Introduction to Evolution

Let's begin our discussion of evolution by trying to follow some of the reasoning that led Darwin to the formulation of his theory:

1. All living things have a tendency to over-reproduce. If an animal population is to remain the same size, every pair of parents must on average produce two surviving offspring during their lifetimes. If they produce more than two, the population will increase, and because each offspring can itself become a parent, the increase in population has the potential to become explosive (exponential). If on average parents in a population produce fewer than two surviving offspring, the population will decline. If that trend continues long enough, it will become extinct. In fact, animals tend to produce more than two offspring. A codfish produces a million or more eggs at a single spawning. A robin lays four eggs at a time; and a robin may produce several clutches of four eggs each during its lifetime. Unless something checked this general tendency to over-reproduction, the world would eventually become impossibly overcrowded.

2. Nevertheless, despite each species' capacity for over-reproduction, populations of most kinds of animals tend to remain more or less stable from one generation to the next.

3. Within species, individuals differ. To some extent, these differences are passed on to descendants via inheritance, i.e. some differences among individuals within species are genetically determined in whole or in part.

4. Given the relative stability of populations in the face of the capacity for profligate over-reproduction, it is apparent that some individuals are more successful than others in producing offspring and/or some offspring are more successful than others in becoming adults. In other words, competition occurs. "Survival of the fittest" in Darwinian terms simply means that those individuals possessing characteristics which make them more capable of surviving and reproducing will do so and will be more successful; that is, they will be better represented in the next generation than will those individuals who are less fit. This is natural selection: the differential reproduction of individuals within a species, from one generation to the next. Natural selection follows irresistibly from the facts of (1) each species' potential for over-reproduction, (2) the absence of such over-reproduction in the long term, (3) the existence of inheritable differences among individuals, and (4) the resulting competition among individuals which causes some individuals to be reproductively more successful than others.

5. Finally, under certain conditions, the process of natural selection (operating by the differential representation of each individual's offspring in succeeding generations) has the capacity gradually (across many generations) to produce changes in the average characteristics of a species. Such change is what is meant by evolution.

Note that natural selection can lead to evolutionary change only if the differences among individuals reflect some degree of underlying genetic difference. If all individuals in a population of animals were genetically identical, it would not matter which ones were producing more offspring: the characteristics of the next (offspring) generation would be the same as those of the previous (parental) generation.

Microevolution

Sexual reproduction produces new combinations of genes each generation, but by itself it does not cause genes to change in frequency. It does not cause evolution.

Let's imagine that a particular gene has two alleles (or forms), a1 and a2 , that occur with the frequencies p and q respectively. If a1 and a2 are the only two alleles of this gene in the population, then by definition p + q = 1. Let us further imagine that sexual breeding occurs at random among individuals possessing these alleles. The frequencies of diploid offspring following such mating can be expressed as:

Where: will be the frequency of offspring with the genotype a1 a1 (a1 homozygotes), 2pq is the frequency of individuals with the a1 a2 genotype (heterozygotes), and is the frequency of a2 a2 homozygotes. The equation is called the Hardy-Weinberg Law. It can be derived intuitively as follows. If breeding is random, the probability of getting an a1 a1 individual is equal to the probability of an a1 sperm finding an a1 ovum. This will be equal to . Similarly the probability of getting an a2 a2 individual is . The probability of getting an a1 a2 individual is, however, equal to the probability of an a1 sperm finding an a2 ovum (pq) plus the probability of an a2 sperm finding an a1 ovum (qp). Since qp = pq, the probability of getting an a 1 a 2 individual is 2pq.

This result will hold generation after generation. Thus, although sexual reproduction allows individuals to produce offspring with a diversity of genotypes, it does not alter the frequency of genes; it does not cause evolution.

We will now consider 3 of the most important ways that evolution may occur while omitting consideration of several other less important factors.

1. Genetic Drift, or the alteration of gene frequencies by chance. We have said that the alleles a1 and a2 have the frequencies p and q in a parental generation, and we have shown why on average the offspring of such a generation should also have the frequencies p and q for the two alleles. However, the process of random mating is by definition a random process; and sampling error can occur.

