BOTANY, VOLUME 3 - EVOLUTION AND SYSTEMATICS - 2007

10. EVOLUTION

10.2. Types and Causes of Natural Variation

10.2.1. Natural Selection

Examples of intraspecific genetic variation observed in nature are rarely random; instead, they typically correlate with environmental conditions and the Specific characteristics of the plant species under study. As early as the beginning of the 18th century, researchers noted intraspecific variations such as differences in flowering times among forest trees of northern and southern origins. However, prior to the experiments conducted by Swedish botanist G. Turesson in the 1920s, there was no consensus on whether such intraspecific variation had a genetic basis or was largely the result of immediate habitat conditions. By cultivating plants of diverse origins—belonging to widespread species found across various ecological niches—under identical conditions, Turesson demonstrated that the differences observed in natural habitats were, in most cases, at least partially preserved in cultivation. From this, he concluded that genetic variation for the studied traits exists within species. Because material from different sites with similar ecological conditions (Fig. 10.25) (e.g., dune populations of the umbrella hawkweed, Hieracium umbellatum) consistently exhibited similar traits and differed from material originating from sites with contrasting conditions (e.g., coastal cliff populations), Turesson established that intraspecific genetic variation correlates with habitat conditions. These findings and Conclusions have since been confirmed and refined by numerous comparative studies. Consequently, it is now well established that environment-correlated intraspecific genetic variation is a widespread phenomenon.

Class="center">Fig. 10.25. Ecological races of the California yarrow (Achillea lanulosa, tetraploid) collected from different elevations (1,400 m, 3,350 m, and 2,100 m) along a ~60 km transect across the Sierra Nevada and the adjacent Great Basin (approx. 38° N). Approximately 60 individuals from each population were grown from seeds in Stanford (30 m). The diagrams illustrate the inherited variation in SHOOT height, with mean values indicated by arrows alongside a representative individual from each population.

Along a west–east transect through California—spanning from the coast across the high elevations of the Sierra Nevada mountains to the bordering Great Basin—genetic variation correlated with climatic factors has been documented in numerous species. For instance, yarrow plants (Achillea lanulosa) originating from various elevations of the Sierra Nevada differ in height from plants of the same species growing in the Great Basin (see Fig. 10.25). A Study of the widely distributed North American grass Andropogon scoparius revealed that northern forms require a longer photoperiod (15 hours of light per day) to stimulate flowering than plants of southern origin (14 hours per day). To understand this observation, it helps to know that summer day length increases with latitude, and that flowering initiation in many plants is photoperiod-dependent. Correlations between genetic variation and specific soil properties have also been demonstrated in Achillea borealis. In experimental settings, forms collected from serpentine soils thrive when grown on serpentine soils (serpentine being a rock type that gives rise to magnesium-rich, lime-deficient soils). Conversely, forms of the same species that do not naturally occur on serpentine soils grow poorly when cultivated experimentally on such substrates. Similarly, it has been shown that populations of Anthoxanthum odoratum from lime-poor habitats require significantly less calcium for growth than conspecific plants originating from lime-rich habitats. Numerous species (e.g., Agrostis capillaris, Anthoxanthum odoratum, Festuca ovina, Mimulus guttatus, Plantago lanceolata, Rumex acetosa) possess genotypes that enable them to grow on soils heavily contaminated with heavy metals such as cadmium, copper, and zinc, whereas other genotypes lack this tolerance. Intraspecific differentiation can correlate with biotic factors as well as abiotic ones. Like many other species, white clover (Trifolium repens) contains a cyanogenic glycoside whose enzymatic breakdown—triggered, for example, by tissue damage—releases hydrogen cyanide. This species is polymorphic for this trait. It has been shown that the relative frequency of cyanogenic (glycoside-producing) versus acyanogenic (glycoside-free) genotypes correlates not only with microclimatic habitat Temperature (Fig. 10.26) on a local scale, but also with the Abundance of grazing snails on a broader scale. Unlike plants from undisturbed areas, specimens of Plantago lanceolata inhabiting heavily trampled paths tend to exhibit a prostrate or ascending growth habit rather than an upright one. In addition to the examples mentioned above, intraspecific differentiation has been documented with respect to

photosynthetically active radiation, drought tolerance, frost hardiness, resistance to parasites and herbivores, pollinator attraction, and competition with neighboring plants, among other factors (Fig. 10.27).

