BOTANY VOLUME 3 - EVOLUTION AND SYSTEMATICS - 2007

11. SYSTEMATICS AND PHYLOGENY

The goal of systematic research is to organize the infinite diversity of organisms encompassing countless forms and lifestyles. To achieve this, one must be able to recognize species and group them into higher-level taxonomic ranks (genera, families, etc.). Systematics also involves describing individual species and higher taxa, assigning them distinct names, and creating identification keys. Building upon The Theory of biological evolution established by Charles Darwin (see Chapter 10), systematics groups species into genera, genera into families, and so on, in a way that reflects the natural evolutionary relationships among organisms and groups of organisms. The sole objective basis for grouping provided by nature itself is the evolutionary history of organisms. In recent years, the goal of uncovering phylogenetic relationships has advanced significantly, as systematic analysis increasingly relies on data contained within organismal DNA. Today, this information—combined with data on "traditional" traits, such as morphological characteristics—is successfully interpreted using highly mathematical and therefore powerful Methods, alongside statistical testing of proposed phylogenetic hypotheses. Consequently, systematics has evolved into a methodologically thoroughly modern science.

Systematics is the first and most crucial step in studying biological diversity. It provides a reference framework for all other biological disciplines and all branches of our economy that deal with organisms or their products, ensuring the precise identification and naming of every Organism and, consequently, clear relationships among them. Systematics is also an indispensable foundation for understanding the evolutionary contexts of biological phenomena. For example, if a cytologist investigates the evolutionary Water/144.html">Origin of the nuclear membrane, a geneticist explores The Emergence of introns, a morphologist attempts to reconstruct The Structure of the earliest flowers, a physiologist studies the evolution of various photosynthetic pathways, or an ecologist analyzes THE ORIGIN OF diverse plant soil requirements, the phylogeny reconstructed by the systematist—identifying ancestral and derived organisms—serves as essential foundational knowledge of the subject.

This chapter first introduces the methods of systematic research, followed by An Overview of the structure and systematics of Bacteria, Fungi, and plants.

11.1. Methods of Systematics

11.1.1. Species Recognition

The species is the fundamental unit of systematics. This does not imply that the species must also be the primary unit of evolutionary change (see 10.3). While numerous species concepts exist (see 10.3.1), most species are recognized in practice based on discontinuous phenotypic variations. Thus, if a group of individuals can be clearly separated into two subgroups—for instance, based on leaf length, pedicel pubescence, flower color, and fruit size (see Fig. 10.30)—and the variations of the traits in question are continuous within these subgroups, the subgroups can be regarded as distinct species. This method is as objective as the reliability with which the observed discontinuity of metric and meristic traits can be calculated and replicated. In all such cases, it is first necessary to demonstrate that one is working with a homogeneous and comparable representative material. Nevertheless, the degree of phenotypic discontinuity required for species recognition is determined subjectively. Phenotypic discontinuity is a documented fact, whereas assigning the identified subgroups of a study population to variants, subspecies, or species remains merely a formal nomenclature issue. It is universally accepted for species that the individuals belonging to them share a common ancestor and are therefore not polyphyletic.

The phenotypic traits used for species recognition are fundamentally the same (see 11.1.3.1) as those applied to group species into higher taxonomic ranks. In practice, most traits are Anatomical and morphological, since initial species identification is most often performed using museum material—specifically herbaria. The clear expression of such traits, both in nature and in museum collections, simplifies assigning a given individual to a particular species. The immense importance of botanical museums (herbaria) lies primarily in the fact that only in these collections, often accumulated over hundreds of years, does a researcher have access to sufficiently representative material of the plant group under study. The world's largest herbaria contain approximately 6 million plant specimens, and thanks to established inter-herbarium material exchange, researchers can temporarily obtain required specimens from various collections. At the same time, it is critically important to study living plants, as certain traits, such as chromosome count, can only be examined in such material.

11.1.2. Monographs, Floras, and Identification Keys

Important outcomes of systematic research at the species level include monographs, which are systematic treatments of a specific group of related forms (e.g., a monograph of the genus Primula), and floras, which are systematic treatments of plants inhabiting a particular geographic region (e.g., the flora of Germany). Monographs and floras, on the one hand, provide descriptions of the studied plants and, on the other hand, facilitate plant identification through identification keys. Monographs additionally include all formal aspects of taxonomic Treatment (the botanist who first described the taxon as the author of the species, genus, etc., the date of description and publication name, synonyms, etc.). Descriptions may vary greatly in detail, but they must clearly illustrate both intra-specific or intra-group Variability and the boundaries distinguishing them from other species and groups.

