Transcript Evolution - Language Log

Road Map
• Evolution:
variation + selection = adaptation
example: LDH in killifish
• Evolved behaviors
– Tropisms in bacteria & fruit flies
– Gibbon songs
• Instinctive learning
– Zebra finches, humpback whales
• Culture as evolution?
– “memetics” vs. genetics
– arguments pro and con
• Evolution of group behavior
– Genetic
– Cultural
10/7/02
Genetic variation and adaptation
• Genetic variation is ubiquitous
• Is a variant worse, neutral or better?
– often this depends on the environment
– Including the attitudes of conspecifics
• In a given environment:
– some variants may be “selected for”
• i.e. produce more offspring
– Result: the population adapts to the environment
– (though much variation is selectively neutral
• The population is (part of) the environment
– so the population adapts to interact with itself
– especially
• in social species
• mating behavior
10/7/02
Adaptation in Fundulus Heteroclitus
(Mummichog)
• Also known as “killifish”
• 5-7 inches, olive green to blue back, white belly
• Lives in salt marshes and tidal creeks
from Labrador to Mexico
10/7/02
LDH
• Lactate dehydrogenase
• Enzyme that catalyzes conversion of
lactate to pyruvate
• A key step in energy production in living
things
• Often the limiting factor in speed or
endurance
10/7/02
LDH variants
• All enzymes have varying effectiveness at
different temperatures
• Some variants work better at lower
temperatures, some work better at higher
temperatures
• Fish are cold-blooded, so their body
temperature depends on the water
temperature
• Mummichog genetic LDH variants
– vary in effectiveness at different temperatures
– are selected according to location (I.e.
temperature)
– result: adaptation to different environments
10/7/02
Mummichog LDH efficiency
The mummichog B locus for lactate dehydrogenase (LDH) has two
common alleles, LDH-Ba and LDH-Bb.
Catalytic efficiency varies with temperature:
the bb genotype “wins” at lower temperatures,
while the aa genotype “wins” at higher temperatures.
10/7/02
Mummichog LDH geography
In the Gulf of Maine, nearly all mummichogs have the
genotype bb at the LDH-B locus.
Off South Carolina, the aa genotype rules.
In between, it’s in between.
10/7/02
Adaptation can be rapid
• In Long Island Sound, the aa (warm-water)
genotype has an overall frequency of 15%
• But for Mummichogs living in the thermal
effluent of a power plant, aa frequency is 30%.
• Efficient adaptation to environmental variation: a
mummichog micro-habitat
• Number of generations: <20
• Like beak variation among Darwin’s Finches on
the Galapagos Islands
– major adaptation to climate change in <5 generations
10/7/02
Bacterial taxis
Taxis: “ability to move in a desired direction”:
Phototaxis: towards or away from light
Chemotaxis: towards or away from chemicals
Magnetotaxis: with respect to magnetic field
Simple but important behavior:
• finding food
• avoiding harm
In colonial and multicellular creatures,
chemotaxis is crucial in creating structures
(“morphogenesis”)
10/7/02
Basic mechanisms of bacterial taxis
• Alternation of (straight-line) “runs”
and (random) “tumbles”
– 3D random walk [animation]
– attractants bind a chemoreceptor
that increases time between tumbles
– repellants decrease time between tumbles
• About 40 genes in the basic system
– plus additional genes for new sensors
– Can orient towards or away from:
• nutrients, poisons, oxygen, light, pH, temperature, etc.
• Categories (which chemicals? which wavelengths? etc.)
and responses (towards or away?)
evolve adaptively
10/7/02
Rhodospirillum centenum
likes it warm and dark…
10/7/02
Tropisms in higher organisms
• Mechanisms are more complex
– (and largely unknown)
• Result is similar
– attraction towards some things
– repulsion from others
• Such preferences may be “learned”
– by individuals: what we normally think of as
learning
– by (sub-)species: through genetic variation and
selection leading to adaptation
10/7/02
Natural behavior variation in
Drosophila Melanogaster
Fruit flies show natural variation in preferences
Pupae hatching in experimental “habitat maze”
– separated themselves on four dimensions of preference
•
•
•
•
up/down
light/dark
acetalldehyde/ethanol
time of the day of emergence (early/late)
– 16 preference combinations
end up in different places in the maze
10/7/02
(Un)natural selection
leads to “speciation”!
