Behavioral Ecology attempts to understand behaviorin terms of its functional design. - Why has it evolved?

Hyp: Behavior is adapted to ecology by evolution

Selection within generations, ofvariants that perform better &heritable response across gen’s:

•Depends on survival and rate of repro given alive.

• Approx.: etc. or timeshort

babiesor

lifetime

babies

clutch

babiesor),r(),R(

•Often approx by short-term indices of performance that are

investedenergyortime

harvestedenergynet

)( expected to be correlated with fitness, like

•All of these measures of performance can be thought of as or efficiency investment

return Microeconomics of max. some currency by optimizing tradeoffs: constrained costs & benefits

Fitness (exp. lifetime repro of genotype) is ultimate measure of performance.

A simple optimal decision – control problem

Suppose there are two things you can choose to do: this I or that AYour continuous decision (control) variable is: x = how much time you allocate to this I.

You have a constraint : your total time available is T, so if you allocate x to this I you can only allocate y(x) = T-x to that A

{this is a tradeoff constraint}This is a troubling problem because both I & A contribute to your fitness w

and you do so want to maximize your fitness!

Suppose the fitness effects of I and A are additive, such that:

xyAxIxWw If there is an optimal x* that maximizes fitness w, then at x* :

dy

dA

dx

dI

dy

dA

dx

dI

dx

dy

dy

dA

dx

dI

dx

dA

dx

dI

dx

dw

*x

10

rule) (chain 0

“marginal value” of a little more I

“marginal value” of a little more A

A simple optimal decision – control problem

Suppose the fitness effects of I and A are additive, such that:

xyAxIxWw and the rates of return for I & A constant, such that:

dcTbxca

dxTcbxaxWw

Then: which can only = 0 if a=c, unlikely!ca

dx

dw

0at out min0 if

at out max 0 if

xca

Txca

This is why linear systems tend to “max out” at the extremes w/o interior optima, & why “diminishing returns” (nonlinear systems) are so important in decision theory

xxy investment on returnor benefit

costor investmentx

? ; :Suppose 0 maxw

x

xywxw

x

y

dx

dy

x

y

xdx

dy

dx

xdw

*x

02

“marginal rate”at x*

long-term averate given x*

In the optimal load size problem, Marginal Value Theorem says: add units (of time & load) of diminishing value until next lower unit would be lower than ave rate of return from just better units.

Predict foragers: more choosy when good types more abundant, less choosy when good types less abundant.

Optimal diet algorithm for discrete food types says same thing: keep adding types of decreasing profitability (e/t) until next lower type is less profitable than ave rate of return from diet of search and harvest only better types.

Does the Optimal Diet Algorithm help us understand the behavior of real animals?

Sort of!

But, the model assumes perfect knowledge,

andpredicts discrete thresholds

(no partial preferences)

Always find partial preferences• imperfect knowledge• lag while learn• variation in prey & predators

Many animals become more or less choosy as best food becomes more or less abundant - qualitatively correct.

Central place foraging in starlings - while feeding babies in nest:

• Postulate parents efficient - in some sense

• Hyp: max. a simple currency:

triptime

tripprey

time

prey

•Predict decision variable = load size given travel time & loading curve constraints.

What load size maximizes possible

?slopex

y

time

prey

Marginal Value Theorem:At opt, slope of loadingcurve (marginal value)equals long term averagerate of return (per trip)

Loading curve,cumulative harvest in patch diminishing returns:

• prey depletion• prey evasion• load interferes

tp

time in patch

Too

sho

rt

Too

long

trip

preyy

trip

timex

x

y

tt

travel time to & from

patch of prey

Does Marginal Value Theorem help us understand central place foraging -

load size in starlings?

Sort of!

Kacelnik exp’s: vary distance from nests to artificial patches & vary loading curves by dribbling out mealworms at different, diminishing rates

trip

preyy

trip

timex

tt tp

Predict:• if lower harvest rate • stay longer, smaller load

• stay longer, bigger load

Observe: predictions qualitatively correct,

but not quantitatively correct

More recent work indicates that starlings don’t measure time or amount on accurate, objective scale.

• if longer travel time tt

In the optimal load size problem, Marginal Value Theorem says: add units (of time & load) of diminishing value until next lower unit would be lower than ave rate of return from just better units.

Optimal diet algorithm for discrete food types says same thing: keep adding types of decreasing profitability (e/t) until next lower type is less profitable than ave rate of return from diet of search and harvest only better types.

The math boils down to this: given an opportunity to pursue a low quality option (2: pick up penny), should it be rejected because search for better options (search for 1: quarters) expected to be more profitable in long run?

