Giant taxon‐character matrices: quality of character constructions remains critical regardless of size

Giant morphological data matrices are increasingly common in cladistic analyses of vertebrate phylogeny, reporting numbers of characters never seen or expected before. However, the concern for size is usually not followed by an equivalent, if any, concern for character construction/selection criteria. Therefore, the question of whether quantity parallels quality for such influential works remains open. Here, we provide the largest compilation known to us of character construction methods and criteria, as derived from previous studies, and from our own de novo conceptualizations. Problematic character constructions inhibit the capacity of phylogenetic analyses to recover meaningful homology hypotheses and thus accurate clade structures. Upon a revision of two of the currently largest morphological datasets used to test squamate phylogeny, more than one‐third of the almost 1000 characters analysed were classified within at least one of our categories of “types” of characters that should be avoided in cladistic investigations. These characters were removed or recoded, and the data matrices re‐analysed, resulting in substantial changes in the sister group relationships for squamates, as compared to the original studies. Our results urge caution against certain types of character choices and constructions, also providing a methodological basis upon which problematic characters might be avoided.     

Meta-analysis of character utility and phylogenetic information content in cladistic studies of ammonoids

While previous workers have argued persuasively that ammonoid workers should use cladistic approaches to reconstruct phylogeny, relatively few cladistic studies have been published to date. An essential yet challenging part of cladistic analysis is the selection of characters. Are certain types of characters more likely to show homoplasy? Are certain aspects of shell anatomy more likely to contain phylogenetically informative characters? Are datasets with more characters inherently better? To answer these questions, a meta-analysis of character data from published ammonoid phylogenies was performed. I compiled 14 datasets, published between 1989 and 2007, representing parsimony-based phylogenetic analyses of ammonoids. These studies defined a combined total of 323 characters, which were grouped into categories reflecting different aspects of anatomy: shell size and shape, ornament, suture, early ontogeny, body chamber and apertural modifications. Tree searches were re-run to determine overall tree statistics, parsimony permutation tail probability (PTP) tests were calculated to assess the phylogenetic information content of the matrices, and retention and rescaled consistency indices for each character were calculated. My analyses revealed that studies with higher character/taxon ratios did not necessarily produce trees with more information content and less homoplasy, as measured by retention or rescaled consistency indices, because additional characters were often parsimony-uninformative. Rather, studies with relatively few characters could produce high-quality trees if the characters were well-chosen and character states carefully defined. Characters related to the body chamber and adult aperture typically had retention indices of either 0 or 1, rarely in between, indicating that they either worked perfectly or not at all. Suture characters tended to have higher indices than shell shape or ornament characters, suggesting more phylogenetic information and less homoplasy in the suture line than in shell traits. These results should aid in the selection of characters for future cladistic studies of ammonoids.     

Misinformative characters and phylogeny shape

The discrepancy between theoretical and observed distributions of tree shapes in recent surveys of phylogeny estimates has lead to investigations of possible biological and methodological causes. I investigated whether the phylogenetic quality of characters is related to the tree shape on which they evolve. Simulated evolution revealed shape-related tendencies for characters to indicate correct cladistic relationships; these differences were measured by examining the characters directly, without deriving any phylogeny estimates. Tree stemminess indices correlated strongly with character quality when characters evolved either speciationally or phyletically. Tree balance was a significant correlate of character quality under speciational evolution but not under phyletic evolution. These results help explain the findings of other simulation studies. With additional study of the behavior of evolving characters and their interaction with phylogenetic methods, we might be able to increase the accuracy of tree estimation and compensate for potential biases related to shape. These results give further reason for caution in trusting phylogeny estimates.     

