On quality of evidence in phylogeny reconstruction: a reply to Zrzavý's defence of the 'Ecdysozoa' hypothesis

Zrzavý’s arguments against the critical analyses of data supporting the Ecdysozoa hypothesis (Wägele et al., J. Zool. Syst. Evol. Res. 37, 211–223, 1999) are discussed. Zrzavý does not understand that the same basic principle of a priori weighting can be applied to sequence data as well as to morphological characters. Quality of evidence is the same as probability of homology, which is estimated from the number of discernible identical details. In sequences it is the number of identical nucleotides. Spectral analyses, dismissed by Zrzavý, visualize patterns of putative homologies present in alignments and also the number of positions supporting splits by chance alone. In cases in which old phylogenetic signals for a given monophylum are eroded in a gene, plesiomorphies and chance patterns will have strong influence on tree topologies and spectra. If plesiomorphies are a cause of errors, the addition of taxa that shorten internal branches is a remedy, although, in many cases such taxa may be extinct. The place of a priori estimations of data quality in a sequence of steps necessary for a phylogenetic analysis is shown. Morphological complexity is used as a proxy for a complex genetic basis and is used as a major criterion to compare characters of the Ecdysozoa and the Articulata. The details associated with the character ‘complex cuticle’ are discussed. Neither moulting nor the known components of the cuticle are novelties occurring only in Ecdysozoa. A published total evidence analysis is used to show that the number of coded characters does not necessarily reflect the quality of the data set. Zrzavý’s misunderstanding of the role of evolutionary scenarios is clarified and the importance of the use of additional biological data for plausibility arguments is explained. Plausibility arguments in favour of the Articulata hypothesis rely on facts found in functional morphology and in the fossil record. Zrzavý’s critique follows the actual mainstream but does not uncover logical mistakes or erroneous data analyses in the work of 86 . It is concluded that the Articulata hypothesis is a well‐founded alternative to the Ecdysozoa; it is based on much better morphological evidence and supported by plausibility arguments that currently do not exist for the Ecdysozoa.     

SPECIES BOUNDARIES IN THE TOXIC DINOFLAGELLATE PROROCENTRUM LIMA (DINOPHYCEAE,PROROCENTRALES), BASED ON MORPHOLOGICAL AND PHYLOGENETIC CHARACTERS1

Wild and cultured specimens of Prorocentrum lima (Ehrenb.) F. Stein from 26 widely different areas in 13 countries were examined in order to determine consistent characters for delimiting species boundaries in this taxon. The morphological characters valve shape, valve size, valve ornamentation, number and shape of valve pores, number and shape of marginal pores, and periflagellar platelets were observed using LM and SEM, and two molecular genetic regions were sequenced. We identified stable morphological characters that were consistent among wild specimens and all cultures, which were valve shape, valve ornamentation, and number and arrangement of periflagellar platelets. All cultures of P. lima identified by these characters formed a monophyletic group in phylogenetic analyses based on the two genes, which, however, included the species Prorocentrum arenarium. P. arenarium was determined to be within the range of morphological variation of P. lima, and therefore we synonymize the two taxa. Within this monophyletic group, P. lima was divided into several subclades in the all phylogenetic analyses. There were no morphological characters specifically related to any one subclade. The subclades appeared to correlate broadly to sample collection regions, suggesting that geographically separated populations may have become genetically distinct within this epi‐benthic species. We have emended species boundaries in P. lima.     