Predict foragers: more choosy when good types more abundant, less choosy when good types less abundant

Reject poor type (2: penny) if:

)1(

25

)(1

1

111

1

2

2

shsh ttt

e

t

e

exp. value of search,opportunity cost ofpursuing poor type

Reject poor (2: pennies) if good (1: quarters) are abundant:

t s1 < 24 value poor type < exp value of search

Benefit ofpursuingpoor type

Foraging theory: Hyp: animals have evolved under selection to be economical-efficient foragers.

Adjust choosiness to max. expected return?

less choosy more choosy

ave value acceptable item

Acceptable items time

Value value items time item h time= x

Optimum

reject poor and worse accept better and best

else waste time handling else waste time searching

In richer environ, be more selective:

•If warm & fat - risk averse, prefer average meal w/o variation•If cold & hungry - risk prone, prefer to gamble on variable meals

Why?

The utility (value) of resources is not proportional to the quantity.

util

ity

quantity

Twice as muchmore than

twice as useful

lose ave win

u(mix, wins & loses)

u(ave quantity)

Gambling makesmore sense if u(ave) is low, little to lose

Twice as muchless than

twice as useful

u(mix, wins & loses)

u(ave quantity)

Gambling makesless sense if u(ave)

payoff is high,little to gain

Jen

sen’

s in

equ

alit

y

Lottery tickets - a tax on the poor!

yellow-eyed juncos (Caraco et al. 1990)

Recent advances in our understanding of risk-sensitive foraging preferencesrisk-sensitive foraging preferences.Bateson M. Proceedings of the Nutrition Society 61 (4): 509-516 NOV 2002... animals are sensitive to the variance associated with alternative food sources ...Whether animals are risk-averse or risk-prone appears to depend on a range of factors, including the energetic status of the forager ...

Animals move to improvemove to improveprospects for survival and/or repro.

At ESS no benefit to moving.

Defend if: benefits > costs (less competition) (of aggression)

Habitat selectionHabitat selection - where toexploit resources

&TerritorialityTerritoriality - when to defend

1. 1. Habitat selectionHabitat selection: simple model: IdealIdeal freefree - individual choices

If equal demand/player & each player max’s expected return:If equal demand/player & each player max’s expected return: -population equalizes supply/demand across patches-population equalizes supply/demand across patches -ratio of players -ratio of players matchesmatches ratio of supply & ratio of supply & players get equal returnplayers get equal return

Expected supply/demandsupply/demand or service/customer

1/3

1/2

IdealIdeal - know alternatives - know alternatives

FreeFree - to choose best option - to choose best option

A gamegame : each player’s moveeach player’s move should depend on other player’s movesshould depend on other player’s moves

?

(ideal free habitat selection - continued)

Harper’s ducks:Harper’s ducks:

Two patches of bread bits supplied at equal rates:

Two patches of bread bits supplied at 2 to 1:

Many other exp’s, esp w/ fish, show similar match supply/demandmatch supply/demand

Relative number ducks matchesmatches relative supply rate

Relative number ducks matchesmatches relative supply rate

(Harper’ ducks - continued)

Ideal? How?Ideal? How?

Initially duck pop matchesInitially duck pop matches proximate cuesproximate cues - throws,

1/2 ducks (=16) in patch 1

EventuallyEventually ducks figure new relationship between

prox. cue = throws andult. payoff = bread

relative demand matches rel supplyrelative demand matches rel supply,1/3 ducks (=11) in patch 1

Can’t assume ideal, esp. in novel, altered environments.Can’t assume ideal, esp. in novel, altered environments.

Exp: Same # throws, but 2 bits per throw in patch 2

12

Gersani et al. 1998.Density-dependent habitat selectionDensity-dependent habitat selection in in plantsplants. !!!!!!Evolutionary Ecology 12: 223-234.

We split the root of a young pea … so that half grew in one pot and half in an adjacent pot ... “fence-sitter” … Each root-half was exposed either to no competition in its pot or to competitor plants sharing its pot. There were one, two, three or five competitor plants.

The fence-sitter shifted its root system from the pot with competition to that free of competition in proportion to the number of competitors … so that the ratio between the roots was similar to the ratio between the resources in the pots.

This result is analogous to the habitat-matching rule of the ideal free distribution of populations … plants invest in each of their roots until plants invest in each of their roots until the uptake rate per unit root biomass the uptake rate per unit root biomass is equal for all roots.is equal for all roots.

Abstract:Pea plants exhibit density-dependent habitat selection ...

