Evolutionary Ecology Final

Plasticity and Environment

1. Not all phenotypic changes over time in a population represent evolutionary responses, they could reflect plastic responses by individuals to a change in the environment. 2. Heritability of traits that are plastic can change between environments. E.G. The heritability of internode length will be estimated to be low in this population at low conspecific densities (when plastic response is low, low GxE) and higher at high conspecific densities (when plastic response is high, high GxE, variation is high), even though we are looking at the same genotype. Heritability of plasticity may also change between different environments. 3. Genetic variation in plasticity of any trait means that plasticity of the trait could also evolve, and so is subject to all the same kinds of evolutionary ecological questions that we may ask

Darwins

Another estimator of evolutionary rate, more applicable for macro-evolutionary studies. measure the proportional amount of change in mean trait in a lineage per million years (arbitrary time scale). Used by paleontologist. Measured as differences in log mean trait value at beginning and end of the time interval measured in million years.

Normal distribution

Can be described by two parameters: mean/average (central tendency) and it's width/standard deviation (spread).

Gene flow

Can bring alleles to new locations, and the alleles may increase or decrease fitness. E.g. if you're a dark mouse moving to a white beach you will get eaten have no babies. Evolution of perfect adaptation is more effective if there is no migration. IF there is a high rate of migration this will cause the two populations to become more similar. If a white mouse and brown mouse mate their offspring will have an agouti gene which prevents dark colour from being expressed, so it will make a light mouse in the forest which is maladaptive.

Broad sense heritability

H^2 - Vg/Vp

Physiological adaptation

Individual organism processes in its lifetime (acclimation, plasticity). Organisms frequently physiologically adapt to their local environmental conditions (sweating to cool down etc). Physiologists regularly refer to adaptive responses by individuals to changes in their conditions. This process occurs in an individual to change itself in its own lifetime. Think of it as more of a acclimation or plasticity, not evolutionary. Occurs in individuals.

Genetic Effect

NOT heritability. The relationship between an individual's genotype and its phenotype, determined through individual development.

Breeder's Equation

R = h^2(S). Therefore how much the population changes depends on selection differential and heritability. R only predicts evolution in a single generation, can't extrapolate further than the data you have.

Genotype

The alleles present at a genetic locus, what is inherited

Oligogenic Adaptive Evolution

The genetic architecture of quantitative traits involves many loci of small effects and few loci of large genetic effect on phenotype. A more dynamically changing form of selection (that replaces mutation as the way that genetic variation is maintained in a population). Fluctuating selection is responsible for maintaining genetic variation.

Senescence

The process of aging. A decline in physiological capacity with age because of reduced reproduction and traits closely linked to survival (physical ability, mental ability etc). It has high evolutionary potential because it may affect individual fitness. For most organisms reproductive potential and survival tend to increase with time early in life, but after a certain age individual reproduction and survival begin to decline with age.

Trait vs. Phenotype

Trait is a character of an organism (e.g. eye colour, height, weight). Phenotype is the measurable trait(s) of an organism, resulting from an interaction between its genotype and its environment (what selection acts on). So we consider a trait when we consider a character of an organism/species, but the phenotype is what we measure and what selection directly acts on.

Selection

differential survival/fecundity of some phenotypes

Selection Thinking

-Observe a trait (clearly identify the trait you want to understand) -Use principle of natural selection as a starting point to generate hypotheses about why a particular trait might exist (and do individuals differ in this trait) -Hypothesize that structures and other features of an organism increase performance at some ecologically relevant task-> speculate on how the trait functions to affect performance -That task increases some component of fitness -> spculate again about how that performance could influence fitness Trait X-> (affects) Ecological performance Y -> Individual fitness Z If data don't support the selection hypothesis then maybe our feature is not an adaptation (could be by-product of previous evolution or selection on correlated trait, or maladaptation aka constraint of evolutionary history and physiology) not everything is an adaptation.

Issues w Studying Senescence

1. Data can be hard to get. Sample sizes are smaller in older age classes. This causes a statistical consequence, our estimates about features of these older classes are less reliable (less precise) 2. Natural selection acts on a cohort of individuals as they age. This means that the poor quality individuals will be removed from the cohort as they age. So the surviving individuals are more homogeneous. Imagine burst speed and predator avoidance success vs. age, the older individuals may have better escape bc that's what allowed them to live so long. We may have a biased estimate of senescence traits if we study old survivors. This will be hard to fix.

Two ways to estimate linear Selection (S)

1. Estimate S as the slope of a linear regression of standardized values of relative fitness (w) vs. phenotype (z) for each individual. w = B + Sz 2. Compare mean phenotypes between samples taken "before" vs. "after" selection in the population. Data: frequency of phenotype (z) before and after selection. S=Mean(after)-Mean(before) NOT the difference in phenotpye between those that died and those that survived. Linear selection affects the distribution of the trait in the population by shifting the mean phenotype in the direction of the slope. Will also reduce phenotypic variance in a population by removing types preferentially from one end of the phenotype spectrum.

Other mechanisms that drive phenotypic change in a population over generations

1. Sexual selection acting on heritable variation (genetic change in a pop'n over generations) 2. Drift (chance effects on fitness: genetic change in populations over generations) the effects of chance is much stronger in small pop'ns, in large pop'ns chance events act on many individuals and tends to average out 3. Non-genetic changes: cultural evolution, multi-generational maternal effects, plastic phenotypic response to changing local conditions. Features are changing but not evolving per se. -E.g. culturally transmited information causing phenotypes to change through learning or multi-generational maternal effects from one generation to the next. -phenotypic plasticity: members of some pop'ns change in their own life times to changes in their local environment. So if the environment that a population exists in changes over time then the population will change too because of these plastic responses (e.g. more food becomes available so people grow taller) If we want to study adaptive evolution we have to rule out options 2 and 3

Natural Selection

1. Variation in some trait occurs among individuals 2. Variation in the trait among different individuals influences fitness through an effect on the ability to perform some ecologically relevant task 3. Results in a change in frequency of the trait that influences frequency within a generation (the distribution of phenotypes in a cohort changes within the lifetime of a cohort) Trait-> Performance-> Survival or Reproduction Conditions 1 and 2 cause natural selection (3) to occur in a population

Selection Differential Standardization

1. the magnitude of the slope (S) of the fitness function is determined by the scale of measurement of the trait. The simple differences in the 'size' of each trait has a large effect on the slope of the fitness function (relative fitness vs. trait). WHen you change the scale of measurement by changing units or using different traits, then the slope of the fitness function changes too. Therefore we use a standardized value of the trait (z). The variation in the trait will now we centered on 0, and negative and positive values. 2. Selection acts through fitness differences among the individuals in a specific population not on absolute fitness, so get mean as absolute/mean fitness of individuals in the population. The most fit has a value of 1, and less fit having one less than 1.