For the moment, let's consider an analogy. Let 's imagine that a population's gene pool for the alleles a1 and a2 consists of a large bag containing black and white marbles with the proportions of black (a1 ) marbles being p and the proportion of white (a2 ) marbles being q. Moreover, for the moment, let's imagine that p = q = 0.5; in other words that the number of black marbles is exactly equal to the number of white marbles. We will now draw marbles from the bag at random. Let's imagine that we draw out samples of 10 marbles, an operation equivalent to that involved in 5 random matings to produce 5 diploid organisms. The probability of getting various combinations of black and white balls can be figured as follows.

Let:

N = the total number of marbles drawn

r = the number of white marbles drawn

p r = the probability of drawing exactly r white marbles out of a total of N.

Elementary probability theory tells us that:

With N = 10 and p = q = 1/2, this equation becomes:

Using this formula, the probabilities for various r's are as follows:

r 0 = r 10 = .00098

r 1 = r 9 = .00977

r 2 = r 8 = .04395

r 3 = r 7 = .11719

r 4 = r 6 = .20508

r 5 = .24609

Note two things about this demonstration:

1. With this small sample size, the probability that the number of black marbles will be exactly equal to the number of white marbles is only about .25.

2. There is a small, but finite chance, that no black or no white marbles will be drawn. This is analogous to the extinction of one of the alleles. With these small samples, the probability that one or the other allele would go extinct by chance is about equal to .002 (or about .00098 x 2).

This example is analogous to events that might occur in a small population or deme of sexually reproducing organisms. In such a population, 2N gametes are drawn from the population at random to produce the next generation, where N is the number of offspring comprising the next generation. If 2N is small enough, the proportions p and q of the alleles a1 and a2 can fluctuate considerably from generation to generation.

This kind of fluctuation is called genetic drift; and there are three circumstances under which it might be an important evolutionary factor. In continuous drift, a small population might remain small for many generations. Under these circumstances, sampling error would be effective in each generation. In intermittent drift, the population is reduced to a small size only occasionally. If mortality is random at the time of the reduction (something which is likely if many individuals are killed by some essentially random event like an earthquake or volcanic eruption), the survivors may by chance constitute a relatively nonrepresentative sample of the population. The third factor is called the founder effect. New demes are occasionally started by a small number of individuals that have accidentally found their way into some region which is suitable for their species. These founders may by chance constitute an unusual genetic sample of their species. When this happens, new demes will tend to be different both from the parent deme and from one another. Many population biologists consider that this founder effect or founder principle may be an important factor in the origin of new species.

Without going into the possible quantitative estimates, it should be intuitively obvious to you that drift will be important only in small populations. The chance that alleles will become either lost or fixed in a relatively small number of generations is considerable in populations having an effective size of less than about 100 members.

Another apparent point that should be stressed is that evolution by genetic drift is without direction. All possible changes are equally likely. It is very unlikely that any substantial net change in any particular direction will occur.

2. Gene Flow. A very rapid way in which the genetic composition of a deme can be changed is by matings with genetically different immigrants from another deme. Again this should be intuitively obvious. Gene flow between demes living in somewhat different environments constitutes an inertial factor which slows the evolutionary divergence of behavioural and other adaptations. However, it also tends to keep the total genetic diversity of the species high and to heighten diversity within demes.

3. Selection, the evolutionary principle postulated by Darwin, is undoubtedly the most important evolutionary force. It is the only evolutionary principle that can assemble and hold together groups of genes over long periods of time.

Selection can be defined as changes over successive generations in the relative frequency of certain alleles as a result of differences in the reproductive success of the phenotypes which they produce. Differences in reproductive success can be due to a variety of factors including:

a. differences in the capacity to survive competition with animals possessing other genotypes.

b. differences in the capacity to survive the onslaught of predators, parasites and diseases.

c. different capacities to survive long-term changes in climate and other changes in the physical environment.

d. variable reproductive competence including differential tendencies to mate, differential fecundity, and differential viability of offspring.

e. differential tendencies to disperse and differential probabilities of survival in new environments.

