Transcript 0.5 - UMK

13. Altruism and sociality
Primitive animals are all the same. There is no individualistic behaviour.
Higher animals evolved individualism. The highest birds and mammals evolved
individualistic characters (moods), motions and fears.
Classical population genetic does not predict individualism because it focuses on
optimisation and equilibrium states that are the same for all members of a population.
Evolutionary theory has to explain:
•
Altruism (the help of others despite of own costs)
•
Cooperation of related and unrelated individuals
•
The evolution of cheating
•
Sexual selection (the existence of differentiated sexual behaviour and mating rituals)
•
Biased sex ratios (the prevalence of either males or females in a population)
•
The existence of highly altruistic insect societies (eusociality)
•
The existence of infanticide in many mammals and birds
•
The existence of homosexuality in many mammals and birds
•
The appearance of common beliefs and religion in man
The unit of selection and evolution
Unicellular organisms
Multicellular organisms
Higher taxonomic level
Species
Classical population genetics (Fisher,
Haldane, Sewall Wright)
Population
Higher taxonomic level
Group
Species
Family
Population
Individual
Cell
Organelle
Genome
Gene
Nucleotid
Wynne Edwards (1962)
to explain cooperation
A more liberal view sees any trait
inducing carrier of information as a
potential unit of evolution. These
include genes, individuals, and
even groups but not species.
The basic unit is the gene as
the smallest essential carrier
of information
C. Richard
Dawkins (1941-
The game theory approach
The classical hawk and dove game
Assume two players:
• a hawk that will always fight until
injured or until the opponent retreats
John F. Nash
(1928-
• a dove that will always retreat.
John Maynard
Smith (1920-2004)
Contests are associated will potential
benefits (B) and potential costs (C).
Hawk v. Hawk: Each
contest has a 50%
chance to win. The net
gain is the difference
between benefits and
costs of the contest
Dove v. Hawk: The
dove will always loose
The pay-off matrix
Hawk
Dove
Hawk
½(B-C)
B
Dove
0
½B
Hawk v. Dove: The
hawk will always win
Dove v. Dove: Each
contest has a 50%
chance to win. There are
no costs
The pay-off matrix
Hawk
Dove
Hawk
½(B-C)
B
Dove
0
½B
The idea behind game theory is now to define equilibrium
conditions that define which game (strategy =
behavioural phenotype) will have the highest payoff in
the long run.
Maynard Smith defined such equilibria that cannot be
beaten by other strategies as evolutionary stable
strategies (ESS).
Populations of individuals playing an ESS cannot be
invaded by immigrating individuals or by mutants playing
other strategies.
The fitness
Is H an ESS?
W(H)  pW(H, H)  (1  p)W(H, D)  W0 
BC
 (1  p)B  W0 
2
B  pB / 2  pC / 2  W0
p
p
BC
 (1  p)B  W0  (1  p)B / 2  W0  B  pC
2
If B > C, H is always an ESS because per
definition 0 ≤ p ≤ 1.
W(D)  pW(D, H)  (1  p)W(D, D)  W0 
Is D an ESS?
(1  p)B / 2  W0  B / 2  pB / 2  W0
For H to be an ESS W(H) > W(D)
BC
 (1  p)B  W0  (1  p)B / 2  W0  B  pC
2
For D to be an ESS W(D) > W(H)
If B > C, D is never an ESS
p
The pay-off matrix
Hawk
Dove
Hawk
½(B-C)
B
Dove
0
½B
What is if costs are higher than benefits C > B?
H : B  pC
D : B  pC
At equilibrium we have
B  pC  p 
B
C
For C > B an ESS is to play hawk with probability p
and dove with probability 1-p.
Even simple games favour mixed strategies.
This is the start of individualistic behaviour.
The Retaliator game
The Bourgeois game
(fight when meeting a hawk and retreat when
meeting a dove)
(fight when owner, retreat when intruder)
Hawk
Dove
Retaliator
Hawk
½(B-C)
B
½(B-C)
Dove
0
½B
½B-e
Retaliator
½(B-C)
½B+d ½B-¼C+g
Retaliator and a mixed strategy are the two
ESS of this game. Realization depends on
the initial frequencies of players.
Hawk
Dove
Bourgeois
Hawk
½(B-C)
B
¾B-¼C
Dove
0
½B
¼B
Bourgeois
¼(B-C)
¾B
½B
The Bourgeois is the only ESS of this game.
Local mate competition
In 1967 W. D. Hamilton proposes that in the long run organisms should preferentially
invested in the cheaper sex.
The cheaper sex is the one that promises more offspring at equal costs.
pM rM CM  pF rFCF
p: probability to produce a son; r: expected
reproductive success, C: cost of reproduction
Which sex to produce?
The probability that
a son reproduces
is high
The probability
that a daughter
reproduces is low
For a proper choice a female
• needs knowledge about the actual sex ratio and
• must have the ability to control which sex she produces
Many Hymenoptera and some
other insects have these abilities
Mammals and birds perform
selective infanticide
Two examples of sex ratio allocation
Figs and fig wasps
Parasitic wasps
Agaonidae are closely
connected to figs.