An identification key can be constructed using a dichotomous or polytomous principle. Consider, for example, three traits with different states (see 11.1.3.2):

1 - 0 Petals red

1 - 1 Petals yellow

2 - 0 Fruit a berry

2 - 1 Fruit a capsule

3 - 0 Plant annual

3 - 1 Plant perennial and the distribution of traits among species:

In this case, a dichotomous key—the most commonly used form of identification keys—appears as follows:

1. Petals red 2.

 Petals yellow 3.

2. Fruit a berry  Species A

 Fruit a capsule Species B

3. Plant annual Species C

 Perennial plant, species D

In a polytomous key, all characters for a single species are grouped together:

 Species Character combination

A 1-0; 2-0; 3-1

B 1-0; 2-1; 3-0

C 1-1; 2-0; 3-0

D 1-1; 2-1; 3-1

The advantage of a polytomous key is that it allows species identification even when, for instance, flower color is unknown, since the combination of fruit type and life form alone is sufficient to distinguish all four species. In a dichotomous key, by contrast, species A and C, or B and D, cannot be distinguished without knowing the flower color. Admittedly, this drawback can be mitigated by including questions about two or more characters rather than just one in the leads and antitreads of the key. This is routinely done in well-constructed keys, but situations may still arise where the required character cannot be observed. Polytomous keys are increasingly provided in electronic format. For example, the interactive keys offered in the computer program 'Intkey' (by M.J. Dallwitz) offer numerous advantages over dichotomous keys. Above all, identification can be initiated using any character or any combination of characters. Information contained in monographs or floras is also partially available in the form of computer Databases.

11.1.3. Investigation of Relationships

Any taxonomic group, regardless of its rank, is generally called a taxon. The degree of relationship between taxa is determined by the relative age of their last common ancestor. Thus, taxon B is more closely related to taxon C than to taxon A if the last common ancestor of B and C is younger than the common ancestor of A and B (Fig. 11.1).

Fig. 11.1. Relationships. The degree of relationship is determined by the relative age of the last common ancestor. Taxon B is more closely related to taxon C than to taxon A because the last common ancestor (2) of B and C is younger than the last common ancestor (1) of A and B

Knowledge of Phylogenetic relationships among taxa is important from three Perspectives. First, relationships (which until recently could only be formulated as more or less well-supported hypotheses) constitute the sole objective basis for Classification and are therefore of paramount importance to the systematist. Second, a classification based on phylogenetic relationships has significant predictive power. This means that a specific property of a species is more likely to be found in its close relatives than in its distant ones. For example, certain morphine Alkaloids of the opium poppy (Papaver somniferum) are known only in the very closely related graceful poppy (Papaver gracile) and the slightly less closely related bracteated poppy (Papaver bracteatum), but are absent in more distant poppy species. However, phylogenetic relationship does not always have predictive power, as characters can be lost. Third, knowledge of relationships is the only way to reconstruct the evolution of characters in any given species. If, for instance, plant physiology addresses the question of whether C₄ plants evolved from C₃ plants among the Amaranthaceae only once or multiple times, this can be answered by examining not the character itself, but the phylogenetic relationships of the taxa possessing this trait.

11.1.3.1. Characters

The groups of characters used for Species Identification and, moreover, for investigating relationships—that is, for uniting species into higher-level taxa—are extremely diverse.

The historical development of systematics demonstrates that the expansion of observable characters is closely linked to The Development of new observation methods, increasing ease of use, and the general state of theoretical knowledge. Since antiquity, the primary basis for comparison comprised plant habit characters; then, from the 16th–17th centuries through C. Linnaeus and into the 19th century, macroscopic characters of flowers and fruits took precedence. With the widespread use of the Microscope in the 19th century, The Study of thallophytic PLANTS AND THEIR reproductive Organs began (E. M. Fries, H. A. de Bary, A. Pascher, W. Hofmeister, J.-B. Payer, etc.), along with the anatomical characters of SHOOT-bearing plants (e.g., B. N. Solereder, C. R. Metcalfe). Following the ESTABLISHMENT OF THE theory of evolution, paleobotany became an increasingly important source for investigating relationships (H. Solms-Laubach, R. Kidston, W. Zimmermann, et al.). As early as the second half of the 19th century, The Significance of distribution ranges for phylogenetic research was recognized (A. Kerner v. Marilaun, R. v. Wettstein, et al.). In the 1920s, data on chromosome number and structure began to be utilized in systematics (e.g., E. M. East, E. V. Babcock), and in the early second half of the 20th century, phytochemical data (R. E. Alston, R. Hegnauer, et al.) and Electron Cell/15.html">Microscopy data (I. Manton, e.g., regarding Algae) gained greater prominence. Over the last two decades of the past century, data on Protein Structure (e.g., L. D. Gottlieb — Isoenzymes) and Nucleic Acids (e.g., C. R. Woese, J. D. Palmer, M. W. Chase, D. E. & P. S. Soltis) became increasingly important, and they now dominate modern systematics.

