that were scored in experiment 1.

5) In experiment 2, the strike orientation only reports whether a prey was caught head or tail first, while in experiment 1 we also have lateral catches.

Notes: In experiment 2, we never observed lionfish fail. Missing values are given as NA. The only source of missing values is the impossibility of scoring the variable during video analysis. This could be due to technical issues (e.g., fogged camera) or lionfish/prey behaviour. For instance, in some cases a lionfish was approaching a prey with a downwards angle, making it impossible to score the end angle or the catch orientation.

Files and variables

File: Data_analysis_experiment_1.xlsx

Description:

Variables

• Date: the date on which a trial took place.
• Lionfish_ID: the identity of a lionfish involved in an approach.
• Approach: the serial number of an approach of a lionfish to a prey. Please note that in this experiment there may be multiple approaches of the same lionfish towards the same prey (see definition of Trial for full clarification).
• Lionfish_size: the total length (cm) of a lionfish involved in an approach.
• Prey_size: the total length (cm) of a prey involved in an approach.
• Blowing: 0 means that a lionfish did not show blowing behaviour, 1 means that it did.
• Start_angle: the angle (degree) between lionfish and prey at the start of an approach.
• Blowing_angle: the angle between a lionfish and a prey when the lionfish started blowing.
• End_angle: the angle between a lionfish and a prey at the end of an approach.
• Catch_orientation: head-first (hf), tail-first (tf), and perpendicular (p).
• Strike_distance: the distance (cm) between lionfish and prey at the end of an approach.
• Trial: the trial in which an approach took place. The start of a trial was defined as a lionfish entering the arena. The end of a trial was defined as a lionfish either catching a prey or losing interest in it. At the end of a trial, a lionfish was guided outside of the arena. Sometimes a lionfish missed the prey, but then approached the prey again within the same trial (i.e., the lionfish missed, but did not lose interest in the prey and tried to catch it again). In these cases, the trial number was the same, but the approach number was different.
• Catch_success: whether a lionfish ended up catching a prey (1) or not (0).

File: Data_analysis_experiment_2.xlsx

Description:

Variables

• Date: the date on which a trial took place.
• Lionfish_ID: the identity of a lionfish involved in an approach.
• Approach: the serial number of an approach. In this experiment, a trial started when a lionfish entered the arena and ended when a lionfish lost interest in a prey or struck. Regardless of whether a strike was successful, the lionfish was guided outside the arena after one strike, so there are no multiple approaches within a trial in this experiment. Hence, there is no variable Trial either.
• Lionfish_size: the total length of a lionfish involved in an approach.
• Prey_size: the total length of a prey involved in an approach.
• Blowing: yes or no depending on whether a lionfish blew or not in the course of the trial.
• Start_angle_prey: prey angle at the start of an approach.
• End_angle_prey: prey angle at the end of an approach.
• Start_angle_lionfish: lionfish angle at the start of an approach. Not at the start of blowing!
• End_angle_lionfish: lionfish angle at the start of an approach.
• Catch_orientation: head or tail depending on whether a prey was caught head-first or not.
• Strike_distance: the distance between a lionfish and the prey when the lionfish struck
• Swimming: swimming or sitting depending on whether a prey was moving or not at the start of a trial.
• Initial_orientation: orientation of a prey with reference to the lionfish head when an approach started.

Code/software

The data were analysed in R 4.4.1, with packages and versions described in the Materials and Methods section of the paper.

Access information

Other publicly accessible locations of the data:

• none

Data were derived from the following sources:

• All of the data were collected experimentally.