Depositing eggs into the
ovaries they pollinate figs.
Sex ratio
ro
0.6
0.5
0.4
0.3
0.2
17 species of fig
wasp species
(Agaonidae)
0.1
0
0
0.2
0.4
0.6
0.8
z
Sex ratio second female
Males are wingless and
mate only with the local
clutch
1
Proportion
of of
fruits
Proportion
fruitparasitized
parsitized
Sex ratio is defined as the proportion of males
Secondary
parasitism of the
parasitoid wasp
Nasonia vitripennis
parasitoid of blow
and flesh flies
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
y = 0.27x-0.43
0
1
2
3
4
Offspring second female /
Offspring first female
5
Selective infanticide in man
The sex ratio is the proportion of males: SR = males / (males + females)
The normal cross cultural sex ratio at birth is 105 males to 100 females = 0.512
(range 101 to 107: 100)
Some reported sex ratios in childhood of preindustrial societies:
Inuit Eskimos: 0.67
Yanomamö Indians: 0.56
Cashinahua, Peru: 0.60
Rajput caste, India: > 0.9
Upper class medieval Florence: 0.57
Selective infanticide in man is found in nearly all cultures.
Often it serves to
• stabilize population size
• to adjust sex ratios to marriage probabilities in cases of highly unequal reproductive success
• to adjust to a culturally preferred gender (frequently the male gender)
Reciprocal altruism
Reciprocal altruism beween non-related
individuals needs:
Blood sharing in the vampire bat
• Long term association of group members.
• Donorship can be predicted from past helping.
• Roles of donors and recipients reverse.
z
• Benefits of the recipients outweigh donor costs.
Percentage of prefeeding weight
• Donors can detect cheaters.
130
Exponential vampire bat
weight loss function due
to starvation
120
• Primary social groups contain 8 to
12 adults with depending young.
110
100
• 30% of the blood sharing events
involve adults feeding young other
than their own.
Weight Donor
lost
90
Recipient
• Blood sharing intensity depends on
the degree of relatedness.
Weight gained
80
Time lost
Time gained
70
0
20
40
60
Hours
Benefits outweigh costs
80
• Blood sharing is often reciprocal.
• Cheaters have not been observed.
Cooperative breeding and helpers at the nest
In the pied kingfisher Ceryle rudis
primary and secondary helpers at the
nest occur.
Helpers occur in many higher bird
species and help adults to raise the
offspring.
Male
helpers
S
Primary helpers are older sons that are yet
unable to breed.
Additional young
fledged per helper
P
They increase their fitness via their younger
sisters and due to additional experience.
>5
Number
of adults
providing
care
4
Secondary helping males are unrelated to the
pair they help.
3
2
0
1
2
3
Young fledged
4
5
Secondary helpers increase their fitness due to
the chance to become the widow’s mate if the
breeding male dies.
The evolution of cheating or the Prisoner’s dilemma
Assume two prisoners have the alternative either of defect the other or to
cooperate. Defection means shorter imprisoning.
B>C
The pay-off matrix
Defect Cooperate
0
Defect
0
Cooperate
0
B(A)
B(B)
0
C(B)
C(A)
If both prisoners defect they do worse than if
both cooperate. However cheating the other
is superior irrespective of what the other
makes.
Hence pure cooperation can never evolve.
Now assume an iterative game where the players
play many times. What would be the best strategy?
In the long run there are several possible strategies
One EES is Tit for Tat (defect if prior being defected
and cooperate if the other prior also cooperated).
Defect CooperateTit for Tat
Defect
0
e
0
Cooperate
d
g
g
Tit for Tat
0
g
g
The prisoners dilemma cannot fully be
resolved analytically.
The first software solution was provided
by Rapoport in 1980.
The program played
Tit for Tat or reciprocal altruism.
The other EES of this game is always
defect.
Inbreeding
What is the probability for a children to get a certain allele from their grandparents?
Grandparents GM1
A,B
Parents
P(C)=0.25
GF1 GM2 GF2
C,D
E,F
M
F
Ch
Childrens
G,H
P(C)=0
P(C)=0.125
The probability that Ch gets allele C is 0.125.
GF1 is already inbred
GM1 GF1 GM2 GF1
A,B
P(C)=0.5
C,C
M
E,F
F
GM1 GF1 GM2 GF1
A,B
P(C)=0.25
C,D
E,F
M
F
Ch
C,D
P(C)=0.25
P(C)=0.25
The probability that Ch gets allele C is 0.25.
The mean probability to get an allele X from
one of the members of a lineage is called
the coefficient of inbreeding.
C,C
P(C)=0.5
Sewall Wright defined this
coefficient as
n
Ch
P(C)=0.5
The probability that Ch gets allele C is 0.5.
rlm   21 Li (1  rl )
i 1
rl→m is the path from l to descendent m
and L the length of path i.
Inclusive fitness
In the Hawk - Dove game the EES for C > B was
B<pC → pB>C
P was the probability of a trait to occur. This is formally identical
with the probability of a gene to occur via descent, it is identical
to the coefficient of inbreeding.