• Two “opposite” strains were mated for
25 generations
– early/up/dark/acetaldehyde vs.
late/down/light/ethanol
• Result: reproductive isolation and
habitat specialization
• Conclusion: behavioral preferences
were genetically mediated
10/7/02
Gibbons
• Arboreal apes
– tropical rain forests of southeast asia
– 12 species in four (sub-)genera
• subgenera are somewhat more different
than humans and chimps
– brachiation
– monogamy
• like 3% of mammal species
• 90% of bird species
10/7/02
Gibbons and us:
Primate Phylogeny
Among the apes, only gibbons and humans have pair bonding.
Also, only gibbons and humans sing…
10/7/02
Gibbon duetting
All species of gibbons are known to produce elaborate,
species-specific and sex-specific patterns of
vocalisation often referred to as "songs" (Haimoff, 1984;
Marshall & Marshall, 1976). Songs are loud and
complex and are mainly uttered at specifically
established times of day. In most species, mated pairs
may characteristically combine their songs in a relatively
rigid pattern to produce coordinated duet songs. Several
functions have been attributed to gibbon songs, most of
which emphasise a role in territorial advertisement,
mate attraction and maintenance of pair and family
bonds (Geissmann, 1999; Geissmann & Orgeldinger in
press; Haimoff, 1984; Leighton, 1987).
10/7/02
The female “great call”
The most prominent song contribution of female gibbons
consists of a loud, stereotyped phrase, the great call.
Depending on species, great calls typically comprise
between 6-100 notes, have a duration of 6-30 s. The
shape of individual great call notes and the intervals
between the notes follow a species-specific pattern.
. A female song bout is usually introduced by a variable but
simple series of notes termed the introductory sequence;
it is produced only once in a song bout. Thereafter, great
calls are produced with an interval of about 2 min. In the
intervals, [are] so-called interlude sequences consisting
of shorter, more variable phrases … The typical female
song bout hence follows the sequential course
ABCBCBCBC…,
10/7/02
Male duet contributions
As a rule, adult males do not produce great calls, but "male short
phrases" only. Whereas female great calls remain essentially
unchanged throughout a song bout, males gradually build up
their phrases, beginning with single, simple notes. As less
simple notes are introduced, these notes are combined to
increasingly complex phrases, reaching the fully developed form
only after several minutes of singing …
During duet songs, mated males and females combine their song
contributions to produce complex, but relatively stereotyped
vocal interactions… Both pair partners contribute to an
introductory sequence at the beginning of the song bout (A).
Thereafter, interlude sequences (B) and great call sequences
(C) are produced in successive alternation…
During great call sequences the male becomes silent and does not
resume calling until near or shortly after the end of the female's
great call, when he will produce a coda.
10/7/02
Gibbon song samples
• Hylobates Lar
– white-handed gibbon
– Female “great call”
with male “coda”
• Hylobates Muelleri
– gray gibbon
– Female “great call”
with male “coda”
10/7/02
H. Lar
Hybrid Songs
H. Muelleri x H. Lar:
H. Lar x H. Muelleri:
H. Muelleri:
10/7/02
Phylogeny of singing in primates
Singing is rare in mammals. It occurs in members of 26 species in four primate genera:
Indri, Tarsius, Callicebus, Hylobates. These are 11% of primate species and 4% of
primate genera. Since the four singing genera are widely separated, they are thought to
have evolved singing independently.
In all singing primates, both males and females sing, and duetting usually if not always
occurs. All singing primates are monogamous (with the possible exception of humans).
Most bird species sing; often bird song is mostly male; duetting bird species are also
usually monogamous.
10/7/02
Are humans monogamous?
Are humans monogamous?
In most mammalian species, sexual access is either determined
by rank… and results in polygyny; or else … two individuals
become “attached” to one another and then isolate themselves
from other members of their species…
[In humans] what is common is… cooperative, mixed-sex
social groups, with significant male care and provisioning of
offspring, and relatively stable patterns of reproductive
exclusion, mostly in the form of monogamous relationships.
