Grebes and flamingos: standards of evidence, adjudication of disputes, and societal politics in avian systematics
2010; Wiley; Volume: 27; Issue: 4 Linguagem: Inglês
10.1111/j.1096-0031.2010.00339.x
ISSN1096-0031
Autores Tópico(s)Genetic diversity and population structure
ResumoCladisticsVolume 27, Issue 4 p. 391-401 Free Access Grebes and flamingos: standards of evidence, adjudication of disputes, and societal politics in avian systematics Bradley C. Livezey, Corresponding Author Bradley C. Livezey Corresponding author:E-mail address:livezeyb@carnegiemnh.orgSearch for more papers by this author Bradley C. Livezey, Corresponding Author Bradley C. Livezey Corresponding author:E-mail address:livezeyb@carnegiemnh.orgSearch for more papers by this author First published: 07 October 2010 https://doi.org/10.1111/j.1096-0031.2010.00339.xCitations: 1AboutSectionsPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL Share a linkShare onFacebookTwitterLinked InRedditWechat Abstract The recent proposal of a sister-group relationship between the Neoavian grebes (Podicipedidae) and flamingos (Phoenicopteridae) is chronicled, and morphological evidence claimed to be supportive of the grouping is examined. The hypothesis arose from an exiguous amalgam of molecular inferences, advanced in part by a pervasive, unsupported superiority conferred upon sequence data, and adopted by several societal committees on avian classification. Morphological characters marshalled specifically to support the hypothesis were found to be erroneous, and associated phylogenetic analyses, where given, were ambiguous. A combined analysis of large data sets for morphology and RAG-1 sequences found flamingos and storks to be sister groups but with reduced support. This example illustrates problems attending the synthesis of contradictory evidence and evaluation of unprecedented hypotheses, and reveals the informality by which revisions are adopted. Procedures for rational synthesis of evidence are needed for progress during this challenging but promising period of diversified phylogenetics, without which disputes will be dominated increasingly by polarized, intransigent prejudice regarding methods and data.© The Willi Hennig Society 2010. The current spate of papers treating phylogenetic methods and resultant inferences is both welcome and stimulating, but the tide of diverse phylogenetic analyses appears to have outpaced logical criteria for genuine consensus. Symptomatic of this disparity are the critiques and responses that follow the appearance of most phylogenetic analyses of scale or potential importance (e.g. Frost et al., 2008; Braun and Huddleston, 2010). Where significant differences arise among authors, methods, or data sets, it unfortunately remains that neither means for objective reconciliation of disputes nor empirically grounded criteria for comparative evaluation of disparate inferences exists. Although congruence among studies is the hope of diverse systematists, most apparent consensuses principally reflect predispositions among investigators as much as genuine empirical congruence. The debates of current systematics, while making for lively reading, can be acrimonious despite congruence (Braun and Huddleston, 2010; Mayr and Manegold, 2010), and often result in poor reconciliation. The recently proposed relationship of the cosmopolitan grebes (Podicipedidae: five genera, 22 modern species) and pantropical flamingos (Phoenicopteridae: three genera, five modern species), in part because of its prima faciae implausibility, epitomizes these and related issues of morphological systematics (cf. Assis, 2009). Here we review the origin and promotion of the recent, counterintuitive hypothesis regarding grebes and flamingos, and examine the evidence (especially morphological) regarding the grouping. The sociological and philosophical dimensions of this publicized dispute are reviewed, both as a reflection of inferences drawn where evidence is equivocal and with respect to the classificatory actions that proceed therefrom. My aim is to raise awareness of problems bearing on such collisions of evidence, and to attempt a cladistic synthesis of divergent classes of signal with respect to this particular