(ideal free habitat selection - continued)

Expected supply/demandsupply/demand service/customer

1/4

1/5

Twice as big

1

22

1 1 1

Differences in demand/individual (or competitive ability) shouldDifferences in demand/individual (or competitive ability) should result in matching of - demand to supply but result in matching of - demand to supply but

- not individuals to supply- not individuals to supply - individuals do not get identical returns - individuals do not get identical returns

Unequal competitors?Unequal competitors?

?

Too many fish are inpoor patch {undermatching}

But, if measure competitive demand of

individuals find:demand matches supplydemand matches supply- many poor competitors

(slow eaters) in poor patch

Input match - 1 to 2 (=.33) Random - at start

Foraging site selection by juvenile coho salmon:Ideal free distributions of Ideal free distributions of unequal competitorsunequal competitors.Grand, TC. 1997. Animal Behaviour 53:185-196.When individuals differ in competitive ability, ideal free distribution theory predicts that animals should be distributed between habitats such that the distribution of their relative competitive abilities (or 'weights') matches the distribution of resources.

Empirical support for a despotic distributiona despotic distribution in a California spotted owl population Zimmerman et al. 2003. Behav. Ecol. 14:433-437.

{contrast w/ IFD – individuals defending best territories have higher repro success}

Territorial species such as the spotted owl are predicted to follow an idealideal despoticdespotic distributiondistribution. However, debate exists on whether wild populations actually meet the assumptions of an ideal distribution, such as perfect perceptual abilities (i.e., the ability to recognize high- and low-quality sites without error). We investigated whether occupancy rates of California spotted owl territories … positively correlated with a qualitative "potential fitness" (denoted by λpf) estimated from survival and reproduction of territorial owls. [We calculated occupancy rates of individual spotted owl territories as the proportion of years that a territory was occupied by a pair of owls. We calculated a potential fitness value (λpf) for each territory from estimates of survival and fecundity of owls that occupied those territories. ]Spotted owls in our study tended to occupy territories with the highest λpf, supporting the assumption of ideal perceptual abilities. However, this relationship was noisy …

Science News of the Week {Science’s News & Views}Birds Spy on Neighbors to Choose Nest Sites. Jay Withgott Information is power, even for birds. Faced with tough choices, animals that know how others have fared in comparable situations can make better decisions.

On page 1168, {Doligez e al. 2002. Science 297:1168-1170.} researchers report that collared flycatchers decide where to nest and whether to return the next year based in part on knowledge of their neighbors' reproductive success. …A team … tested experimentally to what extent birds make use of information gleaned by watching their neighbors, which ecologists call "public informationpublic information.“The researchers took nestlings from some nests and added them to others, creating some plots of woodland with supersized broods and others with measly numbers of young.

The manipulation had a marked effect. Outsiders preferentially moved to to plots augmented with nestlings, apparently judging these plots to be productive. {immigration}

But youngsters on these plots were smaller. EEmigrants picked up on both cues …They fled both treatment plots at high rates …

Public information: From nosy neighbors to cultural evolution. Danchin et al. 2004. Sci.305: 487-491.

Predators (foragers) respond to distribution of preyPrey respond to predatorsrey respond to predators (and counter moves) (and counter moves)

Food (bottom up)Food (bottom up) and the risk of predation (top down)the risk of predation (top down) both contribute to fitness, but calories and the probability of death are in different currencies, so optimization models have to work with the common currency of fitness which is often harder to measure and model that calories in foraging models.This is where dynamic programming modelsdynamic programming models play an important role.

Predation risk breaks size-dependent dominance in juvenile coho salmon ...and provides growth opportunities for risk-prone individuals.Reinhardt UG. 1999. Can. J. Fish. & Aquat. Sci. 56:1206-1212.

Groups of coho salmon … fry in stream tanks formed size-determined dominance hierarchies … smaller fish occupying inferior feeding positions.

… … simulated predation risksimulated predation risk … … allowed small fish to gain access allowed small fish to gain access to better feeding positionsto better feeding positions.. … …

foo

d

Shocking electric kingfisher model

Reinhardt continued ...

Smaller fry that chose to be in the exposed pool had greater growth rates than those that mainly occupied the refuge, …

… the presence of predators creates opportunities for the expression of alternative behavioural strategies ...

Differences in risk taking and growth among small coho fry … may support … a split into different life history trajectories.

far nearfar near

The little Miss Muffet effect: Quantifying the effect of predation risk on foragingThe little Miss Muffet effect: Quantifying the effect of predation risk on foraging Riley C, Dill LM JOURNAL OF INSECT BEHAVIOR 18 (6): 847-857 NOV 2005 The ability of the ideal free distribution (IFD) to predict patch choice of female houseflies was determined by examining their distribution between two patches containing unequal amounts of food. {sugar cubes}The effect of predation risk was quantified by examining fly distribution between patches of equal food, with one containing spiders. Results were used to predict how much extra food must be added to the risky patch to offset the risk of predation.