William D. Hamilton
(1936-2000)
Hamilton’s rule of inclusive fitness
rB  C
A simple example
Assume a new gene A that promotes parental care.
The probability of transmitting A from mother to daughter
is 0.5.
Even if the mother would die due to parental care (cost =
1) two additional raised offspring (B = 2) satisfy
Hamilton’s rule.
0.5 = 1 / 2
Parental care should therefore be widespread in animals.
In cockroaches (Phoraspis
and Thorax) the young bite
wholes in the mothers
thorax to feed from their
haemolymph.
Kin selection and the evolution of sociality
Members cooperate
Part of the members loose
but retain reproductive
own reproductivity in favour
ability
of other group members
Individualistic life
→
Sociality
→
Eusociality
(superorganisms)
Joined parental care
→
Cooperative breeding
and defence
Most bacteria and
True multicellular
→ organisms (Metazoa,
Colonies
→
single cell
eucaryotes
Fungi, Plantae
Most ‘primitive’
animals and
plants
→
Social spiders,
isopods,
many insects,
many fishes
Higher birds and
mammals
Often intensive common
parental care, aunt behaviour,
playing groups, and group
defence
→
Isoptera (autapomorphy)
Some Aphidae and Thripidae
At least 14 independent
lineages of Hymenoptera
Eucalyptus ambrosia beetles
(Australoplatypus
incompertus)
Sponge shrimp (Synalpheus
regalis)
Naked mole rats
(Heterocephalus glaber and
Cryptomys damarensis)
All termites (Isoptera).
They have male and
female workers and
different casts.
Some Aphidae and Thripidae
(Homoptera) have sterile
soldiers. Sometimes
rudimentary parental care.
All ants (Hymenoptera).
They have female workers
only and highly
differentiated cast systems.
Two species of mole rats have
non-reproducing workers and a
queen. Colonies have up to
300 members.
Some eusocial Apidae and
Vespidae (Hymenoptera).
They have female workers
only.
Some bumble bees and other
Apidae species may be either
solitary or eusocial depending
on environmental conditions.
What favours Hymenoptera to become eusocial?
Hymenoptera are haplo-diploid organisms
The haplo-diploid system
Queen King Daughter Son Brother
Fertilized eggs become females
Unfertilized eggs become males
Queen
A,B
King
C
Daughter
0.5
0.5
0.75
0.25 0.25
King
0
1.0
1.0
0
0.5
Queen
1.0
0
0.5
0.5
0.25
The diploid-diploid system
Son
A
Daughter
A,C
Son
B
Daughter
B,C
Hamilton’s rule of inclusive fitness
rB  C
Queen King Daughter Son Brother
Daughter
0.5
0.5
0.5
0. 5 0.5
King
0
1.0
0.5
0.5
0.5
Queen
1.0
0
0.5
0.5
0.5
Queen - daughter
0.5 
C
B
Queen - sister
C
0.75 
B
Given that costs and benefits of reproducing are similar it pays for a hymenopteran female more
to invest in her sisters than in her own brood.
This explains why eusocial Hymenoptera all have sterile female workers and never sterile males.
For instance a hymenopteran female helps her sister at the cost of no reproduction.
At equlilibrum the number of surviving offspring should be 2. Hence C = 2
The sister raises one additional offspring
0.75 
2
 0.67
2 1
0.5 
2
 0.67
2 1
Even for one additional offspring of the sister it pays to resign of own offspring
But be careful
Most of the haplo-diploid Hymenoptera are solitary.
The theory requires that queens a priori invest more in daughters than in sons.
Interestingly, many Hymenoptera are able to decide whether to lay male or female eggs.
They are able to control sex ratios
Termites are diplo-diploid
Eusociality and
monogamy
Phylogenetic analysis
shows that all
ancestral eusocial
hymenopteran
species were
monogame.
Polygamy has derived
after the transition to
eusociality.
Polygamy never
occurs in species with
totipotent workers.
From Hughes et al. 2008
Today’s reading
The game theory site:
http://www.holycross.edu/departments/biology/kprestwi/behavior/ESS/ESS_index_frms
et.html
Selfish gene theory: http://en.wikipedia.org/wiki/Gene-centered_view_of_evolution
The evolution of eusociality:
http://www.thornelab.umd.edu/Termite_PDFS/EvolutionEusocialityTermites.pdf
Biology and sexual orientation:
http://en.wikipedia.org/wiki/Biology_and_sexual_orientation
http://www.newscientist.com/article/mg20427370.800-homosexual-selection-thepower-of-samesex-liaisons.html
Biased sex ratios in man: http://huli.group.shef.ac.uk/lummaaproceedins1998.pdf
and http://www.jstor.org/cgibin/jstor/printpage/00664162/di975349/97p0109i/0.pdf?backcontext=page&dowhat=Ac
robat&config=jstor&[email protected]/01cce4405a00501c7b1f1&0.pdf
and http://en.wikipedia.org/wiki/Gender_imbalance
Figs and fig wasps: http://www.figweb.org/Interaction/index.htm