Forum – Taxonomic Stability is Ignorance

Absolute nomenclatural stability is undesirable in phylogenetic classifications because they reflect changing hypotheses of cladistic relationships. De Queiroz and Gauthier's (1990: Syst. Zool. 39, 307–322; 1992: A. Rev. Ecol. Syst. 23, 449–480; 1994: Trends Ecol. Evol. 9, 27–31) alternative to Linnaean nomenclature is concluded to provide stable names for unstable concepts. In terms of communicating either characters shared by species of a named taxon or elements (species) included in a taxon, de Queiroz and Gauthier's system is less stable than the Linnaean system. Linnaean ranks communicate limited information about inclusivity of taxa, but abandonment of ranks results in the loss of such information. As cladistic hypotheses advance, taxa named under de Queiroz and Gauthier's system can change their level of generality radically, from being part of a group to including it, without any indicative change in its spelling. The Linnaean system has been retained by taxonomists because its hierarchic ranks are logically compatible with nested sets of species, monophyletic groups, and characters. Other authors have offered conventions to increase the cladistic information content of Linnaean names or to replace them with names that convey cladistic knowledge in greater detail; de Queiroz and Gauthier sacrifice the meaning of taxon names and categorical ranks in favor of spelling stability.     

Phylogeny of Recent Canidae (Mammalia, Carnivora): relative reliability and utility of morphological and molecular datasets

Zrzavý, J. & ?i?ánková, V. (2004). Phylogeny of Recent Canidae (Mammalia, Carnivora): relative reliability and utility of morphological and molecular datasets. — Zoologica Scripta, 33, 311–333. Phylogenetic relationships within the Canidae are examined, based on three genes (cytb, COI, COII) and 188 morphological, developmental, behavioural and cytogenetic characters. Both separate and combined phylogenetic analyses were performed. To inspect the phylogenetic ‘behaviour’ of individual taxa, basic phylogenetic analysis was followed by experimental cladistic analyses based on different data‐partition combinations and taxon‐removal analyses. The following phylogeny of the Recent Canidae is preferred: (1) Urocyon is the most basal canid; (2) Vulpes is a monophyletic genus (including Fennecus and Alopex); (3) the doglike canids (DC) form a clade (=Dusicyon + Pseudalopex + Lycalopex + Cerdocyon + Atelocynus + Chrysocyon + Speothos + Lycaon + Cuon + Canis), split into two subclades, South American and Afro‐Holarctic, with uncertain position of the Chrysocyon + Speothos subclade; (4) Canis is paraphyletic due to the position of Lycaon and Cuon. Otocyon and Nyctereutes are the most problematic canid genera, causing an unresolved branching pattern of Otocyon, Vulpes, Nyctereutes and DC clades. Reclassification of the two basal species of ‘Canis’ into separate genera is proposed (Schaeffia for ‘C.’ adustus, Lupulella for ‘C.’ mesomelas). Although the morphological dataset ranked poorly in both separate and simultaneous analyses (measured by number of minimum‐length topologies, relative number of resolved nodes in the strict consensus of all minimum‐length topologies, consistency and retention indices, nodal dataset influence, and number of extra steps required by the data partition to reach the topology of the combined tree), the morphological synapomorphies represent nearly one quarter of all synapomorphies in the combined tree. Among the hidden morphological support of the combined tree the developmental and behavioural characters are conspicuously abundant.     

Phylogeny of the eureptantic nemerteans revisited

In this paper, we revisit the phylogeny of the eureptantic nemerteans. Three species ( Kameginemertes parmiornatus , Drepanophoriella histriana , and Polyschista curacaoensis ), not present in the original analyses by Härlin & Sundberg (1995) , are included, and in the light of the new results we discuss the phylogenetic taxonomy as well as biogeography of the Eureptantia . The biogeography is assessed by dispersal-vicariance analysis ( Ronquist 1997 ), and the new phylogenetic taxonomy is based on developments ( Härlin 1998b, 1999b ; Härlin & Sundberg 1998 ) of nomenclatural ideas originally presented by de Queiroz & Gauthier (1990, 1992) .     