Evolution and character congruence in some western Indian Ocean Phelsuma: Numerical analysis of biochemistry, shape and scalation

The evolution of five island populations of Green gecko, representing inter- and intra-specific divergence, was studied using biochemical data, scalation and shape. The data were numerically analysed using ordination analyses for the phenetic classification and Wagner trees to hypothesize the phylogeny. These studies revealed three phenetic groups corresponding to three mono-phyletic lineages. The numerical analysis of morphological data agreed with the numerical analysis of biochemical data. It is concluded that the classification based on biochemical affinities differed from the previous classification based on conventional analysis of morphology due to methodological and philosophical differences rather than differences between morphological and biochemical evolution.
The ordination analyses were very congruent between data sets (biochemical, shape, scalation, total) and the Wagner trees were generally congruent between data sets. Some Wagner trees based on scalation data were incongruent. The phenetic and cladistic classifications corresponded to each other but differed from the conventional classification. The phylogenetic analysis of the total data set indicated that the three specific lineages showed relatively equal anagenesis. However, anagenic divergence differed markedly between character types. It is suggested that a range of character types be used when studying anagenesis.     

Root causes of phylogenetic incongruence observed within basal sauropodomorph interrelationships

The relationships among basal sauropodomorphs are controversial. Results of cladistic analyses vary from a fully paraphyletic assemblage to a monophyletic core‐prosauropod. We apply the comparative cladistics method to three published cladistic analyses of sauropodomorph dinosaurs, in order to identify root causes for differences between phylogenetic results. Except for three taxa (Saturnalia, Thecodontosaurus, and Efraasia) and one clade (Gravisauria), the remaining genera are recovered with conflicting positions. The comparative method is based on indices that allow for the quantification of the degree of similarity in characters and character states among analyses. A comparison of primary data, character selection, and scoring highlights significant discrepancies in data sets. Our results suggest that one character out of two varies from one analysis to the other. These are the root causes for the phylogenetic incongruence observed. The hurdle of the phylogenetic definition of the clade Sauropoda, which has been defined in four different ways, is also treated. We concur with several recent papers following the first node‐based definition of Sauropoda. © 2015 The Linnean Society of London     

How much can cladistics tell us about early hominid relationships?

Although cladistic analysis has been used to compare hypotheses of relationships among early hominids, the outcomes of different studies have depended entirely on the assumptions made by different investigators. Problems include the close genetic relationship of early hominid taxa, small fossil sample sizes, possible correlations among characters, and a lack of understanding about the evolutionary factors affecting characters. This study investigates the interaction of some of these problems affecting early hominid phylogenetics. Monte Carlo simulations of character state evolution in closely related taxa demonstrate that the sample sizes and close genetic relationships of early hominids do not permit cladistic analyses to obtain unequivocal results. Even with unrealistically good assumptions about the evolutionary dynamics affecting characters, the probability of the most parsimonious hypothesis being true is unacceptably small. In the face of these problems, even phylogenetic statements that are supported by a strong consensus of cladistic studies may nevertheless be in error, and such errors are likely to confound the placement of new specimens and taxa. Advancement in our knowledge of hominid phylogeny can depend only on a fuller understanding of the natural history and evolutionary dynamics of traits.     

Comparative cladistics

Current strategies to compare or synthesize morphology-based cladistic hypotheses do not empower individual cladists to (i) understand the origin, authorship, or structure of character data, (ii) efficiently locate and collate previously published character data, or (iii) effectively compare character data from competing cladistic hypotheses. This paper outlines the requisite terminology, methods and indices to effectively compile and compare morphological character data between competing cladistic hypotheses and to isolate and measure the most important factors behind differing cladistic results—character selection and character-state scoring. When the procedures outlined here are facilitated by appropriate software, morphology-based cladistics may overcome long-recognized limitations in data comparison and synthesis.     

A methodological investigation of hominoid craniodental morphology and phylogenetics

The evolutionary relationships of extant great apes and humans have been largely resolved by molecular studies, yet morphology-based phylogenetic analyses continue to provide conflicting results. In order to further investigate this discrepancy we present bootstrap clade support of morphological data based on two quantitative datasets, one dataset consisting of linear measurements of the whole skull from 5 hominoid genera and the second dataset consisting of 3D landmark data from the temporal bone of 5 hominoid genera, including 11 sub-species. Using similar protocols for both datasets, we were able to 1) compare distance-based phylogenetic methods to cladistic parsimony of quantitative data converted into discrete character states, 2) vary outgroup choice to observe its effect on phylogenetic inference, and 3) analyse male and female data separately to observe the effect of sexual dimorphism on phylogenies. Phylogenetic analysis was sensitive to methodological decisions, particularly outgroup selection, where designation of Pongo as an outgroup and removal of Hylobates resulted in greater congruence with the proposed molecular phylogeny. The performance of distance-based methods also justifies their use in phylogenetic analysis of morphological data. It is clear from our analyses that hominoid phylogenetics ought not to be used as an example of conflict between the morphological and molecular, but as an example of how outgroup and methodological choices can affect the outcome of phylogenetic analysis.     