Case Study: Testing agents of selection in the medium ground finch

13 species of finch on the galapagos, distinguished ecologically by principal resources. Galapagos experiences wet and dry seasons each year, but occasionally wet seasons are missed and drought lasts a year. Hypothesized that dry season reduces availability of seeds causing competition that favours individuals with larger beaks you can eat harder seeds. Seed availability and hardness are the source of selection on finch beaks, therefore population abundance should fluctuate with seed availability and mean surviving beak size shifts to larger beaks during/following drought years. AKA the members of the population are affected by variation in the source, and the distribution of traits in the members of the population varies as the source of selection varies. Remember, with only one occurrence of correlation between seed hardness and beak size you have a sample size of one , therefore the effect could be by chance and we need replication. It is possible that in wet years the seeds are smaller, and therefore its harder to crack a small seed with a big beak, so in wet years there is selection for smaller beaks. This could be better evaluated if you ad a bird feeder and manipulate the availability of differently size seeds.

Reznick's Transplant Experiment

200 guppies were transplants from high predation pool to pool above water fall, after only 4 years (males) and 7 years (females) male guppies were older and larger at maturity and females had larger young. Males evolve faster, perhaps more genetic variation. Found guppies in high predation pools had faster swimming burst performance than those that had be introduced into low predation environments. IN low predation environments pregnant female guppies can become fatter and still escape predation

Fitness function vs. Selection

A fitness function is the relationship between the trait values of fitness of individuals (w and z). There has to be a slope to the relationship for it to be selection. If there is no slope it is a fitness function but it is not selection. If there is a slope it is selection and a fitness function.

Fitness or Adaptive Landscape

A graph with fitness as the y axis, but there are two x axes (one x and one z) so that fitness functions of two traits can be described. This is one of the best ways to think about the evolution of combinations of traits (lots of traits must evolve together, like the use of multiple fingers for example). Two phenotypic traits may evolve in a way that both are required to provide functional performance for the organism. We can now consider these traits in 3D and wonder about the shape of the surface made between the two single trait fitness functions. What does a fitness landscape look like? Anything at all, it depends on the functional relationships between the traits and the effects of that relationship on fitness. The height for any combination of the two traits indicate that combination's fitness relative to any other combination. Fitness landscapes express how selection could act on combinations of traits in a population of individuals placed anywhere on the landscape. If a single point in the center is higher than all others selection is stabilizing for mid values of both traits. If it varies a bunch it is complex landscape of many possible good combinations. These are complex because the fitness function for one trait depends on the specific value of the other trait.

Genetic correlation among trait measured in two environments

A trait measured in each of two environments can be statistically treated as two different traits. Can then consider genetic correlation among the two trait as index of their evolutionary independence. If genetic correlation is positive then its easier to select both traits to be larger. Conversely if genetic correlation in negative then can select in opposite directions. Traits measured in two environments can be thought of as two genetically correlated traits. A trait measured in two environments can be thought of as two different traits that are genetically correlated with each other. If zero correlation then traits can evolve independently. If positive can select both to increase. If negative then if you select for one to increase the other may decrease. E.g. Growth rate in two environments

Adaptation vs. Evolutionary Response

Adaptation results when selection acts on phenotypes in a generation causing allele frequencies (that influence phenotypic variation) to change between generations. Quantitative genetics focuses on the genetical aspect of this process. Selection does not require genetic variation, it requires phenotypic variation because it sorts phenotypic variants through the functional effects of phenotype on relative fitness. Evolutionary responses to selection require heritability (correspondence between variation in genotype and phenotype in a population). With heritability, selection sorting among phenotypes can also sort underlying genotypes in a population

Functional Trait Analysis

Also called a performance analysis. Gets directly at whether a trait functionally actually affects fitness.

Limitations of Selection Differentials

An issue with estimating selection is that selection differentials only tell you the statistical relationship between relative fitness and a phenotype. By itself it tells us little about the ecological factors that cause the variation in fitness among individuals. Determining the selective agents takes more ecological investigation. Determining the targets of the selection, the phenotypic traits that selection acts directly on, also takes more work because we have to establish the functional utility of the trait relative to the source of selection. The key to discovering if a trait is a target of selection is to seek evidence that the trait functionally influences fitness of individuals under selection.

Causes of variation among individuals in a population

Any natural pop'n is composed of individuals with differences in their shared features (coloration, body size, body proportions, metabolic rates, maturation schedules) Causes: 1. Ontogeny -development-> differences based on age and development stage (e.g. babies heads being bigger than bodies when born and grows slower than bodies) phenotypic features change in most organisms as they develop from immature to mature forms 2. Environment -most organisms have the ability to acclimatize to their environment (e.g. larger adults when nutrition is good) 3. Genetics -of interest because they contribute to heritability (3rd condition for evolution by natural selection) These three factors can influence each other. There can be genetic differences among individuals that affect their development and genetic differences can influence individuals to have different phenotypic responses to environmental conditions.

Maternal Effects on Evolutionary Response

Anything that influences h^2 or S will influence the response to selection. Without maternal effects if the selection remains constant then the evolutionary response will be constant increase in the trait mean over generations, when selection for the trait ceases the trait will no longer evolve. If we add maternal effects there will be an effect on the rate of phenotypic evolution under selection. When maternal effects are positive then the rate of evolution increases. Positive maternal effects amplify evolutionary responses, and when maternal effects are negative then the rate of trait evolution decreases. The effect of increasing or decreases the resemblance of parents and offspring is to sharpen or soften respectively the ability of selection to distinguish phenotypic differences among individuals in the population. In the maternal effects model the effects of selection may be more complex. Selection can act on the trait of interest in every generation but it can also act on any genetic variation in the maternal effects that influence the trait. Some genotypes express strong maternal effects to environment, others weak. The maternal effect may be an evolvable trait of the population that also influences offspring phenotype. So the evolutionary change in a trait that is influenced by maternal effects is the combined effects of selection on the trait and selection acting in the previous generation on maternal effects that influence the trait in offspring. Evolutionary responses in populations are harder to predict when plasticity or maternal effects also are evolving at the same time. Maternal effects influence evolution by affecting the resemblance of offspring to parents and by allowing selection to potentially carry over between generations.