case. Historical overview Perceptions of relationships The systematic affinities of grebes have been controversial, although a majority of ornithologists for almost two centuries have allied them with loons (Linnaeus, 1758; Nitzsch, 1840; Sundevall, 1872; Garrod, 1873, 1874; Fürbringer, 1888; Gadow, 1892; Howard, 1950). A number of other authorities dismissed the many traits shared by grebes and loons as convergences (Sibley and Ahlquist, 1972, 1991), but typically did not propose a viable alternative relationship. Storer (1960, 1971), strongly influenced by the locomotory descriptions by Stolpe (1932, 1935), joined in the suspicions of functional convergence to dismiss morphological evidence of a close relationship between grebes and loons (e.g. Storer, 1971; Cracraft, 1981, 1982; Cracraft et al., 2003). Sibley and Ahlquist (1991, p. 244) shared suspicions of convergence regarding grebes and loons, stating: "Whatever may be the virtues of cladistic analysis, Cracraft (1981) classification indicates that it does not consistently separate homology from analogy, as indicated by his placing the loons and grebes together ….". In an addendum, Sibley (1994, p. 90) reaffirmed his support for the ciconiiform affinity of flamingos, and while admitting similarities between grebes and loons, stated that the grebes "… have no close living relatives …". The phylogenetic position of flamingos also has been subject to considerable debate (Sibley et al., 1969), with a majority of early workers (Gadow, 1877, 1879, 1892; Reichenow, 1882) placing flamingos as closely allied with the storks (Ciconiidae). Others, principally on the basis of webbed feet and fundamentally distinct rostral lamellae (Jenkin, 1957), associated flamingos with waterfowl (Anseriformes), whereas some authors vaguely associated flamingos with both Ciconiiformes and Anseriformes (Huxley, 1867; Fürbringer, 1888; Wetmore, 1951). During the late twentieth century, several fossils (Juncitarsus, Palaeodus, Elornis, and Presbyornis) prompted the proposal of several possible allies of flamingos, including Anseriformes, Charadriiformes, and Gruiformes (Feduccia, 1977, 1978; Olson, 1979, 1985; Olson and Feduccia, 1980a,b; Peters, 1987; Ericson, 1999). Plausibility of the latter was diminished further by Livezey and Zusi (2006, 2007) and by species-level analyses of allegedly related orders (Livezey, 1997, 1998, 2009, 2010a). Emergence of a novel hypothesis Van Tuinen et al. (2001) used DNA hybridization in the first analysis to suggest a close affinity between grebes and flamingos, despite the general recognition of DNA hybridization as unreliable for phylogenetic inference (Cracraft, 1987, 1992; Houde, 1987; Mindell, 1992). Citations of Sibley and Ahlquist (1991) as consistent with the pairing (e.g. Fjeldså, 2004) were mistaken; Sibley and Ahlquist (1991, Fig. 356) clustered (upgma) grebes with a narrowly delimited Pelecaniformes, and flamingos with a controversially constituted Ciconiiformes. Subsequent molecular analyses varied regarding this grouping (Cracraft et al., 2004): some allied grebes with flamingos (cf. Cracraft et al., 2004; Hackett et al., 2008), whereas others were ambiguous or contradictory on this point (e.g. Mindell et al., 1997; Van Tuinen et al., 2000; Cracraft, 2001; Cracraft et al., 2004; Fain and Houde, 2004; Slack et al., 2006). Mayr and Clarke (2003) undertook a phenotypic analysis of deep (ancient) divergences of Neornithes (43 taxa, 148 characters), in which grebes were joined with loons (Gaviidae) unless the loons were excluded from analysis. Mayr (2004) followed this with a narrower cladistic analysis (17 families, 70 characters), in which the Gaviidae were inferred to be the sister group of the grebe–flamingo couplet. Mayr (2004) repeated anecdotal suspicions of morphological convergence between grebes and loons and composed evolutionary scenarios for fossils consistent with a union of grebes with flamingos. Manegold (2006) joined with the proposal of two additional features uniting grebes and flamingos, although he did not perform a formal analysis. Mayr (2004, p. 240) concluded, and reasserted (Mayr, 2007b), the couplet to be "… one of the best supported higher-level clades within modern birds …". Fjeldså (2004, p. 19) enthusiastically endorsed the new proposal in a section entitled: "Why taxonomists went wrong in the past". Fjeldså (2004) also speculated upon possible evolutionary convergence and tallied differences between groups, methods common during the phenetic era (Sibley and Ahlquist, 1972, 1991; Olson and Feduccia, 1980a,b). The numerous differences between grebes and flamingos seem to have escaped such scrutiny, a tally of which would be impressive but, like those of similarities, inconclusive without an analysis. Storer (2006) contested the recently proposed kinship between flamingos and grebes by similar descriptive means. Mayr (2007b) responded by using ad hoc assessments of characters, a mode of argumentation for which he had chastised Storer (2006). In a morphological synthesis of higher-order birds and related non-avian Theropoda (150 neornithine taxa; 2954 phenotypic characters), Livezey and Zusi (2006, 2007) found strong support for grebes and loons as sister groups. Mayr (2004, 2005, 2006, 2007a, 2009), however, disputed this finding. Furthermore, (Mayr, 2005; p. 523) attempted to bolster the support for a grebe–flamingo linkage by evolutionary scenarios and reassessments of fossils possibly related to one or both groups, evidence explicitly evaluated under the "new light" of this hypothesis. Finally, Mayr (2008, p. 67) concluded that the tandem exemplified "…when morphology, molecules and fossils coincide". Covevolutionary patterns Hopkins (1942) considered patterns of feather lice (Mallophaga) to be supportive of an alliance between flamingos and Anseriformes (Seebohm, 1890; Shufeldt, 1889, 1901). Storer (2002), however, discerned no clues to the relationships of grebes from a study of parasitic Metazoa of the group. Johnson et al. (2005) reinterpreted the data of Hopkins (1942) and inferred accordance with the newly proposed affinity between flamingos and grebes. Unfortunately, ectoparasites are subject to extralineal transfer by association among host taxa of similar habitat (Mayr, 1957; Stresemann, 1959), and Johnson et al. (2005) confirmed the ambiguity of such data and the influence of the presumed direction of host-switching—flamingos to ducks or the reverse. Current status of evidence Molecular evidence In addition to the contradictory findings among molecular works regarding this proposal, a suite of more subtle problems attend the molecular evidence in this case. The molecular case for a sister-group relationship between grebes and flamingos courts the artefactual affinity in that both taxa manifest higher-level, apomorphic divergence, and includes a significant redundancy of included sequences and related interdependency among studies (Van Tuinen et al., 2001; Cracraft et al., 2004; Ericson et al., 2006; Brown et al., 2008; Morgan-Richards et al., 2008; Pratt et al., 2008). The repeatedly employed mitochondrial genes are typified by evolutionary rates that are unsuitably high for reliable reconstruction of ancient nodes (Stanley and Cracraft, 2002; Cracraft et al., 2004), and appear to be prone to greater homoplasy than morphological characters (Griffiths et al., 2004). Sequence analyses allowing the couplet also are compromised by exclusion of key taxa (Van Tuinen et al., 2001; Morgan-Richards et al., 2008; Pratt et al., 2008). Finally, increasingly employed Bayesian methods are prone to overestimation of support statistics (Cummings et al., 2003; Erixon et al., 2003; Simmons et al., 2004) and problems of unrealistic prior distributions (Pickett and Randle, 2005; Steel and Pickett, 2006). Global solutions are not merely conglomerates of analytically independent two-taxon terminals, but global optimizations of n-taxon problems. The interdependence of constituents of a global optimum is readily substantiated by the deletion of taxa from analyses and the impact of jackknifing, and these in turn are affected by alignments and the evolutionary models employed. Therefore, restriction of comparisons to a single contentious couplet (e.g. grebe–flamingo node) ignores the analytical properties and realism of the global topology of which the couplet is a part. Moreover, the early divergences among Neoaves remain weakly resolved in some molecular studies (e.g. Gibb