Fig. 10.26. Clinal variation in white clover (Trifolium repens). The frequency of allele A, which is responsible for cyanogenic glycoside production (with the alternative allele designated as a), within populations (represented by circles with black [= A] and white [= a] sectors), correlates with January isotherms from the Mediterranean to Northern Europe (left map), and with altitude in the Alps (circular diagrams on the right).

Fig. 10.27. Ecological differentiation in seed plants: A — ecotypes of the two-styled sorrel (Oxyria digyna, Polygonaceae) and their distinct physiological reaction norms showing mean rates of Photosynthesis (circles) and Respiration (triangles)—measured in mg CO₂ per square decimeter of leaf surface area as a function of temperature—for the southern alpine (1) and northern arctic (2) races; B — clinal variation in Scots pine (Pinus sylvestris): under identical cultivation conditions, 52 plants of European origin exhibit a strong correlation between needle dry weight (as a proxy for cold hardiness) and the day length on the first spring day (mean temperature +6 °C) in their natural habitats (serving as an indicator of geographical latitude, continentality, and growing season length).

The detailed spatial distribution of genetic variation in nature can take several forms. If an environmental factor that correlates with a plant trait is discontinuously distributed in nature (for instance, alkaline versus acidic soils overlying calcareous and silicate rocks in the Alps), plants may exhibit more or less discrete variation. Conversely, if an environmental factor changes continuously (such as declining mean temperatures with increasing altitude), continuous genetic variation is likewise expected in the plant material under study. These two patterns of intraspecific variation have long been recognized as ecotypic (ecotype) and ecoclinal (ecocline) variation, respectively. Distinguishing between these patterns can sometimes be complicated when the analyzed trait correlates not with a single environmental factor, but with two or more spatial variables changing in different directions.

The observed patterns of intraspecific genetic variation also depend heavily on experimental parameters. For example, comparisons between spatially distant populations generally reveal sharper differences than comparisons between adjacent populations, and comparisons of seed-grown progeny from two populations typically display less pronounced differences than comparisons of mature plants collected directly from their natural habitats.

Natural selection is widely recognized as the primary driver of intraspecific variation that correlates with environmental factors. The Mechanism of natural selection relies on the fact that different individuals (genotypes) within a population vary genetically in their reproductive success within a given environment. Differences in reproductive success manifest as varying probabilities of surviving to reproductive age, and—subsequently—as differential reproductive output once the reproductive phase is reached. These varying chances of survival and differential reproductive outputs can be collectively termed the fitness of a genotype. Thus, the direct target of natural selection is the individual genotype or, more precisely, its realized phenotype, i.e., the individual Organism. If different genotypes within a population possess differing fitness values, natural selection occurs. As a result of natural selection, relative allele frequencies may shift across different developmental stages within a single generation or across successive generations. Changes in allele frequencies over successive generations can be defined as evolution. A concrete (though theoretical) example of evolution via natural selection is provided by the yarrow species Achillea lanulosa along its California transect. If two genetically identical populations with heritable variations in plant height (which influences flower number and development rate to first flowering) were cultivated at the coast versus high alpine altitudes, their development over several generations would likely diverge along different paths. In the coastal population, taller plants with greater flower numbers would, on average, produce more offspring than shorter plants with fewer flowers. Under a relatively constant population size, this would lead to an increase in the relative frequency of tall plants. In the mountains—where tall plants might fail to set seed altogether due to the short growing season, or where only their earliest flowers could produce viable seeds—the relative frequency of short-statured plants would increase. Thus, natural selection accounts for both evolutionary change and genetic adaptation, since plant height Variability can reasonably be regarded as an adaptation to prevailing growth conditions. Consequently, environment-correlated variation should, in most cases, be understood as adaptive variation.