Heritability always depends on Environment

Because h^2=Va/Vp, changing the environment can affect Va, Vp or both. If you put a pop'n in different environments this changes heritability. Suppose that genetic variation in growth in corn is caused mostly by genes for maximum growth rate under good conditions (light, water, nutrients). In this scenario corn growth varies in the good environment, with some profiting more than others, but in the poor environment all are limited by resources and therefore have the same growth. Allele differences in growth rate are only expressed as phenotypic differences in the good environment. Measured under good conditions there is high Va (even though Va is about genotypes it is measured in phenotype) and high Vp. Under poor conditions growth is limited bc there are no stress alleles (might be the reverse of conditions if there are stress alleles and not growth alleles). There is very little Va and Vp. Therefore h2 changes depending on the environment. Heritability of the same trait in different environments will be different. Heritability of two traits in the same population is likely to be different. Heritability of the same trait in two different populations is likely to be different. We can not reliably compare the heritability of a trait between two populations or two environments

Mouse Fur Parallel Evolution

Beech mouse and deer mouse. The lighter coloured subspecies live on light beaches and the darker coloured sub-species live in a darker habitat. Why is blond fur adaptive in beach mice? Why is dark fur adaptive in mainland mice? The majority of the predators use sight and sound to hunt mice, therefore a mouse that blends into its environment may have a better chance of survival. An experiment was set up with beach and mainland habitats w dark and light mice randomly in each, attack was scored by injuries. A QTL analysis of coat color was done for the mice, found 3 regions of the genome associated with phenotypic variation 1) agouti gene, 2) melanocyte-stimulating hormone receptor (mc1R). MC1R has two alleles (R or C) that differ by a single nucleotide subsitution and controls how much eumelanin and pheomelanin you have. Dark hair: lots of eumelanin. Fair hair: more pheomelanin, little eumelanin. Red: lots of pheomelanin. Agouti represses Mc1r causing low signalling to pigment producing cells, so agouti prevents mc1r from inducing pigment. In addition, mutant mc1r C allele does not signal. Beach mice have mutations in Agouti that affect gene expression. There are two populations w subspecies of both colour (same thing for both: dark mice in dark habitat, light mice on beach). Instead of the two light mice being sister taxa, the populations are sister taxa, so light coloured evolved separately in both population. This is classic parallel evolution. The atlantic and gulf coast populations both evolved white fur but the specific mutations responsible are different. In one population mutated agouti causing white colouration boundary to move dorsally in beach mouse, this agouti was not mutated in the mainland mice. The subspecies are called ecotypes.

Variation results in differences in performance, and ecological performance affects fitness

By fitness we usually mean survival and reproduction. These components are often measurable and influence how many offspring an individual can contribute to the next generation. The number of offspring that an individual sends into future generations is a better estimate of fitness but hard to measure sometimes. Fitness could be influenced by.... body size, jaw mass, blood oxygen saturation level. Selection occurs when only some individuals are likely to survive or reproduce (and others not), many aspects of performance can influence survival and reproduction, some directly (probability of survival) and some indirectly (flight performance)

Measuring Genetic Correlations

Can be done by offspring vs. midparent regression but this time for two different traits, unlike with heritability. Response to selection Rcy=ix(hx)(hy)(rgy)(Vpy) r= genetic correlation vp= phenotypic variance

Performance Analysis

Causally linking traits to performance. How does variation in the trait actually affect the fitness? A functional performance hypothesis involves linking a mechanism to the causal link between a trait and fitness advantage. E.g. snail ratio of width to length, "shell shape affects how crushing forces are spread across the shell, more rotund shells have greater ability to resist crushing forces because they are able to spread forces more evenly over the entire shell than elongate shells. Elongate shells cannot as evenly spread crushing forces across the shell and so are more likely to fail." therefore prediction is " elongate shells crush more easily than rotund shells". The weakness of performance trials (e.g. measuring time it took for fish to consume the snail) is that is does not prove that this is what is actually happening in the wild, it just shows the mechanism of the proposed functional hypothesis is plausible. In theory our complete analysis involves three different causal relationships: 1. A selection analysis (selection differential/selection gradient if multiple traits are considered)-> often involves data from individuals in the field 2. Performance analysis of trait variation on performance variation (performance gradient/differential)-> often involves data from individuals in the lab 3. Performance analysis of performance variation on fitness variation (performance gradient/differential) -> often not possible bc we rarely have performance information about individuals that we can relate to fitness under natural conditions. It is sometimes know, as w assignment 1 where tadpoles w known performance measures were put in artificial enclosures w their predators at known densities Answers: is the trait under selection, does the trait influence performance

Direct Vs. Indirect Selection on a Trait

Direct Selection: selection acts on a trait because variation in the trait directly affects fitness through some effect on organismal function and performance. Indirect: some traits are correlated to each other (e.g. individuals with long arms for their bodies often have longer legs as well, so arm length can be under indirect selection because selection acts directly on leg length and arm length comes under indirect selection bc it is correlated with leg length) So S is a great estimate of total selection on the trait, and possible of direct selection on a trait if all indirect components of selection on the trait are 0. Two ways to distinguish direct from indirect selection acting on any trait 1. Use the statistical method of multiple linear regression to isolate the effect of each variable while holding other variables constant 2. Perform manipulative experiments where you vary one trait while keeping the other traits constant

Estimating Selection on Discrete Traits

Discrete traits do not have intermediate phenotypes. Continuous traits have a range of intermediate values between its limits (thought to be affected by many gene loci so they are called polygenic). Discrete traits are only affected by a few or a single loci. We can't use linear regression because there is no continuous variation on the trait Z. % recaptured can be used as a measure of survival. From this data the relative fitness of each trait can be determined (survival potential of the three genotypes). Selection differential on each trait is then measured by the relative fitness of each trait - the relative fitness(1, the average fitness). So the selection differential assessed on discrete phenotypes measures their disadvantages relative to the most fit phenotypes.

Phenotypic Variance (Vp)

Dispersion of measurable trait resulting from an interaction between its genotype and its environment. what selection acts on Vp=Vg +Ve

Evolutionary Ecology

Ecology affects population genetics, adaptation affects behaviour, performance and population demography. Ecology of populations changes with evolution. Adaptive diversification and speciation are driven by this association. Evolutionary ecology is focused on the interplay between the two types of processes. Ecology of organism: the tasks an individual must do in their lifetime to: acquire energy for growth, survive, successfully reproduce. -If diff individuals perform those tasks with different degrees of success then some individuals will survive and leave more offspring-> the ecological process of natural selection acts on ecology of individuals. Natural selection acts on individuals within every generation (bc it is simply a change in frequency of genotypes) -When selection favours some individuals over others and those differences among individuals are heritable then the composition of the population will change over generations-> Evolutionary Process Evolutionary process: change in frequencies across generations due to heritability of selected for genes. Ecology+Heritability=Evolution Tends to result in local adaptation Ecological Process: Natural selection within a generation IN ADDITION: a population's adaptations govern how individuals find food, survive and reproduce-> adaptations govern the behavioral ecology of individuals and govern pop'n demography through effects on birth and death rates (how long they live, how frequently they reproduce) Therefore evolution of traits in a population feeds back and has ecological consequences for that population, so the ecology of a population will change over time.