Inferring and testing hypotheses of cladistic character dependence by using character compatibility

The notion that two characters evolve independently is of interest for two reasons. First, theories of biological integration often predict that change in one character requires complementary change in another. Second, character independence is a basic assumption of most phylogenetic inference methods, and dependent characters might confound attempts at phylogenetic inference. Previously proposed tests of correlated character evolution require a model phylogeny and therefore assume that nonphylogenetic correlation has a negligible effect on initial tree construction. This paper develops "tree-free" methods for testing the independence of cladistic characters. These methods can test the character independence model as a hypothesis before phylogeny reconstruction, or can be used simply to test for correlated evolution. We first develop an approach for visualizing suites of correlated characters by using character compatibility. Two characters are compatible if they can be used to construct a tree without homoplasy. The approach is based on the examination of mutual compatibilities between characters. The number of times two characters i and j share compatibility with a third character is calculated, and a pairwise shared compatibility matrix is constructed. From this matrix, an association matrix analogous to a dissimilarity matrix is derived. Eigenvector analyses of this association matrix reveal suites of characters with similar compatibility patterns. A priori character subsets can be tested for significant correlation on these axes. Monte Carlo tests are performed to determine the expected distribution of mutual compatibilities, given various criteria from the original data set. These simulated distributions are then used to test whether the observed amounts of nonphylogenetic correlation in character suites can be attributed to chance alone. We have applied these methods to published morphological data for caecilian amphibians. The analyses corroborate instances of dependent evolution hypothesized by previous workers and also identify novel partitions. Phylogenetic analysis is performed after reducing correlated suites to single characters. The resulting cladogram has greater topological resolution and implies appreciably less change among the remaining characters than does a tree derived from the raw data matrix.     

THE HOLY GRAIL OF THE PERFECT CHARACTER: THE CLADISTIC TREATMENT OF MORPHOMETRIC DATA

Abstract— Data scored for cladistic analyses may be quantitative or qualitative, continuous or discrete, and show overlapping or non-overlapping values between taxa. Quantitative and qualitative are modes of expression of data, while continuous or discrete refer to properties of the set of numbers that express the data; both these pairs of terms have been confused with overlapping and non-overlapping. The degree of overlap of values between taxa is often used to filter characters in cladistic analyses: if a minimum amount of overlap is exceeded, or a minimum amount of disjunction not reached, characters are rejected as "not cladistic". However, this rests on a confusion between features of taxa and features of individual organisms (attributes). Cladistic characters are features of taxa, and comprise frequency distributions of attribute values over individuals of a taxon. Cladistic characters logically cannot overlap, although taxa may have overlapping attribute values. Thus, a priori rejection of characters that have overlapping attribute values is non-sensical. Such data may still be rejected from consideration for cladistic analysis if it could be demonstrated that they contain little recoverable phylogenetic signal. Few published analyses have empirically tested this. An analysis of overlapping morphometric data from three series of Banksia suggests that, at least in these cases, they map phylogeny almost as accurately as more conventional, qualitative morphological data. While more such tests are required, morphometric data should not be rejected a priori from cladistic analyses.     

Evolutionary patterns in early tetrapods. I. Rapid initial diversification followed by decrease in rates of character change

Although numerous studies have examined morphological diversification during major radiations of marine taxa, much less attention has been paid to terrestrial radiations. Here, we examine rates of character change over phylogeny and over time for Palaeozoic limbed tetrapods. Palaeozoic tetrapods show significant decreases in rates of character change whether the rate is measured per sampled cladistic branch or per million years along phylogeny. Given changes per branch, rates decrease significantly from the Devonian through the Pennsylvanian, but not from the Pennsylvanian through the Permian. Given changes per million years, rates decrease significantly over each boundary, although the decrease is least significant over the Pennsylvanian-Permian boundary. Decreasing rates per million years through the Permian might be an artefact of the method being able to ascribe longer durations to Permian branches than to Carboniferous ones; however, it is difficult to ascribe the general pattern of decreasing rates of change over time to sampling biases or methodological biases. Thus, the results implicate biological explanations for this pattern.     

