The role of skill in animal contests: Analysis of a neglected RHP trait in fighting hermit crabs

Lead Research Organisation: Plymouth University
Department Name: Sch of Biological and Marine Sciences


The resources that animals need to survive and reproduce are often limited in supply. Therefore, evolution has produced aggressive behaviours that allow individuals to take resources from others. The resulting fighting or contest behaviour has been the focus of intense study because the ability to win fights will have a strong effect on an individual's success in evolutionary terms. Most effort has focussed on the questions of why animals should fight and how they should make the decision to quit a fight and relinquish the contested resource. In contrast far less effort has been made to understand how animals fight, and what might differ between the aggressive behaviour of winners and losers. This is important for two reasons. First, our understanding of why animals fight is dependent on measuring their fighting ability. While it seems obvious that larger individuals should be stronger than smaller ones, in nature more fights take place when opponents are evenly matched. Therefore, it could be subtle differences in their aggressive behaviour that really drive differences in fighting prowess. Second, by investigating aggressive behaviour its-self we are directly investigating the traits for winning fights that we assume are produced by natural selection.

In this project we will focus our investigations on a potentially important component of fighting ability that has been almost completely ignored to date: the possibility that winners and losers differ in skill. Skill describes the effectiveness of the movement patterns that animals use to form aggressive behaviours. We will use accurate filming and 3D animation technology to measure the movement patterns of fighting hermit crabs. These small crustaceans fight over the ownership of the empty snail shells that they rely on for protection. One crab, the defender spends most of the fight tightly withdrawn into its shell. The attacker, meanwhile, uses its shell to repeatedly strike the surface of the defender's shell. If the attacker fights well, this shell rapping behaviour will cause the defender to release its internal grip on its shell, allowing the attacker to evict it and take over the defender's newly vacated shell. We know that vigorous rapping (i.e. at a high rate) is important but should skill play a role as well? Here we will ask whether successful attackers that evict the defender fight more skilfully than those that fail to gain an eviction. First we will ask if their shell rapping is more efficient and accurate, by comparing the displacement distances and the positions of strikes between successful and unsuccessful attackers. We will also ask whether precision is important by comparing how concentrated the points of impact are for successful and unsuccessful attackers. Then we will examine how attackers decide when to switch between shell rapping and a less demanding tactic called shell rocking. In the second part of the grant we will look at the factors that could underpin variation in fighting skill. Do more skilful fighters, for instance, also show better performance of physically demanding or cognitive tasks? And how much of an individual's potential for fighting skilfully (technique) is actually expressed during a fight where its opponent should try to hinder skillful fighting?

These studies will represent the first concerted effort to study the role of skill in animal contests. By providing a better picture of what makes a good fighter, we will gain new insights into the evolution of aggression. Furthermore, these studies will have potential broader impacts beyond biology. In sports science, skill is also composed of accuracy, precision, efficiency and appropriate tactics. The statistical approaches we will use for analysing the data on these parameters in hermit crabs could be adapted for use in programmes that seek to identify potential elite athletes or to reveal areas for improvement in individual training programmes.

Technical Summary

We propose to conduct the first set of concerted studies into the role of fighting skill during animal contests. We will use the common European hermit crab as a model organism and test ideas about skill that have recently been developed in the literature. Broadly, we aim to (i) determine the extent to which variation in skill drives variation in fighting ability (resource holding potential, RHP) and (ii) to investigate potential sources of variation in the level of skill that an individual can bring to bear during a fight. Skill can be quantified through three spatial parameters of movement (efficiency, accuracy, precision) plus the choice of appropriate agonistic tactics. We will investigate the spatial parameters of skill by combining scientific rotoscoping and 3D animation techniques to determine displacement distances and the positions of points of impact when an attacking crab raps its shell against the shell of a defending crab. Accuracy and precision will be assessed using double hierarchical general linear models that assess mean and variance level effects simultaneously. To investigate the underpinnings of skill we will manipulate environmental parameters prior to fighting and test for associations between our measures of skill and performance over a range of appropriate physical and cognitive tasks. If skill is an important RHP trait, this will have consequences for our understanding of the evolution of aggressive behaviour. We will provide the first evidence that skilful motor patterns represent an adaptive behavioural trait in the context of animal contests. Furthermore, current tests of contest theory rely on accurate estimates of RHP in fighting animals, which are used in analyses that seek to understand how make strategic decisions are made. If skill is an important RHP trait that has thus far been neglected, this could lead to a reappraisal of how we test the central predictions of contest theory.