Evolution

Evolution is when the traits and frequencies of phenotypes (altered by natural selection) are passed onto the next generation. Evolution by natural selection is when trait values tend to be passed from parents to offspring. This results in descent with modification

Population

Evolutionary ecologists would say that populations with significant FST values should be considered demographically independent.

Galapagos Finches

Few species of birds managed to colonize the Galapagos islands, Finch colonists arrived and have diversified into 13 species. Two Grant profs and their grad student studied the island of Daphne Major. Bc the island is small they can band every bird w colour-coded bands they can read with binoculars. Census 3 times a year before, during, and after breeding season. When they band them they also measure seven morphological characters. There is huge variation in beak depths for finch of the same species (medium ground finch) and similar body size. Most birds have medium beak depths though. During El Nino years (very dry) plants do not produce new seeds. Standard error declines as the number of ground finches declines due to drought. The abundance of seeds decreased with drought. The seeds available during the drought are larger, and only birds with the largest bills are strong enough to break open the pods. The average beak depth became deeper over the period of the drought. When there is plenty of water and a variety of seed types it is bad to have a wide beak, so when conditions changed selection reversed the change to wider beaks. It is best to have a narrow and deep beak, but evolution actually acts to make beaks deeper and wider because wideness is correlated to deepness. Selection can only act on the raw genetic material and is sometimes constrained by correlations. For complex ecological characters selection may act on multiple traits simultaneously and a character may not vary independently of other characters when correlated. There are no finches w narrow and deep beaks.

Genetic Adaptation or Genetic Drift?

Fst measures the degree of divergence among populations for neutral loci (effects due to drift and migration). For neutral alleles (not under selection). Qst measures degree of genetic divergence among populations for quantitative traits predicted to be under selection (requires common garden data). If FST=QST there is no selection, only drift, bc neutral and selected for alleles are diverging at the same rate. if QST>FST selection is occurring for divergence between the populations because selected for alleles are diverging more than neutral alleles. Both are estimated by taking the variation within a population and comparing it to the variance between populations if FST>QST Evolution is due to natural selection that favours the same phenotype in two populations (both are becoming less divergent from one another in their selected for alleles) Is is only accurately telling us what is happening in these populations if the amount of mutation in neutral traits is equal to that in quantitative traits. This is not necessarily true, neutral traits are often single locus traits and quantitative are nearly always polygenic, therefore quantitative traits have the chance to mutate just like the gene for the neutral trait. On average then the polygenic traits experience higher levels of mutation compared to neutral traits simply bc they are made up of more traits. This accounts for QST>FST but not FST>QST. IN general the amount of FST and QST is possitively correlated in population pairs. This is bc the greater time for divergence the greater the value fo ST regardless of whether it is FST or QST. SO the correlation here represents population pairs that have diverged for a long time (high ST) vs s short time (low ST)

Additive Genetic Variance

Genetic variance associated with the average effects of substituting one allele for another

Mendel

German monk studied plants to determine God's laws. Based on manipulative experiments with peas he tested inheritance. Mendelian inheritance involves traits that tend to reflect discrete differences between phenotypes: wrinkly vs smooth peas; dark vs. light moth colors.

Growth rate in two environments

Growth rate in different environments may have a different genetic basis. Growth rate at low food: metabolic efficiency, aggressiveness if interference competition is important. Growth rate at high food: metabolic efficiency, appetite. We would expect a positive genetic correlation between these two traits but it is unlikely to be 1, likely <1 as many loci will be different for two trait.s

Mollie Male Tail Sword Length

Has the male's sword evolved under selection imposed by female choice? Two hypotheses must be true 1. Preference is based on male sword length itself not some other characteristic of males that is related to sword length. Therefore female preference will be stronger for males with increasingly longer tails 2. Females prefer males with longer tail swords. Regardless of absolute length of the tail sword, females will prefer males with the longer tail. Evaluated by artificially increasing or decreasing the sword length in a sample of males and measuring female preference. Manipulated by cutting the tail to desired length by selecting random males from a population. The random female was given the choice between two manipulated males that differed in their tail swords and scored by how much time she spent associating w each. The effect of sword length differences were artificially created in some tests reversing size of male tails so female behavior was based on tail length alone. The female preference hypothesis was supported. but we still don't know: 1. does increased preference for long tailed males lead to increased reproductive success for males w longer tails 2. Is female preference the only source of selection on tail sword length

Horned Lizards-> exaptation or adaptation?

Have a variety of pointed horns along the back of its head. They have a predator, a small bird (shrike) that can swoop down on small lizards and kill them by a bite to the neck, after which they skewer the carcass on a sharp branch in their territory in order to hold it while they feed. Living lizards have larger horns than shrike-killed lizards. It can only be proven if horns originated at the same time that shrikes became a major predator, aka it evolved as anti-predator weapons against shrikes specifically. If they didn't, then horns are an "exaptation"-> they evolved for some other reason and now have come under selection for anti-shrike predator performance. Similar logic: if feathers and flight evolved at once then feathers are an adaptation for selection imposed by flight. If feathers are an exaptation they evolved at a different time than flight, perhaps they first evolved for warmth, but then flight evolved and selection from gliding and flight starting to influence evolutionary change in feathers. This is the current utility definition or adaptation, rather than the original utility definition that phylogeneticists often use.

What limits selection

If there is not much variation among individuals in their relative fitness then there is not much opportunity for selection. As the variance in fitness gets smaller this causes the slope of the line to tend to 0. Greater variation in fitness means a greater opportunity for selection, as long as there is a consistent relationship between fitness and phenotype so the variance in fitness among individuals sets an upper limit on the strength of selection operating on the population.

Evolution of Plasticity

In order for evolution of plasticity we need phenotypic variation in plastic responses and we need this variation to cause fitness differences in the pop'n, and there needs to be genetic variation underlying phenotypic variation. Genetic variation in reaction norms are needed (GxE which is when reaction norms cross). Then we need selection on the reaction norms, from fitness differences (steeper slopes, where the norm changes in a functionally appropriate direction) Often plasticity evolves when individuals experience changes in their environment that they cannot avoid. Loss of plasticity due to costs: 1. If plasticity is costly this makes it difficult to evolve. Maintaining machinery oc plasticity may be costly, ability of an organism to perceive and evaluate an environmental change and decide to respond or not. 2. mismatch cost. If phenotype is mismatched w the environment it finds itself in. Maybe the environment varies too quickly for an organism to respond and catch up. Therefore plasticity is an evolvable trait as long as there is heritable variation in plasticity. If there is little genetic variation in plasticity (low GxE) then even under very strong selection plasticity may not evolve.