et al., 2007), and several works placed the grebe–flamingo couplet within a large, often basal, eclectic polytomy (Van Tuinen et al., 2001; Chubb, 2004; Cracraft et al., 2004; Ericson et al., 2006; Hackett et al., 2008). In analyses consistent with the flamingo–grebe couplet, contradictory or implausible nodes suggestive of systemic problems seldom are given comparable attention, e.g. the union of grebes with cuckoos (Van Tuinen et al., 2000), and the grouping of flamingos with nighthawks, loons with cuckoos, and grebes with sandgrouse (Fain and Houde, 2004). Also, in most studies supportive of the grebe–flamingo group, aberrancies among other taxa that pertain to reliability of the global reconstruction pass without comment. Examples of considerable controversy include: Apodiformes (swifts and hummingbirds) with Rallidae (rails) by Morgan-Richards et al. (2008); and paraphyly of Charadriiformes (shorebirds) with respect to Cuculiformes (cuckoos) and Passeriformes (passerines) by Pratt et al. (2008). More critically, the flamingo–grebe couplet in turn was inferred to be the sister group of: (i) the Mesitornithidae (mesites) by Ericson et al. (2006); or polyordinal assemblages including (ii) Phaethontidae (tropicbirds), Pteroclidae (sandgrouse), Mesitornithidae, and some Columbiformes (pigeons and doves) by Hackett et al. (2008); or (iii) Falconiformes (diurnal raptors), Coliiformes (mousebirds), and some representatives of the Caprimulgiformes (nightjars) and Columbiformes by Brown et al. (2008). Morphological characters Analytical dimensions of morphological works (above) and the quality of underlying characters proved influential for purported support of this hypothesis. Eleven similarities of the grebes and flamingos proposed as synapomorphies by Mayr (2004, 2007a) are numbered below, eight of which showed homoplasy despite the comparatively few taxa analysed. Two addenda by Manegold (2006) are lettered, but lacking a formal analysis, no consistency indices were given. (30) "At least fourth to seventh cervical vertebrae strongly elongate, processus spinosus forming a marked ridge". Attributed a consistency index (CI) of 1.0 (i.e. unique for included taxa), this feature is much more prominent and serially extensive (third through 14th vertebrae) in flamingos than in grebes (fourth through seventh vertebrae), rendering sequential homology of states among taxa questionable. (31) "At least 23 praesacral vertebrae". Accorded a CI = 0.5, the states were delimited by unequally wide ranges of counts, and ordering inflated the influence of the character. Even modal counts—21 (storks), 23 (flamingos), and 25 (grebes)—reveal the ambiguity of intermediacy. Mayr (2006, p. 60; 2009, p. 69) later described this character in terms of "… an unusually high number of cervical vertebrae". Modal counts of the latter proved no more convincing than tallies for "praesacral" vertebrae (latter including virtually invariant thoracic vertebrae), and the states cited as supportive also pertain to some storks, herons, cranes, waterfowl, and cormorants. (32) "At least four thoracic vertebrae fused to [sic] a notarium (CI = 0.25)". This purported apomorphy shows considerable homoplasy at several taxonomic scales (Storer, 1982), varies in number of incorporated vertebrae, and the state cited is shared by galliforms and falconids (Livezey and Zusi, 2006, character 892), neither order implicated as closely related to grebes, flamingos, or each other. (41) "Humerus with a marked oval depression at insertion site of musculus scapulohumeralis cranialis (CI = 1.0)". This impressio insertii also occurs in loons, whereas the associated muscle is absent in Phoenicopterus ruber and most Ciconiiformes (Vanden Berge, 1970), voiding homology and the synapomorphy. The condition in flamingos also was misinterpreted by Olson and Feduccia (1980b). (44) "Ulna, distal end with marked depressio radialis (CI = 0.333)". Flamingos and loons share a state that is absent in grebes (Livezey and Zusi, 2006, character 1524). (46) "Phalanx proximalis digiti majoris very elongate and narrow … (CI = 0.5)". The ratio employed suggested flamingos to be intermediate in state between ibises and grebes, and related features were most