Reproductive pairing is not found in exactly this pattern in any
other species.
10/7/02
--Terence Deacon, “The Symbolic Species”
Gular sac
Some gibbons have developed a large
“gular sac” apparently involved with
breath control and/or resonance.
Gular sac size and song complexity
seem to correlate across species.
Symphalangus syndactylus
(siamang):
“the [siamang] duet is probably the
most complicated opus sung by a
land vertebrate other than man…”
--Marshall and Sugardjito (1986)
10/7/02
Zebra Finch
(Taeniopygia guttata)
• Small songbird (Australia and Timor)
• Highly social (colonies of 20-1000)
• Pair bonding (with frequent “cheating”)
– Male display, females choose
– bond marked by “clumping” and preening
• Males sing, females do not
– part of sexual and territorial displays
10/7/02
Typical Zebra Finch song
original
slowed x 4
• Not intrinsically pleasing to most humans
• nasal quality, repetitive rhythm
• Production requires difficult motor control and
large expenditure of energy
• Female Zebra Finches (and competing males)
are willing to be impressed
10/7/02
Song learning
in Zebra Finches and other oscines
• Song patterns are species-specific
• However, exposure to adult patterns is necessary for
normal sing development
– deafened birds develop highly degraded song
– birds reared without adult male models
develop degraded but species-typical songs
• Sensitive/critical period for exposure
– 20-35 days after hatching
• Active song develops later
– 60-90 days after hatching
10/7/02
Why learned songs?
• Some sub-oscine species have completely
programmed song
– deafened or isolated birds sing normally
• Suggested advantages of learned song:
– more complex or varied song via cultural rather than genetic
development
• females prefer constrained novelty
• promotes exogamy in large colonies
– intra-species varients of song, called dialects, may serve to
segregate populations of the same species
• promotes endogamy in microhabitats
– more rapid adaptation of the song to different acoustic
environments
10/7/02
Humpback Whale Songs
• Samples:
• Variation and change:
– At any one time, all whales sing similar
songs
– Over time, songs change rapidly
10/7/02
Genetic variation and adaptation
• Genetic variation is ubiquitous
• Is a genetic variant worse, neutral or better?
– often this depends on the environment
– including the attitudes of conspecifics
• In a given environment:
– some genetic variants may be “selected for”
(i.e. produce more offspring)
– producing genetic adaptation to the environment
(though much genetic change seems selectively neutral)
• The population is (part of) the environment
– so the population adapts genetically to itself
– especially
• in social species
• In mating behavior
10/7/02
Cultural variation and adaptation
• Cultural variation is ubiquitous
• Is a cultural variant worse, neutral or better?
– often this depends on the environment
– including the attitudes of conspecifics
• In a given environment:
– some cultural variants may be “selected for”
(i.e. produce more offspring grow or get borrowed)
– producing cultural adaptation to the environment
(though much cultural change seems selectively neutral)
• The population is (part of) the environment
– so the population adapts culturally to itself
– especially
• in social species
• in mating behavior
10/7/02
The culture of military tactics
• From Greek taxis “order, arrangement”
• Two main approaches in the ancient world:
10/7/02
The two patterns
• Heavy infantry: like the Macedonian phalanx
– Associated with city-states in the eastern
Mediterranean
• Light cavalry: like the Mongol hordes
– Associated with pastoral nomads on the Eurasian
steppes
• Choice in many dimensions:
– weapons, armor, tactics, lifestyle
10/7/02
Herders vs. farmers
• Mounted archers
• Light bow, no armor
• Success based on
mobility
– lots of practice in
riding and shooting
• No good at
defending fixed
positions
10/7/02
• Foot soldiers
• 70 pounds of arms
– helmet, shield
– spear, sword
• Success based on
shock of mass
formation
– little training required
• Not very mobile
Geography of two military cultures
10/7/02
Comparison of two failed invasions
• In 514 B.C., the Persian king Darius tried to conquer
the Scythian nomads (roughly in the area of presentday Ukraine). He chased them all over their vast
domains, and finally had to return home in frustration.