PRELIMINARY INFERENCES OF THE PHYLOGENY OF BRYOPHYTES FROM NUCLEAR-ENCODED RIBOSOMAL RNA SEQUENCES

Ribosomal RNA sequences and cladistic analysis were used to infer a phylogeny for eight bryophyte taxa. Portions of the cytoplasmic large (26S-like) and small (18S-like) subunit ribosomal RNA genes were sequenced for three marchantioid liverworts (Asterella, Conocephalum, and Riccia), three mosses (Atrichum, Fissidens, and Plagiomnium), and two hornworts (Phaeoceros and Notothylas). Cladistic analysis of these data suggests that the hornworts are the sister group to the mosses, the mosses and hornworts form a clade that is sister to the tracheophytes, and the liverworts form a clade sister to the other land plants. These results differ from previous cladistic analyses based on morphology, ultrastructure, and biochemistry, wherein the mosses alone are sister group to the tracheophytes. We conclude that cladistic analysis of molecular data can provide an independent data set for the study of bryophyte phylogeny, but the differences between the molecular and morphological results are a topic for further investigation.     

HYBRIDS AND PHYLOGENETIC SYSTEMATICS I. PATTERNS OF CHARACTER EXPRESSION IN HYBRIDS AND THEIR IMPLICATIONS FOR CLADISTIC ANALYSIS

Interspecific hybridization is considered common among plants, but the methods of cladistic systematics produce only divergently branching phylogenetic hypotheses and thus cannot give the correct phylogeny if an analysis includes hybrids. Empirical studies of the impact of known hybrids on phylogenetic analysis are lacking, and are necessary to begin to understand the problems that we face if hybrids are often included in cladistic analysis. Examination of the implications of hybrids for cladistics must begin with patterns of character expression in hybrids. This study includes 17 hybrids and their nine parental taxa that are Central American species of Aphelandra (Acanthaceae), analyzed using a set of 50 morphological characters. The hybrids are overwhelmingly intermediate as quantitatively scored for phylogenetic analysis. They express maternal and paternal, and primitive and derived characters in equal frequencies, showing no evidence of predominant inheritance of derived character states as has been assumed by most cladists who have considered hybrids theoretically. Because of their known genetic constitution, hybrids were useful in homology assessment and ordering character states. The parental character set was generally robust, but some changes were made to reflect the special evidence offered by the hybrids. These hybrids suggest that the inclusion of hybrids in phylogenetic analysis will not lead to unresolved cladograms with rampant homoplasy, as has been predicted by other authors. Instead, the patterns of character inheritance in these hybrids lead to the prediction that a hybrid will be placed by phylogenetic analysis as a basal lineage to the clade that includes its most derived parent, with relatively little effect on homoplasy. These predictions will be evaluated by incorporation of the hybrids in phylogenetic analyses, to be reported in a subsequent paper.     

Phylogenetic systematics and biomecha cs in ecomorphology

Synopsis Research in all fields of biology increasingly uses phylogenetic systematics to interpret biological data in an evolutionary context. It is becoming widely accepted that comparative studies of the correlation of biological features, such as ecomorphological studies, must frame their analyses within the context of a phylogenetic hierarchy rather than treating each taxonomic unit as an independent replicate. Recent methods for the interpretation of ecological and functional data in the framework of a phylogeny can reveal the degree to which ecomorphological characters are correlated with one another, and are congruent with hierarchical cladistic groups. An example of the ecomorphology of labrid fishes is used here to illustrate the application of several of these methods. The structural design and mechanics of the jaws of labrids are tested for ecomorphological associations with the natural diets of these fishes. Methods for analysis of the correlated evolution of both discrete and continuous quantitative characters within a phylogeny are practiced on a single ecomorphological data set. Techniques used include character coding, character mapping, phylogenetic autocorrelation, independent contrasts, and squared change parsimony. These approaches to diverse biological data allow the study of ecomorphology to account for patterns of phylogenetic ancestry. Biomechanics or functional morphology also plays a vital role in the determination of ecomorphological relationships by clarifying the mechanisms by which morphologies can perform behaviors important to the organism's ecology. The synthesis of systematics with biomechanics is an example of interdisciplinary study in which information exchange can elucidate patterns of evolution in ecomorphology.     