Selection can act upon the genetic variation in a population in at least three ways. These ways can most easily be represented graphically. In the diagrams which follow, an up arrow means that a range of variability is favoured by selection. A down arrow means a range of variation selected against.

(i) Stabilising selection acts to favour the average individual while selecting against deviants of all kinds:

This is also sometimes called optimising selection. The result is a reduction in variability. To some extent, this process is almost always operative.

(ii) In diversifying or disruptive selection, extreme individuals of both (or all) types are favoured over the average:

This is caused by two or more adaptive modes located some distance apart along a scale of phenotypic variation. If combined with preferential mating among individuals with similar genotypes, this kind of selection might lead to the formation of two new species.

(iii) In directional selection, individuals possessing one kind of extreme genotype are favoured:

This is the only form of selection that has the capacity to "move" a population in a particular evolutionary direction.

Behavioural Evolution

Inertia and Preadaptation. In our discussion of so-called multiplier effects, we said that behavioural traits and particularly the detailed characteristics of complex social organisations can be expected to change rather rapidly in response to long-term changes in the environment. However, given that a species finds itself in a situation where a particular form of behaviour would be advantageous, it is by no means certain that the potentially adaptive behaviour will evolve. Given that several different species find themselves in a similar ecological situation in which several alternative behaviour might be advantageous, one species may fail to evolve at all and become extinct.

Considerations of this kind point to the necessity of considering evolutionary inertia (factors which make certain kinds of evolution unlikely) and preadaptation (fortuitous predispositions that facilitate evolution of some characteristic).

One of the most striking examples of differences in evolutionary inertia producing evolutionary diversity can be found in the evolution of complex social behaviour in insects. Complex colonial societies in which sterile workers assist reproductive castes in rearing their young have evolved at least 12 times among insects: once among the primitive cockroaches to produce termites, and at the very least 11 times in the wasp order Hymenoptera to produce the ants, hornets, several other kinds of social wasps and a number of kinds of social bees. Equally complex forms of colonial societies based on physiological, reproductive, and especially behavioural division of labour have never evolved in any of the other 20 some odd insect orders. For example, there are 600,000 known species of beetles (Coleoptera) and only about 250,000 species of Hymenoptera. Yet, complex sociality (eusociality) has evolved at least 11 times among the Hymenoptera and never among the beetles. Why 11 times among the Hymenoptera? According to many sociobiologists, the likeliest answer is to be found in the hymenopteran genetic system. Hymenoptera possess a genetic system called haplodiploidy in which males are haploid (develop from unfertilized eggs and have N chromosomes) and females are diploid (develop from fertilized eggs and have 2N chromosomes). As a result of this system, full sisters are more closely related to one another than they would be to their own offspring. For this reason, it can be more advantageous for a group of sisters to assist their mother in rearing more sisters than it would be for the sisters to attempt to have female offspring of their own. Thus, the hymenoperan genetic system may have served as a preadaptation favouring the easy evolution of societies founded by female reproductives that produce sterile female offspring that in turn assist their mothers in rearing other nonsterile female and male offspring. Ant, wasp, and bee colonies, of course, are exactly that kind of society. The sterile workers are all females; and they almost always assist their mothers (or other close relatives).

In considering the notions of inertia and preadaptation, it is very important to recall that there are two things that determine whether a species can respond to a changed environment by evolving new behaviour patterns. The first is the amount of behavioural variability in the population. Natural selection acts by causing individuals with some traits to be more likely to reproduce or to produce more offspring than individuals that possess other traits. With respect to behavioural evolution, it operates by making individuals that behave in some ways to be more successful at reproducing than individuals that behave in other ways. For such a process to work, there must be variability. Different individuals must behave in different ways. If all individuals are alike, there is nothing for natural selection to select.