Morphology (see Chapter 4), which describes the External structure of plants, is of great importance to systematics. Examples of morphological character diversity

include: the Cellular Organization of green algae (monadoid, coccoid, filamentous, tissue-like, thalloid, etc.; see Chapters 5; 11.2, Chlorobionta), opposite or alternate leaf arrangement, and the number and arrangement of stamens in the flower (see 11.2, arrangement of floral parts). To assess the structural homologies of morphological characters, it is often necessary to study their development (ontogeny). For instance, multi-stamened androecia can arise in entirely different ways (see Fig. 11.173). Anatomy investigates internal structural features. In this context, Histology (see Chapter 3) deals with tissue structure, such as the collateral or bicollateral arrangement of xylem and phloem in the vascular bundles of angiosperms, or the presence of true Tissues versus plectenchyma in green and red algae (see 11.2, Chlorobionta, Rhodophyta). Cytology (see Chapter 2), another branch of anatomy, deals with the Fine Structure of Cells. Thus, the STRUCTURE OF THE flagellar basal body (see Fig. 11.88) and the detailed course of Cell Division in green algae, or the ultrastructure of sieve tube Plastids in angiosperms (see Fig. 11.167), are of great significance for systematics. Karyology, as a branch of cytology, investigates the Chromosomes of the Cell Nucleus. The chromosomal structure in interphase and predominantly in metaphase—revealed by various staining techniques—as well as the entire chromosome Complement of The Cell, prove to be variable when comparing different taxa. Palynology and Embryology can also be considered subdisciplines of anatomy. Palynology focuses on the structure of spores and pollen grains (see Figs. 11.174, 11.175, 11.176). It reveals differences in The Nature of spore wall layers, the electron-microscopically visible spore surface sculpture, and the number, shape, and arrangement of apertures. The development of sporangia, gametophytes, gametangia, endosperm, and embryo falls within the domain of embryology. Differences in the origin of the sporangial nutritive layer (tapetum), for example, form the basis for dividing ferns into groups, whereas the Development of the female gametophyte (embryo sac) in angiosperms varies in the number of megaspores involved in its formation and the final number of constituent cells (see Fig. 11.187). Phytochemistry studies the Structure of Plant-derived substances. This may involve Photosynthetic Pigments (various chlorophylls and accessory pigments; see 6.4.2), reserve CARBOHYDRATES (starch, inulin), or classes of substances common to a wide range of angiosperms, such as alkaloids, terpenoids, or Flavonoids (see Fig. 11.168). Furthermore, the systematist can obtain important characters from physiology (e.g., C₃ and C₄ Photosynthesis; see 6.5), ecology (e.g., soil requirements), geographic distribution as revealed by plant geography, or phytopathology (e.g., regarding the host Specificity of pathogenic fungi or herbivores; see 9.3, 9.4); fossil forms studied by paleobotany are particularly crucial to consider when examining relationships. Today, experimental systematics—which deals with Hybridization and interprets crossing success as a criterion of relationship—is losing its prominence. Conversely, PROTEIN AND NUCLEIC acid analyses have become especially significant for phylogenetic research today. While serological protein analysis (comparative precipitation of Proteins by specific Antibodies) and amino acid sequencing are also losing ground, isoenzyme analysis (determining the number of genetic loci and the number and quality of enzyme alleles via Electrophoresis; see Box 10.1) remains a standard Procedure for characterizing the genetic constitution of populations or closely related species. RNA and, above all, DNA analyses began with DNA hybridization—a method now practically obsolete—where phylogenetic inferences were based on the photometrically recorded rate of duplex DNA formation from single-stranded DNA fragments isolated from different species. DNA analysis via restriction fragment length polymorphism (RFLP), which cuts DNA into specific fragments (e.g., using 4- or 6-base-pair recognition sites), is also receding into the Background. In RFLP analysis, the DNA of different species is cleaved using identical restriction Enzymes, and variations in the resulting fragments are caused by Mutations within the Cleavage sites. These mutations are utilized as characters. With the now predominant RNA/DNA Sequencing, The nucleotide sequence of a nucleic acid is determined, and every nucleotide position is treated as a character. Today, plant systematics most frequently employs nucleotide sequences of genes or Gene regions from the plastid rbcL (the large subunit of ribulose-1,5-bisphosphate carboxylase/oxygenase; other increasingly used plastid genes include, for example, matK and ndhF, or specific plastid introns and spacers) and nuclear ribosomal DNA. Nuclear ribosomal DNA includes both ITS sequences (internal transcribed spacers) and 16S/18S RNA sequences encoding the small ribosomal subunit. Mitochondrial Genes have also begun to be utilized. The preferential use of plastid DNA over nuclear DNA is due to the fact that most nuclear genes occur in multigene families, which creates problems in their homologization, whereas plastid genes are typically present in single copies. However, plastid DNA is predominantly maternally inherited (see 10.1.2.5). Particularly informative macromutations of DNA, such as insertions or inversions revealed through sequencing or RFLP analysis, are especially valuable because they rarely arise in parallel in different species. For DNA analysis at low taxonomic levels, numerous fingerprinting methods are still employed today (see Box 10.1), sometimes even enabling the differentiation of individual specimens, mostly with the aid of the Polymerase Chain Reaction (PCR).