The Role of Functional Analysis in Phylogenetic Inference: Examples from the History of the Xiphosura

SYNOPSIS. Conventional cladistic analyses of phylogeny can beinterpreted as operating at the level of phylogenetic trees.They assume that all "evolutionary steps" (transitions fromone character state to the next, along a morphocline) are independentand equal, and, on that basis, select the cladogram which isconsistent with the most parsimonious trees. Evaluation of theassumptions of independence and equality requires considerationof hypotheses at the levelof scenarios. In some cases, argumentsbased on functional analysis can suggest revised interpretationsof either homology or polarity. If properly formulated, thesearguments can alter the evaluation of parsimony for trees tothe extent that even the choice of cladogram is affected. Thestructure of scenario level arguments is identical to that ofarguments operating at tree level. Examples of phylogeneticinference in the context of xiphosurans (horseshoe crabs), usingboth comparative morphological and functional analysis, illustratethis approach. In different cases, orthodox interpretationsof relationship are either challenged or corroborated. Althoughthe introduction of functional analysis into the process ofphylogenetic inference may appear to compromise the usefulnessof the reconstructed phylogeny for testing hypotheses concerningthe role of natural selection in evolution, it actually increasesthe strength of such tests.     

Circularity and Independence in Phylogenetic Tests of Ecological Hypotheses

It has been asserted that in order to avoid circularity in phylogenetic tests of ecological hypotheses, one must exclude from the cladistic analysis any characters that might be correlated with that hypothesis. The argument assumes that selective correlation leads to lack of independence among characters and may thus bias the analysis. This argument conflates the idea of independence between the ecological hypothesis and the phylogeny with independence among characters used to construct the tree. We argue that adaptation or selection does not necessarily result in the non-independence of characters, and that characters for a cladistic analysis should be evaluated as homology statements rather than functional ones. As with any partitioning of data, character exclusion may lead to weaker phylogenetic hypotheses, and the practice of mapping characters onto a tree, rather than including them in the analysis, should be avoided. Examples from pollination biology are used to illustrate some of the theoretical and practical problems inherent in character exclusion.     

Parallelism,parsimony, and the phytogeny of the lemuridae

In spite of the increasing popularity of cladistic methods in studies of primate systematics, few authors have investigated the effects of parallel evolution when such methods are applied to empirical data. To counter the effects of parallelism, cladistic techniques rely on the principle of evolutionary parsimony. When parsimony procedures are used to reconstruct the phylogeny of the Lemuridae, nine highly parsimonious phylogenies can be deduced. Further choice among these competing hypotheses of relationship is determined by the extent to which one embraces the parsimony principle. The phylogeny obtained by the most rigorous adherence to the parsimony principle is one which is wholly consistent with traditional evolutionary classifications of the Lemuridae. Moderate levels of parallelism can lead to the generation of several plausible, alternative phylogenetic hypotheses; less than 25% of the characters analyzed here need have evolved in parallel, yet they are largely responsible for the ambiguity of the nine different lemurid phylogenies. This suggests that phylogeny reconstructions based entirely on cladistic methods do not provide a suitable basis for the construction of classifications for groups such as the order Primates, where the degree of parallelism is likely to be quite high.     