Phylogeny in Echinocereus (Cactaceae) based on combined morphological and molecular evidence: taxonomic implications

Echinocereus is a morphologically diverse genus that includes 64 species grouped into eight taxonomic sections based on morphological traits. In previous molecular phylogenetic analyses, the relationships amongst Echinocereus species were not entirely revealed and useful characters to recognize clades were not provided. The inclusion of several sources of evidence in a phylogenetic analysis is likely to produce more supported hypotheses. Therefore, we performed a combined phylogenetic analysis with a set of 44 morphological characters and six chloroplast DNA sequences. Topologies from parsimony and Bayesian analyses were mostly congruent. However, the relationships of E. poselgeri were not consistent between analyses. A second Bayesian analysis using a long-branch extraction test resulted in a topology with the morphological position of E. poselgeri congruent with that in parsimony analysis. Parsimony and Bayesian analyses corroborated the monophyly of Echinocereus, which included eight monophyletic groups. The combined phylogeny integrated into different clades those taxa that were not determined in previous analyses and changed the relationships of some recognized clades. The clades did not recover the recent infrageneric classification. In the present study, a new sectional classification for Echinocereus is proposed based on the eight recovered clades, which is supported by a combination of morphological and molecular characters. An identification key for sections in the genus is included.     

Systematics of naviculoid diatoms (Bacillariophyta): A preliminary analysis of protoplast and frustule characters for family and order level classification

The relationships between 49 naviculoid diatoms, currently arranged in 14 families and four orders were investigated using cladistic analysis in order to test the types of characters used in diatom systematics and to assess how well the current classification reflects possible phylogenetic relationships in this group. Some of the families and orders comprise taxa with different protoplast characters, or taxa with similar protoplast arrangements are placed in separate families or orders. Therefore as both cell wall and protoplast characters were used, three analyses were undertaken; total data, protoplast data and frustule data. The analyses support the recognition of the Mastogloiales (unequivocally) and the Cymbellales (largely) but indicate that some of the familial groupings are more ambiguous. The members of the Berkeleyaceae, Berkeleya, Parlibellus and Climaconeis, were never grouped together and Achnanthes brevipes never grouped with the other monoraphid diatoms, but usually with members of the Mastogloiales (total and protoplast data). Similarly, Round et al.’s familial groupings within the Cymbellales do not emerge from our analyses. Our results support the hypothesis that monoraphid genera have arisen independently from different naviculoid diatoms, and that Achnanthes sensu stricto should be transferred to the Mastogloiales. Some of the problems associated with incomplete information and inaccurate terminology are discussed briefly.     

<Emphasis Type="Italic">Tabaroa</Emphasis>, a new genus of Leguminosae tribe Brongniartieae from Brazil

A new genus of Leguminosae in tribe Brongniartieae is proposed based on a new species endemic to the Caatinga of Bahia state. The new genus is named as Tabaroa L. P. Queiroz, G. P. Lewis & M. F. Wojc. and the new species as Tabaroa caatingicola L. P. Queiroz, G. P. Lewis & M. F. Wojc. A phylogenetic study of the Brongniartieae based on nuclear rDNA ITS and plastid matK sequences supports a closer relationship between Tabaroa and Harpalyce Moc. & Sessé than to the two South American genera Poecilanthe Benth. and Cyclolobium Benth., which are more similar morphologically. Optimisation of selected morphological characters on one of the most parsimonious trees indicates that the sessile ovary and the explosive pollen presentation are putative synapomorphies of the Tabaroa-Harpalyce clade. The genus Tabaroa may be diagnosed by the absence of peltate glandular trichomes; leaves imparipinnate, exstipellate and with opposite leaflets; flowers sessile, not resupinate, grouped in panicles; anthers apiculate; and fruit indehiscent. The only known species inhabits areas of arboreal caatinga on sandy soil in southwestern Bahia, near the boundaries of Dom Basílio and Livramento de Nossa Senhora municipalities.     