Introduction/20.html">DNA Structure offers advantages over other plant characters in many respects. This is primarily because DNA data can be unambiguously coded, allowing the exact position of NUCLEOTIDES within a DNA sequence to be determined. By contrast, assigning a morphological character to one of several alternative states is often very difficult or even impossible due to the existence of transitional forms. This is particularly evident when working with closely related species, where differences are frequently purely quantitative. Another advantage of DNA data is that they allow the comparison of organisms that share very few phenotypic traits (e.g., unicellular algae and seed plants) but possess identical genes (e.g., rbcL). Finally, a vast amount of characters can be derived from DNA structure; however, it should be kept in mind that sequencing a single gene analyzes numerous nucleotide sequences while treating them as a single gene. Conversely, phenotypic analysis encompasses a broad spectrum of diverse traits and thus involves a very large and largely unknown number of genes.

11.1.3.2. Character "Conflicts"

Depending on the taxon under investigation and its taxonomic rank, different groups of characters are employed. For the subsequent Discussion of character utility, it is useful to distinguish between a character and its state (e.g., character vs. character state). For instance, flower color is a character, whereas red, white, blue, etc., are the states of that character. The need to develop methods for evaluating characters stems from the fact that not all characters invariably support the same grouping of objects. When considering multiple characters, character "conflicts" are virtually inevitable.

When examining three taxa—A, B, and C—and two characters with states 0 and 1, for example, it may turn out that character 1 groups taxon A + B against taxon C, whereas character 2 groups taxon A with taxon B + C (Fig. 11.2).

Fig. 11.2. Character "conflict." Characters 1 and 2, with their respective states 1-0, 1-1, 2-0, and 2-1, point to different phylogenetic relationships among taxa A, B, and C: character 1 groups A with B, while character 2 groups B with C

Character "conflicts" occur, first, because when comparing two (or more) taxa, identical character manifestations may be the result of convergence or parallelism (see 10.4)—meaning they evolved multiple times independently of one another—which is not necessarily apparent from the structure of the character itself. While establishing the Homology of a structural feature (such as a morphological trait) between two taxa can be supported by considering its relative position and ontogeny In addition to its specific qualitative attributes, no such auxiliary criteria exist for evaluating an identical nucleotide occupying the same position in the DNA of two taxa. Second, character conflicts arise because the significance of characters for inferring relationships varies depending on the time of their origin relative to the divergence time of the group under study (see 11.1.3.4).

Various approaches are used to resolve character "conflicts." The first option is to assign different "weights" to various characters to solve the problem. While it is undeniable that different characters carry different "weights" due to their varying complexity, it is equally difficult to objectively evaluate these relative "weights." As a result, classification variants emerge that contain a strong subjective component. Consequently, they can differ significantly despite relying on the identical set of characters. The objectification of organism classification concepts is pursued, on the one hand, by numerical systematics and, on the other hand, by phylogenetic systematics.

11.1.3.3. Numerical Systematics

Numerical systematics (phenetics) involves pairwise comparisons of all taxa to determine their mutual similarity and to calculate the structure of similarities within the studied group. In numerical systematics, the analyzed taxa are commonly referred to as operational taxonomic units, or operational taxonomic units (OTUs). The practical procedure begins (as in all other data Processing methods) by determining the expression of characters across all taxa in the studied group. These can then be encoded in a so-called binary data matrix, where taxa are placed opposite characters (Fig. 11.3, A). For instance, for the character "flower color," the manifestations of red or white flowers are arbitrarily coded as 0 and 1. In the next step of pairwise comparison of all taxa, similarity is calculated (or distance, taking 1 as similarity). A simple way to compute this is by dividing the number of characters shared by two taxa by the total number of observed characters (Fig. 11.3, B). The similarity coefficient calculated in this manner is also known as the simple matching coefficient. Finally, in the third step, using all similarity coefficients, taxa are grouped into clusters of progressively decreasing similarity. The phenogram presented in Fig. 11.3, C is the result of so-called cluster analysis, in which taxa are arranged in a hierarchical order. The various groups, or "branches," diverging from a branching point are also referred to as clusters.