Positive Maternal Effect

Increasing resemblance between offspring and parents, amplifies the change in mean phenotype from generation to generation.

Polygenic trait

Influence by many genetic loci. Interaction between alleles (epistasis). Interaction with environment (phenotypic plasticity). Continuous traits

Quantitative Traits

Influenced by many loci each with small effects. Experience both genetic and environmental variation. Continuously varying traits in a population.

Spandrels

Inside a church are many arches that hold stuff up, arches and pillars have functions. Spandrels are triangular features between arches so they are a trait of everything building with arches, but not functional, but pairs of arches simple cannot be created with spandrels between them. Arches and pillars are functional traits but spandrels are a result of the constraints of architecture. The spandrel is a metaphor for traits in organisms that did not evolve directly under natural selection for some performance function. Many traits that start out as byproducts of other traits may by chance later be useful to an organism

Non-linear Selection Differentials (C)

Instead of directional selection (linear selection) we have stabilizing selection or disruptive selection where the intermediates or extremes have the highest fitness. Again, done in two ways: 1. Estimate C from a fitness function (relative fitness (w) vs. phenotype (z)). From standardized values of phenotype (z) and fitness (w). w=B + Sz1 + CZ2 where S=linear selection component C= non-linear selection component 2. Compare variance in phenotype between before and after selection took place samples using frequencies of standardized values of phenotype (z). C= Variance(after)- variance(before) + S^2 When variance(after)>variance(before) C is positive and selection is disruptive on the trait When variance(before)>variance(after) C is negative and selection is stabilizing The S^2 accounts for the fact that linear selection also somewhat reduces variance in the population negative, non-linear selection differential means there was more variation before than after selection occurred

Conclusions on Guppy Adaptation

Introduction experiment more powerful way to test hypotheses about natural selection than correlation approach (better causation). Experimental response to selection had to be assessed in a common garden experiment in the laboratory, and since differences persist for two generations it is genetic in basis and not in response to detection of a predator.

Equivalent fitness of different phenotypes doesn't necessarily imply disruptive selection

It is tempting to interpret sign of disruptive selection as always indicating that two forms are equally fit at the same time. Not always the case, for example in a pop'n of fish that attack by sneaking up on the side of their larger prey mouths are either bent in the left direction or in the right. In populations where most individuals attack from the right it is more advantageous to have a right-sided mouth and attack from the left because prey are more watchful of their right side. Same goes for a population where individuals mainly attack from the left. So directional selection switches direction back and forth favouring one and then the other and frequencies change.

Non-linear Selection Does not always have to be stabilizing

It is tempting to interpret the signs estimated for C or y as clear indications of stabilizing (-C) or disruptive (+C) selection, but that is not always true. E.g. pollen production by wild radish flower enhances fertilization up to a certain value of pollen production by a plant, but above that there is no added benefit to producing more pollen. Kind of non-linear directional selection on pollen production in flowers by pollinators.

Above falls (guppies)

Killifish, k-selecting (fewer offspring with greater parental care), population size near K, intense competition, large adult size, long generation time (older age at maturity), few and large offspring

Two ways of performance analysis data collection

Look at natural variation in the population. Or these traits can be manipulated in an experimental study.

Phenotypic Plasticity

Many traits are flexible, meaning the that the individual can vary during its lifetime depending on some environmental cue, at the population level, Ve. Traits can be classified by how flexible they are. Behavioural and life history traits are very flexible, but morphological traits are only rarely flexible. We can quantify the flexibility of the phenotype w a phenotypic reaction norm which shows the trait values produced under different environmental conditions. Reaction norms can be linear, curvilinear and even discontinuous. Reaction norms can be measured on individuals, genotypes, families or members of whole populations. In order to have systems that can have phenotypic plastic responses they need to be able to receive an external cue from the local environment. If reaction norms show more than one line then there is genetic variation in a phenotypic trait. If individuals have a plastic response then the line must have a slope (cannot be horizontal line, bc then there is no difference in different environments). If all lines are parallel there is no genetic variation in plastic responses by individuals in the population. (low G*E) A negative plastic phenotypic response will include crossing reaction norms. If there is more variation (higher variance) in one environment when compared to another then environmental effects will affect the heritability of the phenotypic trait.

Polygenic trait inheritance (quantitative traits)

Medeldian and quantitative traits are now thought to be different versions of the same genetic determinism model that connects genotype to phenotype, they only really differ by the number of loci that influence the phenotype value. The simplest mendelian trait has 1 locus w two alleles affecting the trait. As we add more loci that affect a trait phenotype we create more intermediate forms between the two homozygous parental forms resulting in phenotypic distributions that are increasing continuous. Continuous traits are referred to as polygenic w polygenic inheritance. Quantitative genetics examines the genetics of continuously varying quantitative traits

Natural evolutionary rates

Most evolutionary rate estimates are very low. There is a negative relationship between the rate estimate by study and size of interval of time over which the rate was estimated, so estimates from long intervals of time tended to be slow and estimates made from short intervals tend to be fast (maybe to do alternating selection pressures)

How variable is natural selection within a system

Most studies only measure selection in one year. IN studies over many years the direction of selection on a trait often changes between successive years, especially for studies of survival but also for mating. Studies of fecundity show changes in direction from year to year, so selection may be inconsistent in nature and even a reversible force on trait evolution rather than a consstant force.

Selection Gradient Analysis

Multiple linear regression. w=β0 + β1z1 + β2z2 One trait is z1 and one is z2. Ideally, we want to know how each trait independently affects relative fitness (w). Multiple regression evaluates the direct effect of z1 on fitness (β1) while accounting for the direct effect of z2 on fitness (β2). These βs evaluate the direct effect of the trait on fitness while accounting for the effects of other traits. If a β is significant the trait is under direct selection. If only S is significant it is only indirectly selected for. Two Assumptions: 1. You still do not have direct evidence that any of these traits functionally affect survival under shrike predation 2. The outcome may depend on what other traits you consider in your model. There could always be some other unknown trait that you haven't considered that could be correlated with your trait of concern. If this unknown trait functional influences fitness then the selection gradient will change.