A trait can be interpreted as a genetic adaptation if it is heritable and enhances the probability of survival or reproductive success—that is, if it increases the individual's fitness. The term "adaptation" is used to denote both the dynamic process of becoming adapted and the resulting state of being adapted. The process of genetic adaptation begins when a novel trait arises through mutation, which is an entirely random event. If this new trait enhances an individual's fitness, the second step in the establishment of a genetic adaptation is the spread of that trait through the population via natural selection.

To conclusively demonstrate The Role of natural selection as a cause of natural genetic variation, observing a correlation between genetic variation and environmental heterogeneity is insufficient. It must be repeatedly shown that different genotypes exhibit differential fitness under varying environmental conditions, or that a genotype possesses higher fitness in its native habitat than in a foreign environment. This can be accomplished

either by subjecting different genotypes¹ to a specific decisive environmental factor and measuring their fitness, or by performing reciprocal transplant experiments with material of diverse origins in nature. Finally, one can assess fitness differences among individuals within a single population under natural conditions.

¹ All experiments are conducted not directly on genotypes, but on organisms, i.e., phenotypes possessing different genotypes. — Ed. note.

The first research approach was implemented using greater plantain (Plantago major). It was observed that plants growing along heavily trodden pathways feature more or less prostrate flower scapes, whereas plants from untrampled areas display upright scapes. This morphological difference persists in cultivation under identical growth conditions. Experimental trampling of plants of diverse origins (applying appropriately weighted metal blocks repeatedly to simulate the pressure exerted per unit area by an adult human) demonstrated that prostrate scapes suffer a statistically significant smaller reduction in fitness (measured as the dry weight of reproductive Organs) under artificial trampling than upright ones. In the control group (without trampling), the fitness of prostrate and upright forms was nearly identical. In the reciprocal transplant experiments, populations of the same species from contrasting habitats were swapped. For instance, coastal populations of the Californian Achillea lanulosa were transplanted to the mountains, and high-altitude populations were moved to the coast. In both cases, the fitness of populations in foreign habitats was found to be lower than in their native environments. Both types of experiments support the Conclusion that natural phenotypic differences observed among populations inhabiting different ecological conditions are the products of natural selection and can be interpreted as adaptive differences. Reciprocal transplantation has a distinct advantage over the experimental analysis of a single environmental factor, as it allows for the measurement of fitness relative to the environment as a whole, bypassing the need to isolate individual environmental factors that are often difficult to identify yet hypothetically crucial for variation.

It remains unclear whether every phenotypic trait of an organism can be interpreted as an adaptation. An alternative possibility is that mutation-driven traits are selectively neutral and maintained through internal genetic mechanisms (such as pleiotropy or genetic linkage). Mutations leading to selectively neutral traits (as they do not impact fitness) can be expected to occur relatively frequently and are eliminated only slowly by natural selection. A certain constancy of traits might therefore indicate that limited trait variation is the outcome of natural selection. Accordingly, one cannot automatically assume that every single organismal feature represents a perfected adaptation. Instead, biological structures or traits typically maintain multifaceted interactions with their environment and should thus be understood as compromises resulting from divergent selective pressures acting in various directions. A single trait encoded by a pleiotropic Gene or a tight linkage group may be maintained by natural selection, whereas another trait encoded by that same gene or gene complex may not, rendering the latter feature non-adaptive. Finally, a trait might represent an ADAPTATION TO ENVIRONMENTAL conditions that no longer exist, leaving insufficient time for subsequent evolutionary modification. For instance, the exceptionally large and tough fruits found in numerous South American tree species (e.g., Crescentia alata, Bignoniaceae; Scheelea rostrata, Arecaceae) are interpreted as adaptations to large mammalian herbivores that went extinct approximately 10,000 years ago.