similar between flamingos and ibises. (56) "Tibiotarsus, distal rim of condylus medialis distinctly notched (CI = 0.5)". This assessment is erroneous. Livezey and Zusi (2006, character 2155) found that loons and grebes shared a state distinct from that observed in flamingos. (62) "Musculus iliotibialis lateralis, pars acetabularis absent (CI = 0.5)". This character is extremely variable at several undisclosed scales (McKitrick, 1991). (63) "Musculus caudofemoralis, pars caudalis absent (CI = 0.333)". In light of the survey by McKitrick (1991), the polarity of this character is in doubt. (68) "Wing with 11 primaries (CI = 0.5)". Uncertainty of homology is not ironic (contra Mayr), in that the state in grebes and flamingos, considered "apomorphic" by him, is shared as well by some storks (Stresemann, 1963). (69) "Eggs covered with a chalky layer … (CI = 1.0)". This character is also found in the Megapodiidae, plesiomorphic galliforms excluded by Mayr (2004). "Prominent caudolateral projection on ventral side of cervical vertebrae (processus ventrolaterales)". The process in flamingos illustrated by Manegold (2006) is not homologous with the facies articularis of "processus postlateralis" of grebes, which is associated with a unique insertio accessorius tendinis m. longus colli ventralis (Zusi and Storer, 1969). This feature also is related to reduction of the processus costalis, a structure prominent throughout the cervical vertebrae of flamingos (Livezey and Zusi, 2006, character 822), obfuscating serial homology. "Ungual phalanges nail-like". This state in flamingos is much more similar to that in storks (e.g. Ciconia and Mycteria) than the unique state found in grebes. A review of the matrix for avian pelvic musculature by McKitrick (1991), credited by Mayr (2004, p. 167) for two myological characters thought to be consistent with the couplet in question, revealed noteworthy omissions. Eight characters tabulated by McKitrick (1991), but excluded by Mayr (2004), united grebes with loons (her characters 8, 11, 12, 27, 34, 42, 62 and 64), one linked loons with flamingos exclusive of grebes (58), and one joined loons and grebes together with flamingos (52). Counter-evidence also is augmented by the placement of flamingos by McKitrick (1991) in her original analysis of the complete myological data set. Given the evidence contradictory to a union of grebes with flamingos using the 2954 characters analysed by Livezey and Zusi (2006, 2007) and the seven unambiguous synapomorphies uniting grebes with loons and ten other osteological characters supportive of that group in the same study, it is very likely that additional contradictory characters were overlooked by Mayr (2004). The characters of Mayr (2008) include the few deemed supportive of the proposal, for which inflated consistencies are an artefact of reduced taxonomic scale (Livezey, 2009). The matrix of Mayr (2004)—the commentary by Mayr (2007a) being without analysis—proved to be a subsampling of the matrix by Mayr and Clarke (2003), including abridgements from 148 to 70 characters and 43 to 17 taxa. Mayr (2004) reported that a heuristic analysis of his matrix found four equally parsimonious trees of length 197, all of which joined grebes with flamingos with 80% bootstrap support. The matrix presented by Mayr (2004) was reanalysed using Paup (Swofford, 2002) and according to the options given therein, except that here the deterministic branch-and-bound algorithm was employed instead of the heuristic protocol used by Mayr (2004). The reanalysis resulted in: (i) the recovery of only two maximally parsimonious trees of length 219, and (ii) confirmation of 80% bootstrap support for the grebe–flamingo couplet. However, the smaller solution set and greater total lengths of its members found here indicated that Mayr (2004) failed to treat multistate codes as polymorphisms. Instead multistate codings were analysed by default as equivalent to missing data, allowing for a wider set of candidates for substitution. Inverse-constrained, branch-and-bound analyses of the matrix of Mayr (2004) yielded a Bremer (support) index of only five steps for the couplet comprising grebes and flamingos; the five-step deficit identified the Gaviidae as the