Flies were found to conform fairly closely to the IFD.

Predation risk had a major effect on their distribution, with fewer flies feeding in the presence of predators.

Addition of extra food to the risky patch was successful in offsetting the risk of predation.

Non-lethal effects of predators on prey:Non-lethal effects of predators on prey:Predator-induced morphological changes in an amphibian: Predation by dragonfliesPredation by dragonflies affects tadpole shape and color.affects tadpole shape and color.McCollum SA, Leimberger JD. 1997. Oecologia 109: 615-621.

Abstract:… Gray treefrog (Hyla) … tadpoles reared with predatory dragonfly … larvae tadpoles reared with predatory dragonfly … larvae differ in shape and colordiffer in shape and color from tadpoles reared in the absence of dragonflies.

By exposing tadpoles to tail damage and the non-lethal presence of starved and fed dragonflies, we determined that these phenotypic differences these phenotypic differences are induced by non-contact cuesare induced by non-contact cues present when dragonflies prey on Hyla.

The induced changes in shape are in the direction that tends to increase swimming speedincrease swimming speed; thus, the induced morphology may help tadpoles evade predators. …

McCollum SA, VanBuskirk J. 1996. CostsCosts and and benefitsbenefits of a predator-induced polyphenism in the gray treefrog Hyla chrysoscelis. Evolution 50:583-593.

Abstract:… gray treefrog (Hyla chrysoscelis) tadpoles … reared in ponds with predatory dragonfly larvae are relatively inactiverelatively inactive … and have relatively large, brightly colored tailfinsbrightly colored tailfins with dark spots along the margins.

… induced phenotypesinduced phenotypes such as this should confer high fitness … when in the presence of predators, but should be costly when the predator is absent.should be costly when the predator is absent.

Our study tested for the predicted fitness trade-offtrade-off … by first rearing tadpoles in mesocosms under conditions that induce the alternate phenotypes, and then comparing the performance of both phenotypes in both environments.

… Tadpoles from the two environments showed significantly different behavior, tail shape, and tail color within two weeks of exposure. ...

McCollum & VanBuskirk continued

These results confirm that the predator-induced phenotype … is associated with fitness costs and benefits that explain why the defensive phenotype is induced rather than constitutive.

… in ponds where there was no actual risk of predation … both phenotypes grew at the same rate, but thethe predator-induced phenotypepredator-induced phenotype had significantly lower survival than the typical phenotype, indicating that induced tadpoles suffered greater mortality from causes other than odonate predation.suffered greater mortality from causes other than odonate predation.

%

surv

ival

We tested the susceptibility of both phenotypes to predation by exposing them to dragonflies in 24-h predation trials. The predator-induced phenotype showed a significant survival advantage in these trials. showed a significant survival advantage in these trials.

Note that phenotypic plasticity allows ‘best of both worlds.’

“hit” jargon from“miss” signal detection theory

“correct reject”

“false alarm”

Species displaying

snakebehav

display (min)

Time to aband

(h)

Dist snake moved (m)d

Chipmunk Hunt 10 2 50

Chipmunk Hunt 2 4 5

Thrush Hunt 19 1 275

Squirrel Hunt 11 6 106

Squirrel Hunt 12 1 150

Squirrel Hunt 14 11 3

Chipmunk Bask 13 10 3

Chipmunk Bask 20 28 6

Chipmunk Bask 6 12 4

Chipmunk Bask 25 11 13

Chipmunk Bask 12 20 3

Squirrel Bask 31 10 4

Pursuit-deterrent communicationPursuit-deterrent communication between prey animals and timber rattlesnakes: between prey animals and timber rattlesnakes: the response of snakes to harassment displays the response of snakes to harassment displays Clark RW 2005. Behav Ecol & Sociobiol 59 (2): 258-261.

... After receiving displays, foraging {hunt} snakes left their ambush sites and moved long distances before locating subsequent ambush sites, indicating that they responded to displays by abandoning attempts to ambush prey in the vicinity of signalers. {Basking snakes keep basking}

Most analyses of prey-predator communicationprey-predator communication are incomplete because they examine only the behavior of the prey. For example, research on interactions between rodents and rattlesnakes has focused on the behavior of rodent signalers, while responses of snakes have been virtually unexamined. Rattlesnakes are ambush predators ...