Depending on the phenotypic outcome of evolutionary change, natural selection is classified into three distinct forms: directional, disruptive, and stabilizing (Fig. 10.28). In directional selection, within a population displaying a normally distributed range of genetic variation, individuals at one of the phenotypic extremes exhibit the highest fitness. Over successive generations, this results in a progressive shift of the variation curve toward that extreme phenotype. In disruptive selection, genotypes located at both extremes of the variation curve possess higher fitness than individuals situated near the center of the spectrum, which initiates the splitting of the population into two distinct groups. Conversely, in stabilizing selection, genotypes exhibiting intermediate population values possess the highest fitness, whereas genotypes at both extremes show low fitness. Such a fitness distribution leaves the overall genetic variation unchanged across successive generations. The example of stabilizing selection clearly demonstrates that the evolutionary outcome of natural selection does not always entail a shift in allele frequencies.

Fig. 10.28. Stabilizing, directional, and disruptive selection. The range of variation (abscissa) in ancestral populations (bottom) reflects the frequency of individuals with different hereditary backgrounds: it either remains stable (A), shifts directionally (B), or splits into two distinct groups (C) as a result of different forms of selection (indicated by arrows).

Intraspecific genetic variation correlates not only with environmental factors. The analysis of genetic variation, particularly at THE MOLECULAR LEVEL (e.g., isozymes), has shown that this type of genetic variation is partially correlated with and governed by the properties that collectively constitute the recombination system of a species. A comparison of intrapopulation and interpopulation variation is of primary interest here. The pollination system generally Functions in such a way that self-pollinating species typically exhibit relatively low intrapopulation and relatively high interpopulation variation. In cross-pollinating species, the relationship is inverse: variation is higher within populations than between populations. These patterns are explained, on the one hand, by the genetic effect of continuous self-Fertilization and, on the other hand, by the varying intensity of genetic exchange (gene flow) between populations in self- and cross-pollinating species. Self-fertilization leads to the loss of genetic variation within a population while simultaneously promoting the isolation of neighboring populations. METABOLISM/18.html">The Influence of gene flow on The Structure of intraspecific genetic variation is also evident when comparing different mechanisms of pollen and diasporae dispersal. Wind-pollinated and wind-dispersed species exhibit relatively greater intrapopulation and relatively lesser interpopulation variation compared to species pollinated and dispersed by animals, as well as species lacking specialized dispersal adaptations. This suggests that gene flow distances, and consequently genetic exchange between populations, are on average greater in wind-pollinated and wind-dispersed species than in those pollinated and dispersed by animals. Life form also affects the expression of variation: annual species show lower intrapopulation and higher interpopulation variation than short-lived herbaceous perennials, which in turn exhibit lower intrapopulation and higher interpopulation variation than long-lived woody perennials. A potential explanation for these patterns lies, on the one hand, in the known correlation of self-pollination and restricted gene flow with short life span, and of cross-pollination and enhanced gene flow with long life span, and, on the other hand, in the fact that over time, populations of long-lived species are more likely to become genetically enriched through gene influx from other populations compared to populations of short-lived species. Finally, in species with asexual or partially asexual reproduction, genetic variation is substantially lower than in exclusively sexually reproducing species.

Box 10.2. Population Genetics

Changes in allele frequencies across successive generations are quantitatively investigated in population genetics. Thus, population genetics seeks to answer the question: what will the genotype or allele frequencies become in a given generation if their frequencies in the preceding generation are known? Allele frequencies can be derived from calculated genotype frequencies. For a single genetic locus with alleles A and a, there are three genotypes—AA, Aa, and aa—with frequencies P, Q, and R, respectively. The frequency p of allele A is calculated by dividing the sum of twice the frequency of the homozygous genotype AA and the frequency of the heterozygous genotype Aa by twice the total number of individuals in the population:

Twice the frequency of the AA genotype and twice the number of individuals must be used because each diploid individual contains two alleles, making the total number of alleles in the population twice the number of individuals. Since the sum of the frequencies P, Q, and R of the three genotypes accounts for all individuals in the population and equals 1, the formula given above can be simplified to: p = P + Q/2. The frequency q of allele a is calculated analogously using the formula q = R + Q/2. Since the sum of both allele frequencies equals 1 (p + q = 1), q can also be determined as: q = 1 - p. Conversely, genotype frequencies can only be calculated from allele frequencies under specific conditions. Calculating genotype frequencies in the next generation from allele frequencies in the previous generation is possible if:

✵ the fitness of the three genotypes AA, Aa, and aa (as in the example) is equal,

✵ genotypes mate at random,

✵ no new alleles arise (through mutations or gene flow),

✵ the population is sufficiently large to rule out random fluctuations in allele frequencies.

In this case, the frequency of AA is p², the frequency of Aa is 2pq, and the frequency of aa is q². This fundamental relationship between allele frequency and genotype frequency in population genetics is known as the Hardy-Weinberg law. It implies that under constant conditions across successive generations, there is no shift in allele and genotype frequencies, and consequently, no evolution occurs. However, one can test whether a population is in equilibrium or not. From the frequencies of genotypes AA, Aa, and aa present in a population, the frequencies p of allele A and q of allele a can be calculated. If the population is in equilibrium, the frequencies of genotypes AA, Aa, and aa must equal p², 2pq, and q². If the observed genotype frequencies deviate from those expected under the Hardy-Weinberg law, the population is not in equilibrium. The cause of this could be unequal fitness among the three genotypes. The Effect of unequal fitness on allele frequencies in subsequent generations can be calculated. For simplicity, it can be assumed that the fitnesses of AA and Aa are equal and set to 1, while the fitness of aa is lower: 1 - s. The letter s denotes the selection coefficient, which equals one minus fitness; in the simplest case, fitness can be calculated as The ratio of the survival rate of a given genotype to the survival rate of the most fit genotype. In our example, genotypes AA and Aa have a fitness of 1, whereas the survival rate of genotype aa might be 90% of that of AA and Aa. Its fitness would then be 0.9, and the selection coefficient s = 0.1. In the initial population, the genotype frequencies are AA = p², Aa = 2pq, and aa = q²(1 - s). The total population size is: p² + 2pq + q²(1 - s). If p is replaced by 1 - q in this formula (since p + q = 1, whence p = 1 - q), the expression for the total population size simplifies to: 1 - sq². The genotype frequencies in the population will then be:

The frequency p₁, modified under the action of selection, will be:

(Frequency of AA + half the frequency of Aa as a proportion of the total population.) The change in frequency during selection:

If initially p = q = 0.5 and s = 0.1, then

Thus, the frequency p₁ equals 0.5128, and the frequency q₁ equals 0.4872. This formula makes it possible to calculate and predict the change in allele frequency across successive generations under random mating and differential genotype fitness. If we assume that the selection coefficient of the homozygous recessive genotype aa = 1, i.e., individuals of this genotype never reach reproductive maturity, the frequency q_n of allele a after n generations can be calculated from its initial frequency q₀:

With an initial frequency q₀ = 0.5, the frequency q_n of allele a after, for example, 10 generations (n = 10) of constant elimination becomes q₁₀ = 0.083. This demonstrates that even very low fitness of a homozygous phenotype cannot lead to the complete elimination of the allele responsible for it, because that allele is retained within heterozygous genotypes. This underscores the great Evolutionary Significance of diploidy and the associated heterozygosity of a vast number of organisms. Such calculations become more complex in cases where genotypes differ not in survival rate but in reproductive success, when mating within a population is non-random, when gene flow occurs between populations, or when genotype fitness changes across generations, for example, in a frequency-dependent manner. Quantitative analysis becomes even more intricate when selection is observed not at a single genetic locus, but at two or more loci that mutually influence one another. At the same time, this corresponds more closely to real-world conditions, since presumably very few phenotypic traits are encoded by a single genetic locus alone.

