Dominance, conflict and association Flashcards
Group living 4
Dominance hierarchies
In many social species, the relationships between pairs of individuals are asymmetrical.
Animals compete for valued resources: food, shelter, mates, status, territory
Observations of competitive interactions between ‘every’ pair in group:
Often seems that one individual tends to supplant all others, whereas another is supplanted by all others. In between the top and bottom ranking animals are animals that supplant some but are supplanted by others = a dominance hierarchy.
Computing dominance
First decide on what observations to make:
–What constitutes a dominance interaction?
–Depends on species, context, resources etc etc.
(See lectures on doing behavioural observations)
Re-arrange recorded observations so that the individual that is never supplanted is at the top & the one that is always supplanted is at the bottom.
Orre-arranged in order until minimum number of supplants appears on the left-hand side of the diagonal.
Properties of dominance hierarchies
*In a perfectly linear hierarchy.
^all dyadic relationships must be asymmetric, all possible triadic relationships must be transitive
(Transitive = sequential)
In reality, few dominance hierarchies are perfectly linear
*Often reversal occur - this is when an apparently subordinate individuals wins an encounter with a normally dominant individual.
Often:
Some individuals of equal status
Some relationships NOT transitive
(see diagram of Northern elephant seal hierarchy in notes)
Why do reversals occur?
Reversal = when an apparently subordinate individuals wins an encounter with a normally dominant individual.
Can occur due to:
Intrinsic effects e.g.
*in some groups two or more individuals may have equal or very similar status.
*changes in age /condition / health.
*Perceived value of resource
*follow on effects.
Extrinsic effects
*e.g. topography/habitat.
*Location (e.g. in own territory vs. out of it)
*Context (relates to what observations you base your calculation of dominance on e.g. competition for food cf. mates cf. grooming, etc).
Random effects:
*mistakes (by the animals)
*mistakes by the observer!
Dominance is not necessarily a fixed attribute, dominance relations are often fluid (flexible) and capable of rapid change
Dominance hierarchies: Quantifying linearity
Landau’s index of linearity (h) (see equation in notes)
- h value ranges from 0.0 to 1.0. (1.0 = perfect linearity).
- values of h > 0.9 denote a strongly linear hierarchy.
More complex modifications of this now exist (e.g. De Vries 1995), but this is the basic principle
e.g. In a study of hyenas: Relatively linear hierarchy observed with some reversals in mid/low ranking individuals with clearer dominancy in alphas
Dominance hierarchies: cautions
–high probability that any data set can be arranged to form a linear dominance hierarchy when none exists in reality (Appleby, 1983).
–especially when group is small – i.e. easy to juggle data until the best hierarchy is obtained.
e.g.: 5 or less individuals: perfect linear hierarchy can be obtained by chance (p > 0.05) even when the underlying dominance relationships are actually random.
Need at least 6 individuals to show statistically significant linearity – more if there are reversals!
–So, need a good sample size of individuals
–Or, repeated observations.
Dominance rank
*Dominance Rank: the index (order) of dominance status assigned to each individual.
–i.e. dominance is measured on an ordinal (ranking) scale,
*No indication of the magnitude of the difference in dominance status between two individuals.
Other ways of calculating dominance
Examples:
Boyd & Silk (1983)
I&SI (De Vries)
David’s Score
^ measuring dominance on an interval scale
-> difference in dominance between 2 individuals can be quantified and tested for statistical significance. This is useful for describing hierarchies that are not highly linear.
The issue with dominance hierarchy models is that they rely on clear win/loss outcomes – in many conflict situations not losing (drawing) can be as important if not more than losing
Analyse any asymmetric interactions involving actor & recipient, e.g. grooming, food-sharing.
Dominance is not necessarily a fixed attribute
Dominance relations are often fluid and capable of rapid change, they are affected by:
*Age
*Location (e.g. in own territory vs. out of it)
*Context (relates to what observations you base your calculation of dominance on e.g. competition for food cf. mates cf. grooming, etc).
What observations should be used?
Fights only (escalated contests)?
Or all aggressive interactions?
Win/lose only? or draws also?
What can we do with dominance scores?
- Relate them to fitness e.g. Assam Macaques (Ostner et al. 2011) dominant individuals have higher mating success.
- Assess how dominance hierarchies affect stress levels e.g. Bergman et al. 2005: Chacma baboons (Papio hamadryas ursinus) - cost of dominance
They noted periods when stable dominance hierarchies existed & when changes occurred in the dominance hierarchy (e.g., when a new male became dominant).
Method:
Feacal samples used to assess glucocorticoid levels of individuals
Conclusion:
Aggression & stress levels lowest during periods of stable dominance hierarchy
Conflict behaviour: Fighting behaviour and game theory
Additional realism:
- Fighting behaviour should also be affected by the (perceived) value of the contested resource
-expect higher intensity fights over more valuable resources.
- individuals differ in their fighting ability = Resource holding power (RHP)
–reflects the fitness budget available to a contestant during a fight.
- Can individuals assess RHP asymmetries?
e.g. Does the potential fitness benefit of a particular mate affect the duration of fights between male red-spotted newts (Notophthalmus viridescens)? (Verrell 1986)
*Staged encounters between size-matched males over a female that varied in body mass
*Record duration of wrestling between the males
*Duration of wrestling was positively correlated with female body length
*i.e. males can adjust level of competition according to potential benefits (percieved value)
Game theory assessment models
Self-assessment models
- An individual engages in a contest until it reaches an internally determined cost threshold.
- First individual to reach its cost threshold loses the fight.
Mutual assessment models
- Individuals assess the relative fighting ability of their opponent.
- Each contestant references its own RHP relative to that of its opponent.
- Individual possessing the lower RHP can withdraw from the contest early on, reducing time, energy & risk of injury from engaging in a contest that it would lose.
- The individual with the higher RHP wins the fight.
- sequential assessment model is a type of mutualistic assessment e.g. seen in deer rut season (see notes)
see in notes, graphs by Gammell and Hardy 2003:
a)Self-assessment models predict contest duration increases with both loser and winner RHP.
b)Mutual-assessment models predict contest duration increases with loser RHP but declines with winner RHP
Determining a suitable game theory assessment:
example: Male fiddler crabs fight over burrows. Size of claw determines dominance
Does a mutual-assessment or self-assessment model best explain contest duration?
Methods:
- Staged contests between males of different RHP (measured by claw size)
- Examined correlations between claw size and contest duration.
Results:
- +ve correlation for both winner & loser
= self-assessment model (compare to graphs above)
Social associations
Dominance
–Separation of individuals (avoidance)
–Coexsitence in social groups
Association
- altruism and codependence
Measuring social associations
Indices of association:
*Quantify extent to which two individuals, A and B, associate with each other.
*Assess number of separate occasions that A and B are seen together,
*Also, the number of separate occasions that A is seen on its own and the number of separate occasions that B is seen on its own.