• Darius invaded Greece in 490 B.C. He sacked a few
cities, and was defeated by the Athenian infantry
(who were outnumbered 3 to 1) at the battle of
Marathon. The Athenians caught the Persians
unprepared by running a mile in full panoply, which
was an unheard-of tactic at the time.
10/7/02
The Greek historian Herodotus describes both invasions,
and admires the ability of the Scythian nomads to win without
pitched battles:
The Scythians indeed have in one respect, and that the very most
important of all those that fall under man's control, shown themselves
wiser than any nation upon the face of the earth. Their customs
otherwise are not such as I admire. The one thing of which I speak is
the contrivance whereby they make it impossible for the enemy who
invades them to escape destruction, while they themselves are entirely
out of his reach, unless it please them to engage with him. Having
neither cities nor forts, and carrying their dwellings with them wherever
they go; accustomed, moreover, one and all of them, to shoot from
horseback; and living not by husbandry but on their cattle, their
waggons the only houses that they possess, how can they fail of being
unconquerable, and unassailable even?
10/7/02
Summary of the example
• Adaptation of weapons, military tactics
– Asian steppe vs. Mediterranean coast
– Herding vs. farming
– Nomadic bands vs. walled cities
• Continuing cultural variation
– proliferation and copying of what worked best
– result: consistent patterns across thousands of
years and miles
• Process was (entirely?) cultural
– some genetic effects are possible
– ~300 generations
• Cultural adaptation to two ecological niches
– fitness depends on the environment
– cultural and geographical
10/7/02
Genetics vs. memetics
• Many similarities:
evolution as “descent with modification”
• Differences:
– Mechanisms of “descent”
sexual reproduction vs.imitation and borrowing
– Mechanisms of innovation
random mutation/recombination
vs. explicit problem-solving
– Rate of change
10/7/02
How do behaviors and abilities evolve?
• Algorithm development by “descent with modification”
• Some convincing (modeled) examples
• General principles could apply to individual or cultural
learning as well as genetic “learning”
• The mapping from genes to behavior is almost
completely mysterious
(but the mapping from genes to anatomy is almost
as mysterious…)
10/7/02
Example:
Strategies for the “Iterated Prisoner’s Dilemma”
• The “prisoner’s dilemma” paradox
• The “iterated prisoner’s dilemma”
– Proposed in 1984 by Robert Axelrod
• The IPD tournaments
• Genetic algorithms
• Applications of GA to the IPD
10/7/02
The “prisoner’s dilemma” paradox
Origin: situation of captured thieves
– if everybody keeps silent, all go free
– if one confesses
• (s)he gets a reward
• everyone else gets a heavy sentence
– if everyone confesses
• everyone gets a moderate sentence
If you analyze the options objectively, your best bet is
to confess. But if everyone confesses, everyone is
worse off than if everyone kept silent.
Generically: total cooperation is better than total noncooperation; but any individual can then better his or
her situation by “defecting”.
10/7/02
PD payoff matrix
(payoff to me)
Temptation >> Reward >> Punishment >> Sucker’s payoff
[Also: (Temptation + Sucker)/2 <= Reward]
10/7/02
Consider the options…
• If you defect
– if I cooperate I pay $100
– if I defect I pay $10
– so my best bet is to defect…
• If you cooperate
– if I cooperate I get $300
– if I defect I get $500
– so my best bet is to defect…
10/7/02
PD without money or cops:
the “furry critter’s dilemma”
10/7/02
Conclusion: nice guys finish last
• PD arguments were used to “prove” that
cooperation could never be an evolutionarily
stable strategy, except perhaps among kin
• “Every man for himself, and the devil take the
hindmost…”
• A depressing conclusion in the context of the
cold war, where nuclear standoff looks very
much like a PD situation.