The second factor that is necessary for natural selection to produce behavioural evolution is that some of the differences between individuals must be genetically determined. For natural selection to work, the offspring of successful parents must more likely to exhibit similar forms of successful behaviour than the offspring (if any) of less successful parents. So far as behavioural evolution is concerned, this means that the behaviour selected must be somehow related to the animal's genetic endowment. However, this point does not mean that the behaviour need be innate in the sense of being unmodifiable by the environment. Natural selection does sometimes tend to select very stereotyped behaviour patterns that experience can modify very little. However, at other times, it operates to modify an animal's capacity to learn something. For example, we know that even with respect to such fundamental matters as the capacity to recognize another member of one's own species, natural selection has produced some species in which the capacity is innate and others in which it is only something that is easily learned. Generally, behaviour is determined by interactions between genes and experience. However, for a behaviour to evolve via natural selection, something about the behaviour must be genetically determined.

A useful way to consider behavioural variability is a model based on the statistical concept of variance. For any characteristic that can be measured on an interval scale, the variance of the characteristic is:

Where:

x = any individual measurement

n = the number of measurements

This concept is particularly useful because a particular variance may be considered to consist of a sum of several component parts or subvariances. With respect to behavioural and other characteristics which involve both genetic and environmental effects, the phenotypic variance seen in a population is often considered to consist of at least three components as follows:

Where:

= the contribution of genes to the phenotypic variance.

= the contribution of environment to the phenotypic variance.

= gene-environment interactions.

The only one of these designated components on which natural selection can act is .

The proportion of the total phenotypic variability for a particular characteristic which is attributable to the average effects of genes in a particular environment is called the heritability of the character. It is symbolized by the term . Heritability can be estimated in ways which we will discuss below. However, first I want to stress that estimates of heritability are always limited to particular environments; and the heritability of characteristics can be different in different environments.

Within a particular group of organisms maturing in a relatively homogeneous environment, total phenotypic variability () is the sum of the variance attributable to genetic differences ( ), variance attributable to environmental influences ( ) , and variance attributable to environment-gene interactions ( ).

Heritability in the broad sense () is then defined as follows:

It is possible to partition genetic variance ( ) into three components such that:

Where:
= variance due to the additive effects of genes contributing to various individual genotypes. Some of the genes cause more of the characteristic to develop, some less. The sum of the effects of these genes assembled in each individual helps to determine the degree to which the characteristic has the potential to develop. This component of variability may be directly selected.= variance due to dominance deviations, i.e. to differences in the degree to which different genes at the same locus will tend to be dominant.
= variance due to epistatic interactions. A gene may block or enhance the expression of another gene at a different locus. For example, the allele b 1 might suppress the expression of a characteristic caused by the allele a 1 at a different locus. In contrast, the allele b 2 might have no effect on the expression of a 1 ; and the allele b 3 might enhance the effect of a 1 .

From these three components of genetic variance, it is possible to separate out a narrower measure of heritability that permits a more direct estimate of the rate at which evolution can occur. This new estimate is called heritability in the narrow sense (); and it is defined as:

The speed at which a trait will evolve in a population will then be the product of and the intensity of the selection process. More precisely:

Where:

R = the response of the population to selection.
= heritability in the narrow sense.
S= a parameter determined by the proportion of the population effected and the strength of the selection pressure.

Two further points should probably be made at this point in an effort to avoid confusion. First, although capacities to learn certain things can very likely be inherited, there is no evidence that specific learned behaviour patterns can be inherited. The notion that behaviours which an individual has learned in its lifespan can be inherited by that individual's offspring is one which has been totally discredited in the 20t h century.

The second cautionary note is to reiterate that we are talking here only about biological evolution. Evolution of customs and cultures can of course occur as a result of a kind of selection that is not genetic. Parents can teach things to their offspring, and offspring can learn by imitating their parents and then learning even more during their own lifetimes. This parent-teaching-offspring type of transmission of behaviour patterns occurs, especially in our own species but also in other kinds of vertebrate animals that live in complex societies. For example, learned traditions about migration routes and certain other matters are thought to exist in herds of caribou, bison, wild sheep, and other ungulates. Nevertheless, biological or genetic evolution has undoubtedly been far more widespread and much more important than the evolution of traditions and cultures. Moreover, even where cultural evolution is possible, it is possible only in animals that have the genetic make-up necessary to permit them to live in close parent-offspring groups where the transmission of culture can occur.