Fig. 11.3. Numerical systematics: A — data matrix for taxa from A to F and 10 characters, each with states 0 and 1; B — pairwise similarities between taxa (ratio of the number of shared characters to the total number of characters); C — phenogram calculated from the coefficients

The method just described can be modified in many ways or used with various options. For example, a certain character (e.g., flower color) may have not two, but a greater number of states (e.g., red, white, yellow; multistate character). This variant can be computed similarly to two-state characters. When coding quantitative characters, such as leaf length, value classes must be established—for example, designating all leaves shorter than 10 cm as 0 and all leaves longer than 10 cm as 1. Such classes are generally defined arbitrarily. In principle, however, one can bypass coding such quantitative characters altogether and treat the simple difference in character expression between two taxa as a measure of similarity (distance) for that character. Besides the very simple simple matching coefficient, A large number of other similarity or distance coefficients can be employed. These make it possible, for instance, to account for whether the commonality between two taxa stems from the absence or presence of a character manifestation. Jaccard’s coefficient, for example, does not account for similarities based on absent characters. When examining DNA sequence characters, corrections are made for the increasing probability of multiple and convergent changes at a single nucleotide as the distance between two taxa increases (e.g., Kantor distance, Kimura distance). Finally, there are numerous cluster analysis techniques as well as so-called ordination analysis. For instance, Principal Component Analysis (PCA), unlike cluster analysis, provides a non-hierarchical representation of similarity relationships. Commonly applied methods—especially for molecular data—include UPGMA (unweighted pair group method using arithmetic averages) and neighbour-joining (NJ). While UPGMA yields "correct" results only when the evolutionary rate is uniform across all lineages, NJ can handle varying evolutionary rates through the continuous recalculation of the initial distance matrix.

Phenetics' claim to objectivity is limited in many respects. Like all other methods, it begins with the Selection and evaluation of characters, for which no objective criteria exist. A specific problem with phenetic methods is that numerous different approaches exist both for calculating similarity coefficients and for subsequent pairwise similarity computations, which can yield disparate results with no objective criterion for choosing among them. Consequently, a fundamental question arises: can similarity be equated with kinship, given that kinship is the sole objective criterion for grouping? Similarity and kinship are identical only when the evolutionary rate is uniform across all lineages and in the absence of parallelisms and convergences. If accelerated evolution in a single Lineage—such as the colonization of a completely novel habitat—leads to strong phenotypic divergence, similarity will inaccurately reflect genealogical relationships (Fig. 11.4). Initially, however, phenetics made no claim to reconstruct kinship.

Fig. 11.4. Similarity and kinship. When evolutionary rates vary, similarity does not reflect kinship. Although A and B are more similar than B and C, B and C are more closely related

11.1.3.4. Phylogenetic Systematics

The method of phylogenetic systematics (cladistics) was developed by the entomologist W. Hennig. The first step involves selecting all characters. The second step of the analysis consists of evaluating characters based on their relative time of appearance. Characters found exclusively within the studied group (the ingroup) (Fig. 11.5) are termed relatively advanced or apomorphic (apomorphy), whereas those already present outside the studied group are considered ancestral or plesiomorphic (plesiomorphy). Depending on whether an apomorphy is present in only a single taxon or in multiple taxa, one speaks of autapomorphy or synapomorphy, respectively, while plesiomorphy corresponds to METABOLISM/2.html">THE CONCEPT OF symplesiomorphy. Evaluating a character state as apomorphic or plesiomorphic is relative because it changes with The Scope of the studied group. For instance, possessing bisexual flowers is a plesiomorphic character for angiosperms as a whole (compared to unisexual flowers), as bisexual flowers represent the ancestral state in these plants (see 11.2. The Flower). For all seed plants, however, bisexual flowers represent an apomorphy, as this character state occurs exclusively within angiosperms and can be used to substantiate the monophyly of this group. Hennig’s crucial breakthrough was demonstrating that only apomorphic characters—not plesiomorphic ones—are valid for establishing kinship. For example, if one were to compare the autotrophic aster (Asteraceae/Asterales) with the autotrophic mullein (Scrophulariaceae/Lamiales) and the parasitic broomrape (Orobanchaceae/Lamiales) and conclude that asters and mulleins are more closely related based on their shared presence of chlorophyll, this Conclusion would be erroneous. The reason is that the presence of chlorophyll is a primitive trait of green plants and, as a plesiomorphy, cannot serve to establish kinship within a small subgroup of them. Chlorophyll has been retained in asters and mulleins, whereas it was lost in broomrape.