COULD A CLADOGRAM THIS SHORT HAVE ARISEN BY CHANCE ALONE?: ON PERMUTATION TESTS FOR CLADISTIC STRUCTURE

Abstract Absolute criteria for evaluating cladistic analyses are useful, not only because cladistic algorithms impose structure, but also because applications of cladistic results demand some assessment of the degree of corroboration of the cladogram. Here, a means of quantitative evaluation is presented based on tree length. The length of the most-parsimonious tree reflects the degree to which the observed characters co-vary such that a single tree topology can explain shared character states among the taxa. This “cladistic covariation” can be quantified by comparing the length of the most parsimonious tree for the observed data set to that found for data sets with random covariation of characters. A random data set is defined as one in which the original number of characters and their character states are maintained, but for each character, the states are randomly reassigned to the taxa. The cladistic permutation tail probability, PTP, is defined as the estimate of the proportion of times that a tree can be found as short or shorter than the original tree. Significant cladistic covariation exists if the PTP is less than a prescribed value, for example, 0.05. In case studies based on molecular and morphological data sets, application of the PTP shows that:

• 1 In the comparison of four different molecular data sets for orders of mammals, the sequence data set for alpha hemoglobin does not have significant cladistic covariation, while that for alpha crystallin is highly significant. However, when each data set was reduced to the 11 common taxa in order to standardize comparison, reduced levels of cladistic covariation, with no clear superiority of the alpha crystallin data, were found. Morphological data for these 11 taxa had a highly significant PTP, producing a tree roughly congruent with those for the three molecular sets with marginal or significant PTP values. Merging of all data sets, with the exclusion of the poorly structured alpha hemoglobin data, produced a data set with a significant PTP, and provides an estimate of the phylogenetic relationships among these 11 orders of mammals.
• 2 In an analysis of lactalbumin and lysozyme DNA sequence data for four taxa, purine-pyrimidine coding yields a data set with significant cladistic covariation, while other codings fail. The data for codon position 3 taken alone exhibit the strongest cladistic covariation.
• 3 A data set based on flavonoids in taxa of Polygonum initially yields a significant PTP; however, deletion of identically scored taxa leaves no significant cladistic covariation.
• 4 For mitochondrial DNA data on population genome types for four species of the crested newt, there is significant cladistic covariation for the set of all genome types, and among the five mtDNA genome types within one of the species. However, a conditional PTP test that assumes species monophyly shows that no significant cladistic covariation exists among the fur species for these data.
• 5 In an application of the test to a group of freshwater insects, as preliminary to biological monitoring, individual subsets of the taxonomic data representing larval, pupal, and adult stages had non-significant PTPs, while the complete data set showed significant cladistic structure.

     

Angiosperm wood evolution and the potential contribution of paleontological data

Wood anatomy is often viewed as a source of independent data that may be used to assess evolutionary relationships among angiosperms. Comparative anatomical studies document suites of correlated characters that have been interpreted as general evolutionary trends, of which several have been asserted to be irreversible. Paleobotanical data summarized by Wheeler and Baas provide broad chronological corroboration of some wood anatomical trends, such as evolution from scalariform to simple perforation plates and long to short vessel elements. However, the focus on general evolutionary trends rather than on analyzing character distribution patterns in a cladistic phylogenetic context obscures a more detailed understanding of the evolution of wood anatomical features. Patterns of character evolution, including the assertions of irreversibility, need to be tested through cladistic analyses. In this paper selected wood anatomical features from families of Magnoliidae and “lower” Hamamelididae are summarized and mapped onto previously published cladograms as a preliminary means of testing previous hypotheses of wood evolution. The results show that many of the characters are homoplasious and have evolved both in accord with, and counter to, the hypothesized general trends in different groups of flowering plants. In general, changes that confirm generalized trends are more common than changes that are counter to those trends. Future studies should combine wood anatomical characters with other features as part of a cladistic analysis. Fossil woods have not yet contributed significantly to phylogenetic studies, but in the very few cases where they have been linked to fossil reproductive structures, the woods have provided a better understanding of wood anatomy in early members of some families. Data from fossil wood expand the diversity of anatomical structure known in some angiosperm taxa and thus provide additional evidence that might be used in phylogenetic analyses. Fossil woods have the greatest potential to affect phylogenetic analyses where they can be linked to other fossil organs. The best chance for establishing such a linkage is through the study of fossil charcoalified woods that co-occur with other dispersed mesofossils.