10/7/02
Axelrod’s innovation
(1984)
• Treat PD as a game with repeating turns
– Endless, or at least players don’t know when the
end will come
• Add up the scores across turns
• Play different strategies against one another
– Human game-playing
– Better, let the computer do it
10/7/02
Payoff matrix for Axelrod’s game
10/7/02
First try
• Axelrod advertised for strategies
– 14 were submitted, some very complicated
– Axelrod added one: Random
• Run a tournament: every strategy plays against every other
strategy 200 times
• The winner:
– “Tit for tat” (submitted by Anatol Rapoport)
– “Cooperate with strangers, and otherwise do whatever the opponent
did last time around”
• If we define a “nice” strategy as one that is never the first to
defect, then the 8 top scoring strategies (out of 15) were “nice”.
• “Forgiving” strategies do better than those that bear grudges
– in fact, “tit for two tats” would have won if it had been entered
10/7/02
Second try
• Analysis of first tournament was published
• Second tournament was opened to any new
entrants
– 62 entrants this time
– number of rounds was left open
• Result: “tit for tat” wins again
– including against “tit for two tats”
10/7/02
More results
• Many other analyses and competitions
• Two attempts at evolutionary simulations
– (1) Evolutionary competition among fixed set of 63
strategies
• “tit for tat” won 5 out of 6 rounds
• A similar “nice, forgiving” strategy won the 6th
• However, no such strategy is “evolutionarily stable”, in
the sense that a uniform population can always be
successfully invaded by an alternative.
– (2) “Genetic algorithm” to evolve new strategies
and let them compete
10/7/02
Evolving PD strategies
• Strategies as pseudochromosomes
– 4 possible outcomes at each stage of the game
– 4 x 4 x 4 = 64 possible 3-move “histories”
– To determine how to act in each of these 64 cases
requires 64 specification of “C” (cooperate) or “D”
(defect)
– Thus a PD “gene” is a string of 64 C’s or D’s
– Add 6 more to cover the first 3 moves
– Total of 70 letters make up a pseudochromosome
10/7/02
Mutations and mating
• Mutation: just change one of the 70
letters from C to D or from D to C, with
some small probability
• Mating: combine two “genotypes” by
selecting a random crossover point
– crossover point of 9 combines 1-9 from
one parent with 10-70 from the other, and
vice versa
10/7/02
Natural selection
• Run a mini-tournament and score each
genotype
• Mate genotypes (with random
mutations) to produce offspring in
proportion to their score on the previous
round of the tournament
10/7/02
Results: overall population fitness improves
10/7/02
Gene pool analysis
•
Five “alleles” evolved in the vast majority of
individuals:
1.
2.
3.
4.
5.
–
10/7/02
Don’t rock the boat: if RRR then C
Be provocable: if RRS then D
Accept apologies: if TSR then C
Forget: if SRR then C
Accept a rut: if PPP then D
Most of the resulting “individuals”
beat tit-for-tat
Emergent properties
of group behavior
• Descriptions at different levels
– temperature, sound vs. molecular motions
– cell vs. macromolecules
– organism vs. cell
– ant colony vs. individual ants
– marketplace vs. individual merchants
• Some key properties only emerge in
group interactions
10/7/02
Why this matters
• Fitness is often determined by emergent
properties
– Fitness in multicellular organisms can’t be
determined by looking only at individual
cells
– Fitness in social organisms can’t be
determined by looking only at individuals
10/7/02
Ant path selection
10/7/02
Ant problem solving
10/7/02
Termite mound
10/7/02
How social insect behavior evolves
• individual properties are genetically controlled
–
–
–
–
pheremone chemistry
when to emit which pheremones
effects of sensing pheremones
the “Termite Machine” is like a Turing Machine
that reads and writes smells rather than symbols
• intelligent group, stupid individuals
– plans and decisions exist only at the group level
– dispositions of individuals are like elements of a circuit
or statements in a program
• evolutionary environment is (mainly) the group
– like the evolutionary situation of multicelled organisms
10/7/02
Adam Smith’s “invisible hand”
[A merchant] generally, indeed, neither intends to
promote the public interest, nor knows how much he is
promoting it. … [H]e intends only his own gain, and
he is in this, as in many other cases, led by an
invisible hand to promote an end which was no part of
his intention. … By pursuing his own interest he
frequently promotes that of the society more
effectually than when he really intends to promote it.
Wealth of Nations, Book IV
10/7/02