Fig. 11.5. Comparison with the sister group. If in the studied group of taxa A, B, and C the character states are 1-0 for A, 1-0 for B, and 1-1 for C, and the 1-0 state in sister group S is considered plesiomorphic, one must postulate only a single character transformation: 1-0 -> 1-1 in the lineage leading to C. If the initial state were instead assumed to be 1-1, two character transformations would be required: either the transformation (1-1 -> 1-0) occurred during the emergence of both A and B, or this transformation (1-1 -> 1-0) occurred in the common ancestor of A, B, and C after the emergence of S, while a reverse transformation (1-0 -> 1-1) occurred during the emergence of C. Comparison with the sister group serves as the basis for THE PRINCIPLE OF parsimony in identifying apomorphic characters

To resolve whether a character state is apomorphic or plesiomorphic, comparison with the sister group (sister group comparison) serves as a vital method. Here, it is generally assumed that the character state found in the closest relative of the studied group (the sister group) is plesiomorphic. The rationale for this assumption is that it is more parsimonious than the alternative assumption of an independent apomorphic manifestation of the character in a non-sister group¹. If, in a studied group of three taxa A, B, and C, the character states are A: 1-0, B: 1-0, and C: 1-1, and the state 1-0 in the sister group S is considered plesiomorphic, only a single character transformation from 1-0 to 1-1 is required in the lineage leading to C (Fig. 11.5). Conversely, if the ancestral state were assumed to be 1-1, two character transformations would be necessary: either the transformation from 1-1 to 1-0 occurred independently twice during the origins of A and B, or this transformation took place in the common ancestor of A, B, and C after the emergence of S, followed by a reverse transformation from 1-0 to 1-1 during the origin of C (Fig. 11.5). Because only an apomorphic character allows inferences about kinship, the identical 1-0 character state in A and B does not indicate a close genealogical relationship between these taxa; rather, A and B have simply retained the character state of their common ancestor.

¹ The authors’ explanation of the cladistic principle of parsimony is imprecise. The discussion should center on the independent appearance of identical character states in both the studied group and its sister group. — Editor's note.

Parsimony as a criterion for selecting informative characters does not imply that evolution itself must be economical or always follow the shortest path. Rather, parsimony is a general scientific principle that requires the minimum number of hypotheses necessary for explanation.

Even when the closest relative of the studied group is known, it is necessary to compare the group not only with its sister group but additionally with other related groups. The sister group together with these other relatives form the outgroup. The Concept of the outgroup is broader than that of the sister group. The challenge in comparing with sister and outgroups lies in the assumption that once-arisen characters remain unchanged within these groups. Because the sister group is geologically of the same age as the studied group, and other outgroup taxa are geologically older, they have had an equal amount (sister group) or even more time (other outgroup taxa) for evolutionary modification than the studied group itself. Therefore, it is highly unlikely that the sister or outgroup remained unchanged since the emergence of the studied group.

Once a character has been evaluated as apomorphic or plesiomorphic In the second step of the analysis, only apomorphic characters are retained in the subsequent step to establish kinship. If all instances of parallel evolution—parallelisms—as well as character reversals have been correctly identified during initial character selection, the apomorphic character states should be consistent. However, this is almost never the case. Character conflicts caused by parallelisms or reversals are termed homoplasy, and a character state is described as homoplastic if it has arisen more than once within the studied group or has reverted to its ancestral state after arising. Existing character "conflicts" are resolved by employing suitable computational methods and applying the principle of parsimony to minimize the total number of required character transformations.

This final step is the cornerstone of what is today most commonly called maximum parsimony cladism, which has largely superseded phylogenetic systematics in its original, newly described form and is implemented in computer software such as PAUP (Phylogenetic Analysis using parsimony). The principle of maximum parsimony involves abandoning the preliminary assessment of character states as apomorphic or plesiomorphic, and thus forgoing comparison with a sister or outgroup; instead, all characters are computed to minimize the number of required character transformations in the resulting cladograms. Nevertheless, the sister/outgroup is still included in the analysis. The most parsimonious cladogram is preferred, which initially lacks a starting point (an unrooted cladogram). The "ROOT" as the starting point of such a cladogram is positioned based on the results of comparing the ingroup and outgroup. Evaluating characters as apomorphies or plesiomorphies using the maximum parsimony method is performed only after the cladogram has been calculated and rooted. Various groups diverging from a branching point, or the "branches" of a cladogram, are also referred to as clades.