Selection Gradient

One way to appreciate that since organisms are composed of many possible traits that may be correlated with each other, we should estimate selection on trait A while statistically accounting for selection on other correlated traits. Other traits may also functionally affect survival, and some/all of the traits of an organism may also be correlated with each other (e.g. lizards with larger body size also have larger horns) THis accounts for the selection differential S=direct selection + indirect selection and only shows the direct selection.

How Strong is Natural Selection

Only a few examples of direct selection being strong (β>0.5). Selection on most traits in natural populations is relatively weak. Weak selection is harder to statistically detect than is strong selection. This suggests that in nature most traits are under relatively weak selection or are only rarely under strong selection. So there may be two ways that selection shapes variation in natural populations: continuous weak selection or rare bouts of very strong selection. Selection data: There is good coverage across broad taxonomic groups but there are biases 1. Studies of selection on morphology far outweigh any other kinds of traits 2. Fitness is most often estimated by mating success. Complete estimates of total fitness including reproductive success, fecundity and survival were rare. So the previous conclusions assume that these issues do not bias our inferences

Estimating Heritability

Parent-offspring regression (slope of best fit estimates h^2), sibling analysis, twin/clone studies, predigree analysis.

Maternal Effects

Parents (mothers usually) can have a large influence on the early environment experienced by their offspring, defined as maternal contributions to offspring phenotypes beyond their direct genetic contribution. Can be done by providing good or poor food, shelter, places to grow and even important information about local conditions. This is a very special cross-generational form of phenotypic plasticity. Parental contributions can affect offspring fitness and offspring phenotype. Up to now we have assumed that resemblance of offspring to parents is entirely genetic due to inheritance, but this is not valid in species where there is any kind of parental care. if we can increase resemblance between offspring and parents we can over-inflate heritability estimates for some traits in a population (e.g. larger beetle parents make a larger dung pile for their offspring so their offspring can be larger) Maternal effects can be post birth or pre birth (how female provisions the eggs with chemicals etc)

Performance Measure of Tree Bark Colour

Performance measure is the temperature of the cambium in the water. North vs. south side of the tree acts as a control to see if the paint itself affects the cambium temperature because the north side does not get any sun in the day (northern latitudes). All naturally white trees and painted trees increase and then decrease temp over the course of a single winter day and didn't thaw during the day. But paint a white tree brown and its cambium temp can increase and decrease through a range that goes above 0 (thaws) if it is on the south side of the tree. Lack of freeze-thaw on the N side suggests that light absorption (brown's trait) causes warming of the cambium tissue and not changes in air temperature. Freeze-thawing causes frost damage to cell membranes and soft tissues. Cambium damage from freeze-thaw was worse on brown-painted compared to the controls. We still don't know id cambium colour influences fitness or whether variation in bark colour is heritable so we don't know if bark colour is an adaptation to daily light cycles in the winter

Environmental Variance

Phenotypic variation caused by the environment. P = G + E

Below falls (guppies)

Pike cichilid, r-selecting (lots of offspring, little parental care), population size well below K, benign competition, small adult size, short generation time (younger age of maturity), many and small offspring (so that a few would manage to avoid being eaten by pike cichlid)

Response to Selection

R. The response to selection is a change in mean over multiple generations. Selection differential is measured in one generation. Response to selection is mean of gen2-mean of gen1, or it is S(h^2), the selection differential multiplied by heritability. Focusing on microevolutionary responses.

Life History Trade-Offs and Guppies

Reznick says you should think of life history evolution in terms of energy allocation to different body processes. Juveniles would be expected to devote most resources to growth in low predation pools where killifish are the only predators. Also some energy to body maintenance/repair as at low adult risk sites bc they may live to an old age. In low adult risk pools adult females devote energy to maintenance and fat accumulation that is not observed in high adult risk of predation pools (doesn't matter here, must devote lots of resources to reproduction).

Reaction Norms

Set of phenotypes produced by a genotype over a range of environments. Invented by plant ecologists that grew exactly the same genotype in each of several environments. In this case the weight of guppies that are good foragers in low vs. high food environments (do better in high food), or the weight of guppies that are good competitors in high vs. low food (do better in low food). Often reaction norms are represented as graphs of a phenotype across environments. If reaction norms do not cross genetic correlation is positive and it is easier to select both traits in the same direction. Here you can see if the environmental condition 1 and 2 have different effects on phenotypes (if the lines are not completely parallel then the effect of environment on phenotype is unequal from one genotype to another) If the optimum is the same line in the graph you can reach joint optimum w an existing genotype which will become fixed. If it is not the same line they can not reach joint optimum with existing genotypes unless mutation. Might become specialist restricted to one environment. Reaction norms cross: negative genetic correlation between performance in each environment (G*E interaction if reaction norms cross) (z1 and z2 will have a negative slope) Reaction norms can evolve if G*E is present. If there is no genetic correlation between traits z1 and z2 there will be no consistent relationship between performance in two environments. G*E interaction so can fix best genotype so that optimal phenotype is produced in each environment (if the optimums are correlated). Any future evolution trajectory is independent as different sets of loci are involved. Evolution of phenotypic plasticity can occur.

Guppie Coloration

Small fish that live in pools. Pools below waterfalls have many species that prey on guppies including the pike cichlid (main predator). Species has sexual dimorphism, males are smaller and brightly colored. Female behavior in high-predation sites is constrained by predator avoidance (they avoid more than males, and have higher foraging rates) and sexual harassment from males trying to mate before they are eaten. Pools above waterfalls have few predators, only killifish which prey on juveniles. Males with pretty colors attract more females and therefore have higher mating success. Colorful low-predation males devote time to pursuing females and attempting sneaker matings, so females experience unsolicited mating attempts that prevent their foraging. Endler found positive directional selection for bright colors in males at killifish sites, as it resulted in higher mating success. Found negative directional selection for dull colors where there were pike cichlids. His transplant experiments showed male guppies transplanted above waterfalls evolved brighter coloration. Killifish can travel overland on rainy nights and get above waterfalls, but only prey on juveniles less than 12mm. More competition for food when Killifish, not pike cichlid are present bc the pike cichlid keeps pop'n density low. The Pike cichlid prefers brightly-colored active fish of all size classes. Bc guppy populations are reduced by the pike cichlid food is abundant and growth is fast

Quantitative genetics

Study of the genetic mechanisms of continuous phenotypic traits. A bell curve can be approximated by only two genetic loci each with two frequent alleles.