Planned Impact

The scientific impact of this work will primarily in the area of animal contests but could also be relevant for sports scientists, and by extension the sports industry. We would disseminate our findings through the organisation of symposia and a stand alone meeting, to which we would invite researchers and representatives from relevant organisations.

We would also seek to engage with the public and both the PI and RI have track records of successful social and traditional media engagement and public talks, all of which are supported by Plymouth University. Due to the large amount of video data that would be collected and the fact that skill (as a general concept) is readily understood, there is also the opportunity to engage with the public directly through a citizen science project. As described in Pathways to Impact, this would be an online based project that would compare viewer perceptions of skill levels with the data on skill that we collect as outlined in the Case for Support. This project could address ongoing questions about the role of anthropomorphic language in animal behaviour and the hope is that it would also stimulate public debate and interest in the subject generally, as well as generating specific interest for the proposed project on skill.

There is a huge public thirst for new information about the natural world, and in this project we will make new discoveries about a species that anyone living in the UK can readily encounter in its natural setting, and where related species are easily accessible globally. Furthermore, the core ideas underlying the research, the links between skill and success in contests, are of fundamental interest to the public. Furthermore, the potential links between animal behaviour and sport are likely to enhance the public's interest in the work and to attract the attention of the media. Therefore, this project is provides excellent opportunities for public engagement with science, in particular demonstrating the importance of fundamental science that can nevertheless be relevant to a diverse range of other areas.
Description Our initial experiment has now been published. In it we differentiate between fine scale and coarse scale accuracy during fights and show that course scale accuracy (striking the opponent in the correct region) is more important than fine scale precision (hitting a narrowly defined spot again and again). This addresses one of the key questions about skill during a fight described in the proposal - understanding the relative importance of accuracy versus precision.

During the first (2020) spring/summer lockdown we were unable to enter labs. We devised an alternative study using existing data on mixed martial arts fighting (MMA) looking at the relative contributions of vigour and skill to the perceptions of fighting prowess of third part observers (judges in this case). We found that they rate vigour more accurately than skill; and there is a mismatch between contest outcomes and ratings of skill. This finding is very relevant to the aims of the project as animal contests where skill is important also involve third part observers with an interest in assessing the performance of the combatants (e.g. a female judging the quality of two fighting males). The study has been published in Biology Letters.

We have now completed the four man experiments on the original proposal. One on accuracy versus precision is already published, and a second on technique versus skill is accepted and in press. The remaining two studies on muscle strength and cognition have completed data collection and manuscripts are in prep, with a view to submitting in 2022.
Exploitation Route These findings will be relevant to researchers working on animal conflict, who may adopt similar approaches to the analysis of their own study systems now that the concept of spatial skill in fights has been validated. It may also influence contest theory as we show that selection may favour broad-scale accuracy over fine-scale precision for movement patterns used in agonistic behaviour. In addition these findings could be relevant to the field of combat sports, especially in the case of our analysis of skill in MMA fights.
Sectors Other

Description Rotoscoping 
Organisation Liverpool John Moores University
Country United Kingdom 
Sector Academic/University 
PI Contribution Introducing the study system to the collaborator, and providing live animals for trial runs of the 3D filming and subsequent rotoscoping.
Collaborator Contribution The collaborator at LJMU has provided additional expertise in the scientific rotoscoping that we are using in this project, and has enabled us to refine and simplify our procedures. For example as a result of the collaboration we will use photogammetry to build virtual models of gastropod shells instead of 3D scanning originally envisaged.
Impact The collaboration is in its early phase. It will enable us to collect better data, more efficiently in due course.
Start Year 2019
Description XROMM 
Organisation University of Liverpool
Country United Kingdom 
Sector Academic/University 
PI Contribution We have introduced the collaborator to the study system and have provided hermit crabs for analysis. The concept is that we will use XROMM equipment in the partner's lab in order to 'see inside' hermit crab shells during live filming for fights. This will provide insights into the behaviour of defending crabs (which has thus far been impossible).
Collaborator Contribution Expertise in using XROMM approaches to reveal movement patterns of animals that are usually obscured. Optimizing XROMM system for use with hermit crabs.
Impact This collaboration is in its initial phase. We have obtained some preliminary XROMM film and aim to generate data in due course.
Start Year 2020