So-called evolutionary systematics attempts, albeit without formalized methodologies, to reflect both similarity and kinship in classification. Angiosperm classification systems still frequently employed today—such as those by A. Cronquist, R.M.T. Dahlgren, A.L. Takhtajan, and R.F. Thorne—represent precisely this evolutionary systematics.

11.1.3.5. Maximum Likelihood Method

In interpreting DNA nucleotide sequence characters, researchers increasingly apply a technique known as the maximum likelihood method.

The starting point of this method is the formulation of a sequence evolution model specific to the studied group. On the one hand, such a model accounts for sequence variations within the group, for example, by estimating substitution rates based on observed variability at each position or by inferring the relative frequencies of transitions and transversions from sequence variations. On the other hand, the model of sequence evolution incorporates additional general assumptions, such as the mutual independence of sequence positions. A Phylogenetic Tree is then calculated that explains the observed sequence variations in the studied group with maximum likelihood based on the adopted, group-specific model of sequence evolution.

11.1.3.6. Statistical Analysis of Kinship Hypotheses

Kinship hypotheses generated via phenetics, cladistics, or maximum parsimony are amenable to statistical testing of their stability to a certain degree. One such approach is bootstrap analysis. By randomly resampling from the original data across numerous iterations, a new data matrix is compiled and subsequently analyzed. This new matrix has the same dimensions as the original one but may differ in that a particular character, for instance, is used three times while two other characters are omitted entirely. A bootstrap index value of 90 for a given clade means that in 90% of replicate analyses of randomized data matrices, the placement of taxa within that clade is identical to their placement in the Analysis of the original data matrix. In cladistics, the Decay index (also known as the Bremer index) is commonly employed. This metric determines the number of additional steps (character transformations) required for a resolved clade (i.e., one depicting genealogical relationships among A, B, and C; see Fig. 11.1) to collapse into a polychotomy (where the genealogical relationships of A, B, and C can no longer be resolved). A Decay index value of 1, for example, would mean that a resolved branch (clade) in the most parsimonious phylogenetic tree becomes unresolved in a tree that is only a single step longer than the most parsimonious one. An increase in the Decay index is interpreted as an increase in the robustness of the kinship hypothesis.

11.1.4. Phylogeny and Classification

The aforementioned approaches to systematic analysis lead to hypotheses regarding the evolutionary relationships among taxa, i.e., taxonomic phylogeny. A particular challenge, inherent to both phenetic and Cladistic Methods alike, is their difficulty in resolving reticulate evolution (i.e., the emergence of a new taxon via the hybridization of diverged lineages), which is so common in plants (see 10.3.3). Neither a phenogram nor a cladogram constitutes a true pedigree or genealogical tree, as they provide no direct information about ancestors or descendants. A taxon's phylogeny can be used to trace character evolution. Simply mapping traits onto a phenogram or cladogram illustrates how a character has changed over the course of the taxon's phylogenetic development. This is exemplified by forms related to greater celandine (Chelidonium majus: Papaveraceae-Chelidonioideae) (Fig. 11.6). Phenograms or cladograms must be based on characters that are independent of those whose phylogeny is being reconstructed.

Fig. 11.6. Character evolution. Mapping traits onto a cladogram shows how a character might change during the phylogeny of a taxon. Within the Chelidonium majus alliance, multiporate pollen grains arose in the common ancestor of Sanguinaria, Eomecon, Macleaya, and Bocconia, whereas the tricolpate pollen grains present in the common ancestor of Hylomecon, Stylophorum, and Chelidonium became multiporate in the ancestor of Stylophorum and Chelidonium. Tricolpate pollen grains re-emerged with the origin of Chelidonium.

When translating phenograms and cladograms into a formal classification, one follows the widely accepted principle of phylogenetic systematics that only monophyletic groups should be recognized. Monophyly (Fig. 11.7) is defined as a case in which a taxon includes all direct descendants of a single common ancestor. Paraphyly occurs when all groups included in a taxon descend from a single common ancestor, but not all descendants of that ancestor are included within the taxon. In polyphyly, direct descendants from different ancestors are grouped into a single taxon.