10.2.2. Genetic Drift

Another factor crucial to the genetic structure of species is chance. Random shifts in allele frequencies from one generation to the next are termed genetic drift. Such events can be triggered, in particular, by a sharp reduction in population size, with the probability of random allele frequency shifts increasing as the population shrinks. As a result of genetic drift, one of the alleles may be lost, leaving all individuals homozygous for the alternative allele at a given genetic locus (fixation). However, the genetic impact of a population bottleneck only becomes noticeable when population size drops below 10 individuals and remains low for several generations, resulting predominantly in Inbreeding, and provided that one of the locus alleles was rare in the initial population. Population size may be small, for instance, because a few individuals or even a single individual (hermaphroditic and self-fertilizing) reached an island via long-distance dispersal or accidental human Introduction and founded a new population (founder effect). Dramatic environmental changes can also drastically reduce population size (bottleneck effect).

An example of the founder effect is the grass Echinochloa microstachya, introduced by humans from North America to Australia. While in North America each population of this species is distinguished by its own characteristic allele combination, 18 of the 20 populations studied in Australia proved to be genetically identical. However, such uniformity is not always observed. For instance, silky bentgrass (Apera spica-venti), introduced from Europe to North America, exhibits a similar breadth of genetic variation on both continents.

Continuous fluctuations in ice sheet thickness during the Quaternary period and the associated climatic changes profoundly impacted the population sizes of all organisms living at that time. This explains why, in refugia, numerous tree species are characterized by relatively high genetic variation, whereas in northern areas colonized by them at the end of the glacial period, populations are genetically impoverished (Fig. 10.29). This pattern of genetic variation could have arisen because the recolonization of glaciated territory did not proceed through the gradual range expansion of previously retreated populations. In areas being newly repopulated, population fluctuations driven by rapid climate shifts constantly occurred, or single individuals / small groups of individuals may have established themselves far ahead of the advancing population front.

Fig. 10.29. Geographic distribution of 11 different plastid types of Alnus glutinosa. Greater genetic diversity is found in glacial refugia in Southern Europe than in the areas recolonized from them. Circles with black and white segments indicate the heterogeneity of populations regarding plastid types.

Observations explaining the pattern of genetic variation through specific plant traits (such as the mating system) or stochastic events raise the question of the role of natural selection. A more general question is to what extent genetic variation is neutral, i.e., to what extent different genotypes share equal fitness. Regarding phenotypic traits, it can generally be assumed that different genotypes frequently exhibit varying fitness. Consequently, the pattern of genetic variation in phenotypic traits is often the result of natural selection. Regarding molecular traits¹, it was postulated—primarily by M. Kimura—that the majority of genotypes present within a population possess equal fitness and are therefore selectively neutral. At the same time, there is no doubt that the effect of most new mutations (see Fig. 10.29) is deleterious, as these mutations are eliminated by selection. Kimura hypothesized that genetic variation at the molecular level is primarily the result of genetic drift rather than natural selection.

¹ Excluding molecular traits from phenotypic traits is an error. — Note by the Editor.

This neutral theory of molecular evolution is supported by rates of Protein Evolution that are too high to be accounted for by natural selection, the remarkably high protein polymorphism within species, the relative constancy of evolutionary rates for Proteins across different lineages, and the higher evolutionary rates observed in non-functional regions of Enzymes compared to functional ones. The line of argument, initially based on the Analysis of Protein sequences, was later refined in light of somewhat different findings from DNA nucleotide sequence analyses.

The question of whether genetic drift or natural selection exerts a stronger influence on molecular evolution depends largely on assumptions regarding population sizes, mutation rates, and selection coefficients. Since these parameters typically cannot be quantified precisely and must be estimated roughly, specific variants of genetic variation can be interpreted through either the neutralist or selectionist models of evolution, depending on the estimates adopted. Consequently, the relative roles of drift and selection cannot be definitively resolved. What is certain, however, is that cryptic mutations or certain non-transcribed DNA sequences evolve neutrally, meaning that both drift and selection shape molecular evolution. Furthermore, genetic drift may also play a role in phenotypic evolution.