Refining molluscan characters: morphology, character coding and a phylogeny of the Caenogastropoda

Midgut morphology of gastropod molluscs has been underutilized as a resource of characters for phylogenetic analysis. The exclusion of these features reflects the inference that they will be uninformative in determining phylogenetic relationships because they are functionally correlated. In general, it has been hypothesized that the style sac form of midgut is an adaptation to microphagy and becomes secondarily simplified in taxa that have adopted a macrophagous/carnivorous habit with a corresponding increase in extracellular digestion (i.e. radular trituration, gizzards and/or foregut glands). This assumption has resulted in the formulation of adaptive scenarios concerning gastropod alimentary systems, which are mapped onto phylogenetic hypotheses derived from other characters. However, any conclusions regarding phylogenetic utility, and therefore homology, must be realized within a cladistic context. For this analysis, a multi‐organ system anatomical data set was assembled for 16 caenogastropods and two outgroups. The data matrix comprises 64 characters and includes many systems poorly represented in previous broad‐based comparative surveys, such as the alimentary and reno‐pericardial systems. In addition, several taxa were included for which no comprehensive anatomical studies have been available (Cyclophoroidea, Ptenoglossa). Phylogenetic analysis with NONA 1.6 resulted in two most‐parsimonious trees with length 188, CI = 0.53 and RI = 0.63, differing only in the placement of Prunum apicinum and Conus jaspideus. The topology of the strict consensus (Macleaniella Theodoxus((Neocyclotus Marisa)((Lampanella((Petaloconchus Strombus)(Crepidula Bithyia)))(Littorina(Neverita(Cypraea(Nitidiscala(Panarona(Prunum Conus(Ilyanassa Urosalpinx)))))))))), is largely congruent with several independent estimates based on both morphological and molecular data, supporting caenogastropod monophyly and monphyly of the Architaenioglossa, Sorbeoconcha and Neogastropoda. To evaluate phylogenetic utility of the midgut, a broad sampling of taxa was included representing a diversity of feeding modes, food preferences and alimentary morphologies. Character optimization revealed that the evolution of midgut structure is highly mosaic, cutting broadly across patterns of feeding, diet and foregut complexity, to a degree previously unappreciated. In addition, features from the foregut (subradular organ, oesophageal folding) and midgut (position of gastric shield) are broadly distributed among large clades of caenogastropods, providing critical basal synapomorphies within the group. This demonstrates that the gut is an unexploited resource for important and informative characters in higher order systematics of caenogastropods. © 2003 The Linnean Society of London, Zoological Journal of the Linnean Society, 2003, 137 , 447–554.     

Evaluating the efficacy of continuous quantitative characters for reconstructing the phylogeny of a morphologically homogeneous spider taxon (Araneae, Mygalomorphae, Antrodiaetidae, Antrodiaetus)

The use of continuous quantitative characters for phylogenetic analyses has long been contentious in the systematics literature. Recent studies argue for and against their use, but there have been relatively few attempts to evaluate whether these characters provide an accurate estimate of phylogeny, despite the fact that a number of methods have been developed to analyze these types of data for phylogenetic inference. A tree topology will be produced for a given methodology and set of characters, but little can be concluded with regards to the accuracy of phylogenetic signal without an independent evaluation of those characters. We assess the performance of continuous quantitative characters for the mygalomorph spider genus Antrodiaetus, a group that is morphologically homogeneous and one for which few discrete (morphological) characters have been observed. Phylogenetic signal contained in continuous quantitative characters is compared to an independently derived phylogeny inferred on the basis of multiple nuclear and mitochondrial gene loci. Tree topology randomizations, regression techniques, and topological tests all demonstrate that continuous quantitative characters in Antrodiaetus conflict with the phylogenetic signal contained in the gene trees. Our results show that the use of continuous quantitative characters for phylogenetic reconstruction may be inappropriate for reconstructing Antrodiaetus phylogeny and indicate that due caution should be exercised before employing this character type in the absence of other independently derived sources of characters.     