Immortal Organisms

THere are examples of organisms that are immortal, a planarian worm can be cut repeatedly in half and if in optimal environmental conditions and nutrients can regenerate itself from each half into two worms. A jellyfish releases gametes that when fertilized form planula larvae that settle to the bottom and develop into a sedentary polyp stage that grows and eventually buds off a new free-floating medusa stage. But when environmental conditions are poor the medusa avoids reproduction by reversing its aging and becoming a young sedentary polyp again. It can then bud off a medusa stage. Technically this means that it could never die, some piece of itself will always be out there.

Agent of Selection

The ecological/environmental factor that causes selection on the trait in the population (e.g. for horned lizards the shrikes feeding on horned lizards are the agent). We hypothesize that variation in the agent of selection should differentially affect survival and reproduction. To test the agent of selection vary the agent, either observationally (the agent already varies in abundance in space or time) or experimentally (manipulate the abundance of the agent). Then ask: -Does this affect survival or reproductive success of individuals in the population? (could this cause selection) -Does the intensity of selection change (does it cause selection) -Does the response to selection change? (does it cause selection)? It is not enough to see if the variation in the source affects survival and reproduction, because that is just saying that is could cause selection. What we need to see in order to say there is evidence that it causes selection, is the variation in intensity of selection and response to selection for the trait of interest to us.

Disposable Soma Hypothesis

The evolution of a separate germline from somatic cells has resulted in immortal germ line cells but at the expense of somatic death. Genes are close to immortal and repeatedly develop a disposable vehicle to travel in (and find other genes for reproduction) then to continuously maintain the vehicle for ever. Bodies are disposable vehicles for the genes in your germ line. This hypothesis proposes a fundamental energetic trade off: An individual can only acquire a finite amount of energy and nutrients in its life time. It can allocate some of this to maintain the soma and the rest for reproduction. This is done so in a trade off since there is a finite amount of energy and nutrients, the more energy allocated to the soma the less is available for reproduction (and vice versa). Different genotypes make different allocations but some of these are better than others. The best allocation to each of these is one that will maximize life time reproductive success. (if mates are rare you might want to invest in the soma so you can look for them, if mates are abundant you might want to invest in reproduction instead). The hypothesis is that it is energetically cheaper to occasionally recreate a new soma than to continuously repair an old soma forever, and this have generated selection to favour the evolution of a separate germ line cells and soma cells. If it is adaptive to occasionally regrow a new soma and not continuously invest in somatic maintenance, then aging results from un-repaired damage to the cellular machinery that makes the soma possible. The evolution of the separation of germline and somatic cells is somatic death and immortal germ line cells. Senescence is not a target of selection, it evolves as a by-product of selection acting on energy allocation between germline and soma.

Correlational Selection

The form of selection that results in two traits changing in a functional way with each other. Traits do not independently or additively affect fitness, particular *combinations* of traits are favoured over other combinations. We tend to focus on the functional consequences and evolution of one trait at a time, independent of the effects of the others, however the correlation of two or more traits together could be adaptive and have evolved directly under selection. E.g. striped snakes tend to flee from a predator w a fast sprint while spotted species tend to flee w short sprint and freeze (reversal). The two traits are correlated. Researchers collected a bunch of garter snakes from the wild and crossed them to make a number of different combinations of skin pattern and escape behavior and then did mark-recapture. Assumed that frequency of recaptures was related to survival. Lowest recapture frequency: striped and reversal. Two best combinations: spotted and reversal and striped and sprint. This suggests that selection by predators may have caused the evolution of trait covariation among the different snake species, so selection could be the mechanism that caused different species to evolve with different trait combinations.

Measuring Correlational Selection

The multiple regression model that we have used before to study selection can be extended to include a test for correlational selection, done by including an interaction term (correlational selection coefficient Ybc) that estimates how much the effect of one trait (c) on fitness depends on the phenotype of the second trait (b). w= βo + β1(b) + β2(c) +Ybc or can include non-linear selection gradients as well w=βo + β1(b) + β2(c) + Y1(b)^2 +Y2(c)^2 + Ybc If there are no significant non-linear selection gradients then selection is directional.

Infinitesimal Adaptive Evolution

The original model of adaptive evolution. Phenotypic traits are made up of many genes, each with a small effect. All under relatively weak but continuous selection so that not too much genetic variation is lost due to selection. Whatever variation is lost is replaced by new mutations. Brought about bc if many traits evolve under selection which favours some genotypes over the others in a population, then what maintain genetic variation in a population over long time periods? Selection should reduce genetic variation over time, so the solution was to think of selection as weak for the most part and all genetic variation that is lost is replaced due to mutation.

Adaptation- Noun

The outcome of adaptation, the verb, resulting in all individuals in a population having some sort of adapative feature for their particular local environmental condition. Feature of populations. Criteria of a trait being an adaptation: 1. Evolving/recently evolved under natural selection 2. Has a function that influences performance at some ecological task 3. Performance affects fitness 4. Heritable variation exists for the trait in the population 5. Current vs. original function and selection (some biologists wouldn't include a spandrel as an adaptation bc it didn't evolve as an adaptation originally, others think that if the trait was modified by selection it's an adaptation, regardless of original function.

Adaptation-Verb

The process of adaptation where natural selection causes genetic change over generations in a population in some trait that functionally influences the fitness of populations. Occurs in populations.

Inheritance

The process that passes DNA from a parent to an offspring, property of individuals across generations.

Heritability

The proportion of phenotypic variation in a quantitative trait that is inherited. 0< h^2<1. Selection only leads to evolution if the traits are heritable. If variation is due to differences in genotype then the survivors of selection pass their successful phenotypes to their offspring. A population statistic about the average correspondence between parents and offspring in your population (must be measured in the same environment). It is a population statistic. The hypothesis of heritability predicts that on average the phenotypes of offspring should resemble the average phenotype of their parents. *Measures the relative importance of genetic variation in determining phenotypic variation in a population*. Heritability will be 0 if all individuals have the same phenotype, bc it relies on a relationship between different parent phenotypes and different offspring phenotypes, if there if no variation there is no heritability, even though there is genetic effects. If genes that control a phenotype are fixed in a population then heritability is 0 even though genes control the phenotype.

Genetic determinism

The relationship between an individual genotype and its phenotype, through development.