next-best sister group for grebes for his abridged data. The likely invalidation of fewer than half of the purported synapomorphies listed by Mayr (2004) or the predictable effect of including just half of the contradictory characters omitted by him would suffice to negate the proposed couplet. By constrast, Livezey and Zusi (2006, 2007) favoured the grebe–loon clade by: (i) a bootstrap support of 100%; (ii) a Bremer index of 52; (iii) an additional 146 steps incurred by joining grebes with flamingos; (iv) seven unique synapomorphies (global CI = 1.0) uniting grebes with loons (Livezey and Zusi, 2007, Table 2); and (v) linkage of flamingos with storks (Ciconia) by a bootstrap value of 100% and Bremer index of 21 steps. Combined-data analysis I performed a combined-evidence analysis of the 2954 morphological characters of Livezey and Zusi (2006) combined with an unpublished, comparable matrix of 2944 bp nuclear RAG-1 by Tree-of-Life collaborators G. Barrowclough and J. Groth. To optimize comparability of data and analytical simplicity, heuristic (global parsimony) searches in PAUP were conducted using evenly weighted data (characters versus nucleotides, and transversions versus transitions). Morphological characters were treated following Livezey and Zusi (2006, 2007); sequences were unordered with gaps treated as missing data. Searches for shortest trees were based on ten random starts; bootstrapping was based on 100 replicates, and Bremer indices were the minima from inverse-constrained searches using ten random starting trees. Families analysed followed those selected by Mayr (2004) or inferred to be relevant by Livezey and Zusi (2007), including the Tinamidae as outgroup. Three genera of Galliformes, one Procellariidae (Pachyptilus) and two Ardeidae were absent from the RAG-1 data and omitted. Five terminals for four other large orders included by Mayr (2004) did not alter findings in preliminary analyses, were deemed irrelevant to the problem (Cracraft et al., 2004), and were excluded: Cariamidae, Recurvirostridae, Laridae, Cathartidae and Opisthocomidae. Of the 33 genera analysed, only three—quail (Lophortyx/Callipepla), bitterns (Botaurus/Ixobrychus) and ibises (Plegadis/Eudocimus)—were represented by different genera in the two data sets. Morphological data for the 33 taxa replicated corresponding findings by Livezey and Zusi (2007), including support linking grebes with loons (bootstrap = 87%, Bremer index = 19). The RAG-1 data separately supported the grebe–flamingo couplet (bootstrap = 63%, Bremer index = 4), and hence a dichotomy of signal pertained. The combined analysis of 5898 characters included 1722 parsimony-informative characters for these taxa (g1|104 = −0.728) and supported two shortest trees (length = 7082, CI = 0.41, RC = 0.31), a strict consensus of which was largely resolved (Fig. 1). In both trees, Phoenicopterus was the sister group of Ciconia (bootstrap = 65%, Bremer index = 10), and Gavia the sister group of Podiceps (bootstrap = 100%, Bremer index = 50). Placing Phoenicopterus as the sister group of Podiceps for the combined data alone required 116 additional steps (2% of total tree length). Figure 1Open in figure viewerPowerPoint Maximum-parsimony analysis of 2954 morphological characters (Livezey and Zusi, 2006) and 2944 nucleotide sequences for the RAG-1 nuclear gene (G.F. Barrowclough and J.G. Groth, unpublished data). Strict consensus tree of two shortest solutions for 33 neognathous genera, using palaeognathous Eudromia as outgroup. Families are shown as terminal taxa after Mayr (2004, fig. 1), with genera in parentheses. Nodes pertaining to flamingos (Phoenicopterus) and grebes (Podiceps) are labelled with bootsrap percentages (above) and Bremer support values (below). Chronicle of perceptions Suspicion approaching paranoia For some, the spectre of convergence is omnipresent, and although only circumstantial, the mere supposition has passed for formal counter-evidence (e.g. McCracken et al., 1999; Sorenson et al., 1999). Attempts to defend such bias (e.g. Hedges and Maxson, 1996, 1997) principally entail subjective suspicions and vague references to functional similarities instead of character analyses. Despite much homoplasy in sequence data (Lee, 1997), however, erroneous influence of sequence data is rejected by unmitigated assertion or is simply not entertained. Mayr (2004, 2007a) disparaged morphological likenesses unfavourable to the preferred hypothesis citing assumed convergence, while emphasizing differences (i.e. negations of similarities) between grebes and loons as countering their close relationship. Chubb (2004) suggested that the "true" phylogenetic affinity of grebes and flamingos had been obscured by dissimilarities of independently acquired aquatic specializations despite their purportedly close relationship. In sum, convergence and divergence and similarities and differences were enlisted in the devaluation of characters inconsistent with a favoured hypothesis of relationship for the couplet. Form, function, environment, and genomes, however, interactively mediate evolutionary phenotypic patterns, and these factors are mutually influential regarding the signs of history sought in phylogenetic reconstruction. Common ancestors of modern lineages share initial states at the point (node) of basal cladogenesis, subsequent to which descendant lineages are subject to independent and phylogenetically constrained evolutionary change. Taxa sharing feeding habits as plausibly might have acquired a feature related to feeding independently, or closely related taxa also may share synapomorphies (possibly secondarily differentiated) together with attributes of life history. If the true phylogeny is known (obviously rarely if ever the case), topological incongruence of characters renders an hypothesis of convergence logically admissible. The reverse process, however, in which an assumption of convergence is used to exclude or favour one of several hypotheses or "trump" contrary evidence a priori, is not logically admissable. Even Mayr (2007b: 1446) conceded that: "there is no way to recognize convergence a priori". Implications for empirical standards A prejudice exists that favours molecular data over morphological evidence (Livezey and Zusi, 2001; Assis, 2009; Livezey, 2010b; Vogt et al., 2010), at times strident (e.g. Hedges and Sibley, 1994; Van Tuinen, 2002), and despite cogent support for pluralism of approach (Hillis, 1987; Bledsoe and Raikow, 1990; Patterson et al., 1993; Baker et al., 1998; Hillis and Wiens, 2000; Wiens, 2004; Smith and Turner, 2005; Pisani et al., 2007). In morphological analyses, some evidently view empirical conflicts not as predictable ephemera of diverse investigations of complex phenomena using signals of variable quality, but as errors routinely correctable by deletion of offending contradictions or methodological finesse. Mayr (2007a, p. 7) asserted that morphological inferences are more worthy if made "… before the molecular evidence was available …", a condition at odds with his presentation of morphological characters construed to corroborate prior molecular inferences (Mayr, 2004, 2005, 2007a). Mayr (2007a, p. 7) also stated that support values "are not necessarily accompanied by phylogenetic accuracy", a suggestion reasonable but untestable in the absence of a genuine standard by which to judge accuracy. In a critique of Livezey and Zusi (2006, 2007), Mayr (2007a) speculated that extensive data "may work well" in molecular studies, but that such richness in morphological studies may be conducive to "severe shortcomings". The latter assertion turned on the baseless claim that variation in number of states per morphological character is analytically misleading, and is at odds with known advantages of expanded databases (Sanderson, 1989; Sanderson and Donoghue, 1989; Pickett et al., 2005; Prevosti and Chemisquy, 2010). The preference shown for narrowed samples of morphological characters by Mayr (2004, 2007a) may reflect, in part, an affinity for the apparent but often false clarity of small data sets. It is incontestable that knowing the answer to a problem facilitates its solution, but certain discovery of truth is unlikely or impossible in real-world phylogenetics. Constructive discourse is difficult where reduction of information or respect for only one kind of evidence is advocated for the solution of complex problems. The objective here is not to argue for a
Referência(s)