Phylogenetic systematics of the nymphaeales

A cladistic analysis was applied to reveal the phylogenetic relationships among the Nymphaeales. Seventeen out of twenty three characters in gross morphology, anatomy and palynology were analyzed, for their evolutionary polarities. From the results of the present analysis, the phylogenetic status of each genus and their relationships were clarified: 1)Nelumbo is a distinct taxon and is presumed to have originated from an ancestral stock of the Nymphaeales; 2)Ceratophyllum has a close phylogenetic relationship withCabomba; and 3) in the Nymphaeaceaesensu stricto, Nuphar and the remaining genral constitute a monophyletic group. A conclusion obtained from the present analysis was that the following three families should be recognized in the Nymphaeales; Nelumbonaceae Nymphaeaceae, and Ceratophyllaceae. The generaBrasenia andCabomba are traditionally classified in the Nymphaeaceae or in the independent family Cabombaceae. However, they should be included in the family Ceratophyllaceae.     

Molecules,morphology, missing data and the phylogenetic position of a recently extinct madtom catfish (Actinopterygii: Ictaluridae)

Taxa missing large amounts of data pose challenges that may hinder the recovery of a well‐resolved, accurate phylogeny and leave questions surrounding their phylogenetic position. Systematists commonly have to contend with one or two species in a group for which there is little or no material available suitable for recovering molecular data. It is unclear whether these taxa can be better placed using analyses based on morphological data only, or should be included in broader analyses based on both morphological and molecular data. The extinct madtom catfish Noturus trautmani is known from few specimens for which molecular data are unavailable. We included this taxon in parsimony and Bayesian analyses of relationships of madtom catfishes based on a combination of morphological and molecular data. Results indicate that using a combination of morphological and molecular data does a better job at providing a phylogenetic placement for N. trautmani than morphology alone, even though it is missing all of its molecular characters. We provide a novel hypothesis of relationships among Noturus species and recommendations for classification within the group. © 2009 The Linnean Society of London, Zoological Journal of the Linnean Society, 2009, 155 , 60–75.     

Review of the taxonomy of the Calanidae (Copepoda) and the limits to the genus Calanus

Conflicting systems of classifying the Calanidae (Brodsky, 1972; Bradford & Jillett, 1974) are analysed and a preferred system recommended. Brodsky's system, based on characters with assumed adaptive and functional significance is rejected as is his claim to have produced a phylogenetic scheme. Bradford & Jillett's system, based on a variety of characters, results in a coherent classification which aims to be neutral concerning phylogenetic relationships among genera. Species of Calanoides, Calanus, Nannocalanus, and Neocalanus are surveyed and the need for further taxonomic work is assessed.     

Cladistics and the land plants: A response to Robinson

The reconstruction of phylogenetic relationships should be based not on belief but on an explicit and logical analysis of all available characters. Hennigian phylogenetic systematics (cladistics) provides a framework for evaluating putative homologies characterizing particular hierarchical levels, determining relationships of taxa sharing congruent patterns of homologies, and constructing a classification based on this information. Fossils can and should be included in the analysis if enough of the relevant characters are preserved; this is not currently possible for early land plants because of the fragmentary fossil record. To avoid circularity, adaptive and functional considerations should be addressed only after a phylogenetic hypothesis based on patterns on patterns of shared homologies is available.     

Three new species of <Emphasis Type="Italic">Luetzelburgia</Emphasis> (<Emphasis Type="Italic">Leguminosae</Emphasis>, <Emphasis Type="Italic">Papilionoideae</Emphasis>) from the caatinga of Bahia,Brazil and an identification key to all species of the genus

Three new species of the Brazilian genus Luetzelburgia Harms (Leguminosae, Papilionoideae, Sophoreae sensu lato (Vataireoid clade)) are described and illustrated under the names Luetzelburgia harleyi D. Cardoso, L. P. Queiroz &; H. C. Lima, L. neurocarpa D. Cardoso, L. P. Queiroz &; H. C. Lima and L. purpurea D. Cardoso, L. P. Queiroz &; H. C. Lima. All three occur in the caatinga vegetation of the State of Bahia, Eastern Brazil. A discussion of the morphological characters distinguishing the new species from their close relatives, as well as an identification key to all known species of Luetzelburgia, are presented.     