Fig. 11.7. Mono-, para-, and polyphyly. A monophyletic taxon (3-4, 2-4, 1-4) contains all descendants of a single immediate common ancestor. Paraphyly is observed when all groups included in the taxon (1-2, 1-3) share a common ancestor, but not all descendants (3-4, 4) of that ancestor are included in the taxon. A polyphyletic taxon contains subgroups (1 and 3) that originated from different ancestors.

One of the many well-known examples of paraphyly is found in the two North American genera Clarkia and Heterogaura (Onagraceae). While Clarkia possesses two whorls

of fertile stamens, a deeply lobed stigma, and a capsular fruit, Heterogaura has only a single whorl of fertile stamens, a capitate stigma, and a nut-like fruit, which is why this taxon was originally described as a distinct genus. Molecular analyses have clearly demonstrated that Heterogaura is closely related to certain species of Clarkia and, therefore, evolved from within the genus Clarkia (Fig. 11.8). If Heterogaura were retained as a separate genus, Clarkia would become paraphyletic with respect to Heterogaura: although all representatives of Clarkia trace back to a single direct common ancestor, the exclusion of Heterogaura means not all descendants of that ancestor are included in a single taxon. Abandoning paraphyletic taxa implies first and foremost that classification should be based on evolutionary relationship rather than overall similarity. Consequently, Heterogaura—despite its morphological divergence from typical Clarkia species—should be classified as a representative of the genus Clarkia, because its closest relatives belong to Clarkia¹. Molecular analyses have also shown that eudicots are paraphyletic with respect to monocots (see Fig. 11.213). Nevertheless, many authors dispute the phylogenetic systematics requirement to disregard plesiomorphic characters in classification.

¹ This taxon is appropriately included in the genus Clarkia because some species of the latter are more closely related to Heterogaura than to other Clarkia species. — Note by the Editors.

Fig. 11.8. Paraphyly (A – F). Clarkia (A – C) and Heterogaura (D – F) differ in the number of fertile stamens, stigma shape, and fruit type. Phylogenetic analysis indicates that maintaining Heterogaura heterandra as an independent genus results in the paraphyly of the genus Clarkia.

11.1.5. Nomenclature

The naming and classification of plants are governed by numerous formal rules set forth in the periodically revised International Code of Nomenclature for algae, fungi, and plants. The formal aspects of systematic research are also referred to as taxonomic research or simply Taxonomy. Due to the highly ambiguous Definitions of this term in literature, we have entirely dispensed with it here. However, the terms "systematics" and "taxonomy" are commonly treated as synonyms.

Plant classification employs a hierarchy of obligatory taxonomic ranks, or categories. These represent Abstract ordering concepts that occupy strictly defined positions within the hierarchy. A species name is a so-called binomial (double name) consisting of the genus name and the specific epithet (e.g., Achillea millefolium). The full species name also includes the authority—the name of the scientist who described the species. In Achillea millefolium L., the letter L stands for Linnaeus. Names of supraspecific taxa consist of a single word (uninominal) (e.g., Achillea). Table 11.1 presents the primary taxonomic categories, their standard endings, and the corresponding names of specific taxa using yarrow (Achillea millefolium) as an example. A new taxon in botany is considered validly published when its description meets specific criteria: the selection of a name in accordance with established rules, the provision of a Latin Diagnosis, publication in a peer-reviewed or publicly accessible botanical publication, and the designation of a type specimen (see below), which must be preserved and made accessible to the scientific community. The conditions ensuring the accessibility of the publication and type to all researchers are defined in the Code. All scientific names of plants are used in a latinized form. Nouns are used for generic names and names of higher taxonomic ranks, whereas adjectives are most commonly used for specific epithets (and other infraspecific categories). The protologue in botany traditionally had to be written in Latin. Through typification, the published name of a type is definitively linked to a specific plant specimen. The type of a species or infraspecific taxon name is typically represented by a voucher specimen preserved in a herbarium. Only in exceptional cases, in accordance with historical tradition, can an illustration or a description serve as the type. Notably, the type specimen is not necessarily "typical" or average for the entire taxon. A genus is typified by a type species, and a family by a type genus. If multiple validly published names exist for a taxon, the principle of priority applies: the earliest validly published name is adopted. This rule does not apply to taxa above the family rank, and non-priority names may sometimes be conserved ("nomina conservanda"). Names published at a later date acquire the status of synonyms of the legitimate name.

In the recently much-discussed "phylogenetic nomenclature," briefly outlined in the PhyloCode, organism naming is guided exclusively by phylogenetic relationship.

Table 11.1. Hierarchical list of principal taxonomic categories, their standard endings, and the taxonomic position in the classification of yarrow (Achillea millefolium)