Narrow sense heritability

The slope of offspring trait vs. mid-parent trait gives you the narrow sense heritability. The proportion of phenotypic variance in the trait that is directly heritable. h^2 = Va/Vp Additive genetic variance/phenotypic variance. Assumed narrow sense in the breeder's equation. The proportion of phenotpic variance explained by additive genetic variation

Common Garden Experiment

This is done to determine the heritability of genes, or the genetic component to genes. Reznick randomly sampled wild guppies from 2 high and 2 low predation sites. Offspring from each wild females reared in a constant environment in the lab. Second generation offspring are mated to lab offspring from the first generation litters of other mothers (from the same intensity of predation environment). He found that those reared from low predation sites were older at maturity and larger at maturity and had fewer offspring than those reared from high predation sites. The first litter showed fewer offspring in low compared to high but the second and third litter did not show this, low and high had the same number of offspring in their litter because they were reared in a common environment. Common environment experiments are helpful so one can tell if a genetic response has occurred. Vp= Vg +Ve If we eliminate Ve by raising everything in the same environment the differences among individuals only reflect genetic variation Vp=Vg, can be used to estimate heritability.

Studying Adaptation

To study adaptation you need to understand mechanistically how the trait affects performance, and how performance affects fitness in a specific ecological context. 1. What is the source of selection? Agent of selection? The ecological or environmental factor that affects some individuals more than others and so could cause natural selection. Can be tested by comparing areas with the source to areas without, or experimentally removing the source. 2. What is the target of selection? The trait that is specifically responsible for fitness differences? We must know: -How does variation in the trait affect performance of the individual at some task in a specific ecological context? -How does variation in performance affect survival or reproductive success (fitness) in the specific ecological context?

Trait Variance

Variance is a measure of spread (differences among individuals in some feature) in a population, concept invented by Fisher (F-test). Variances are additive (include and can be split into genetic and environmental components) Vg= Va +Vd +Vi Vd(dominace) and Vi(epistasis) are often lumped together because they are not simple inherited from parents to offspring like single alleles, they are interactions between sets of alleles that are inherited from both parents. Ve = VE + Vm VE is environment of individuals, broken down into differences within individuals (differences in repeated measures in the same environment due to chance) and differences among individuals (differences in same environment, due to plasticity) Vm= mom's non-genetic influence on her offspring (maternal care) e.g. moms reared in nutrient rich environments will produce offspring w different traits than those reared in low nutrient environments. Mom's environmental effect is passed on to all of her offspring.

Genetic Variance (Vg)

Vg = Va + Vd + Vi Genetic variance is made up of additive (heritability genetic variance), dominance variance and interaction (epistasis) variance. Epistasis= genetic value of an allele at one locus depends on which allele is at another locus. Dominance= genetic value of an allele at one locus depends on the other allele at the locus. Bc we cannot inherit Vd or Vi from a single parent (depends on the combination of parents) we usually only care about narrow sense heritability.

Can we accurately estimate Va from a regression of mid-parent mid-offspring?

We can only get an estimate of additive genetic variance of a trait by removing parental effects, how do we do this? IN organisms where fathers contribute only sperm and nothing else we can perform a father-offspring regression of phenotypes because the father contributes only sperm. Comparing the F-O regression tot he equivalent mother-offspring regression can reveal maternal effect contributions on offspring phenotype. The differences in slopes reveals the amount of maternal effect that contributes to the resemblance between parents and offspring. Only works if fathers do not contribute parental care. When fathers do contribute more than genes we use an experiment: Cross fostering. This is when eggs are switched among nests as soon after laying as possible so that some/all are reared by animals that are not the genetic parents. We can then measure the degree of non-genetic parental effect by the foster parent. We thing regress offspring phenotype against the foster and genetic parent phenotype. The slope of the offspring against genetic parents gets an estimate of heritability without parental care. The slope of the offspring against foster parent is estimate of parental care effects (expected to be a much weaker effect compared to the heritable effect). This only controls for post-birth maternal effects. Common Garden experiments rarely removes all maternal effects (bc of the pre-germination/hatching effects). But the second generation of offspring reared in the common garden environment will tend to express phenotypes with few maternal effects. A common garden 1st generation parent 2nd generation offspring regression will yield an estimate of heritability that is minimally affected by maternal effects. Although maternal effects can span 3-4 generations!

Plastic Response or Evolution?

We have to try to test whether phenotypic differences reflect plastic responses to an environmental change. ONe way is to take offspring of different phenotypes and rear them in a common environment (for 2+ generations bc of maternal effects). We can also look for direct evidence of genetic change in the population for those traits likely under selection 1. when you have a pop'n not under presumed selective pressure and can compare it to a pop'n under selection where the phenotypic change as occurred 2. When we can also track a population over generations - compare frequencies in past to current (harder, 1 is more common)

Reaction Norm for Maternal Effect

Where the y axis is offspring trait and the x axis is the maternal environment or condition (say high or low nutrient condition). The single genotype may be offspring weight. Without maternal effect the horizontal line would not have a slope and with maternal effect it would result in higher offspring weight in high nutrient condition compared to low nutrient condition experienced by the mother. If these genotypes occurred among individuals in a single pop'n it would indicate genetic variation in maternal effects, this means that under some conditions maternal effects may evolve under selection.

Haldanes

a measure of the change in mean value of a trait in a population per generation, made in standard deviation units and applicable to multigenerational microevolution. These allow us to compare differences in rates of trait evolution for organisms with vastly different generation times. Meausre the mean trait value, amount of phenotypic variation in your target pop'n at generation t, then you repeat measurements in same pop'n a known number of generations later. The difference in mean trait values divided by the number of generations provides a per generation change in the trait in the population in standard deviation units.

Lande's Equation for Multiple Traits

deltaz= G(P^-1)S or deltaz=Gβ Where deltaz is a vector of the response to selection, G is the genetic variance matrix, P^-1 is the inverse of phenotypic variance matrix, and S is a vector of the selection differentials, β is a vector of multivariate linear selection gradient. Deltaz is the change in beak depth/width in generation which is equal to genetic variance times β. Gives you a predicted response to selection for multiple traits. We will most likely not use this equation. rp= indirect selection rg= correlated response from shared genes

Size of h^2

h^2=1: The change will be passed on entirely to the next generation. H^2 is almost never 1, usually less, so generation 2 will usually have a mean that is different from individuals with highest fitness in preceding generation. Morphological traits have higher h^2 than life history or behavioral traits. Therefore the response to selection should be greater for morphological traits.

Standardized Selection Differential

i= S /sqr(Vp) where Vp is phenotypic sample variance. S is the selection differential (meanafter-meanbefore). Standardizing allows us to compare selection on different traits. This allows you to ask whether natural selection was stronger on trait X or trait Y or trait Z.

Fitness, traits, and selection

qualities of individuals, in theory they can be measured at the individual level. Both fitness and trait values are continuous (quantitative) Selection on the other hand is an ecological process that acts on a population, so selection is a quality of a population. Selection is a statistical relationship between variation in fitness and variation in some trait among individuals in a single populations. Just because there is variation in fitness among individuals does not imply that selection will occur in the population