TNT version 1.5, including a full implementation of phylogenetic morphometrics

Version 1.5 of the computer program TNT completely integrates landmark data into phylogenetic analysis. Landmark data consist of coordinates (in two or three dimensions) for the terminal taxa; TNT reconstructs shapes for the internal nodes such that the difference between ancestor and descendant shapes for all tree branches sums up to a minimum; this sum is used as tree score. Landmark data can be analysed alone or in combination with standard characters; all the applicable commands and options in TNT can be used transparently after reading a landmark data set. The program continues implementing all the types of analyses in former versions, including discrete and continuous characters (which can now be read at any scale, and automatically rescaled by TNT). Using algorithms described in this paper, searches for landmark data can be made tens to hundreds of times faster than it was possible before (from T to 3T times faster, where T is the number of taxa), thus making phylogenetic analysis of landmarks feasible even on standard personal computers.     

Bayesian and Maximum Likelihood Analyses of Rotifer–Acanthocephalan Relationships

Rotifera is composed of groups with unusual ultrastructural, physiological, and reproductive characters. Our ability to understand the evolution of these features is complicated by the fact that the phylogenetic relationships among the three traditional rotifer groups (Seisonidea, Monogononta, and Bdelloidea) and Acanthocephala remain unresolved. Here, I present maximum likelihood and Bayesian analyses of rotifer–acanthocephalan relationships using both the protein-coding gene hsp82 and a combined data set of hsp82 and ribosomal small subunit (SSU) DNA sequences, using nucleotide and codon based models of evolution. Statistical analysis of the phylogenetic support for any of the likely relationships among rotifer groups suggests that more than a combined hsp82 + SSU data set will be needed to resolve rotifer–acanthocephalan phylogeny with any degree of certainty.     

How many marmoset (Primates: Cebidae: Callitrichinae) genera are there? A phylogenetic analysis based on multiple morphological systems

The marmosets, tribe Callitrichini, are the most speciose clade in the subfamily Callitrichinae, containing 21 species. However, there is no consensus among molecular and morphological systematists as to how many genera should be recognized for the group. To test the morphological support for the alternative generic classifications, this study presents a comprehensive phylogenetic analysis. It is the first such analysis to include all 21 species and employ continuous and discrete osteological, pelage and tegument, karyological and vocal characters. This dataset was combined with nucleotide sequences from two mitochondrial and four nuclear regions. Separate analyses showed that, among morphological datasets, osteological characters were best at solving relationships at more inclusive levels, whilst pelage characters were most informative at the interspecific level. This suggests the presence of different transformation rates for the two character sets. When a single most parsimonious tree was obtained using the 83‐character matrix, three main clades were identified, supporting the division of the marmosets into three genera: Callithrix, Cebuella and Mico. The total evidence analysis that included an additional 3481 molecular characters corroborated most of the morphology‐based clades and also supported a three‐genus classification of the marmosets. This is the first morphological study to support an Amazonian marmoset clade (Cebuella + Mico), which is also strongly supported in exclusively molecular phylogenies, and to synonimize Callibella under Mico.     

Phylogenetic signal in bone histology of amniotes revisited

Legendre, L, Le Roy, N, Martinez‐Maza, C, Montes, L, Laurin, M & Cubo, J. (2012). Phylogenetic signal in bone histology of amniotes revisited. —Zoologica Scripta, 42, 44–53. There is currently a debate about the presence of a phylogenetic signal in bone histological data, but very few rigorous tests have fuelled the discussions on this topic. Here, we performed new analyses using a larger set of seven histological traits and including 25 taxa (nine extinct and 16 extant taxa), using three methods: the phylogenetic eigenvector regression, the tree length distribution and the regressions on distance matrices. Our results clearly show that the phylogenetic signal in our sample of bone histological characters is strong, even after correcting for multiple testing. Most characters exhibit a significant phylogenetic signal according to at least one of our three tests, with the phylogeny often explaining 20–60% of the variation in the histological characters. Thus, we conclude that the phylogenetic comparative method should be systematically used in interspecific analyses of bone histodiversity to avoid problems of non‐independence among observations.