Sunteți pe pagina 1din 17

Society of Systematic Biologists

Phylogeny as a Central Principle in Taxonomy: Phylogenetic Definitions of Taxon Names Author(s): Kevin de Queiroz and Jacques Gauthier Reviewed work(s): Source: Systematic Zoology, Vol. 39, No. 4 (Dec., 1990), pp. 307-322 Published by: Taylor & Francis, Ltd. for the Society of Systematic Biologists Stable URL: http://www.jstor.org/stable/2992353 . Accessed: 26/05/2012 12:05
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Taylor & Francis, Ltd. and Society of Systematic Biologists are collaborating with JSTOR to digitize, preserve and extend access to Systematic Zoology.

http://www.jstor.org

Syst.Zool., 39(4):307-322, 1990

PHYLOGENY AS A CENTRAL PRINCIPLE IN TAXONOMY: PHYLOGENETIC DEFINITIONS OF TAXON NAMES


KEVIN DE QUEIROZ AND JACQUES GAUTHIER Department ofHerpetology, California Academy ofSciences, GoldenGate Park,San Francisco, 94118 California Abstract.-Definingthe names of taxa in termsof common ancestry, thatis, using phylogenetic definitionsof taxon names, departs froma traditionof character-based definitionsby granting the concept of evolution a central role in taxonomy. Phylogenetic definitionsbear on other taxonomic principles and practices,including the following: (1) Names cannot be applied to nonmonophyletictaxa as the result of mistaken ideas about relationships or characters.Such mistakesdo not affect conclusions about the monophylyof taxa but about their contentand/or diagnoses. Because nonmonophyletictaxa can only be named deliberately,theyare easily avoided. (2) Definitions are clearly distinguished fromdescriptions and diagnoses. Definitions are ontologicalstatements about the existenceofentitiesthatresultfrom the relationshipsofcommon ancestryamong their parts; descriptions and diagnoses are epistemological statementsabout how we recognize the parts of those entities. (3) By precisely specifyingthe clade (ancestor) with which a name is associated, phylogeneticdefinitionsclarifythe limitsof taxa, and this in turn clarifiesrelated phenomena such as time of origin and duration. (4) Unlike the case for character-basedintensional definitionsand enumerative extensional definitions,phylogenetic definitionsprovide an unambiguous criterionfor synonymyof taxon names: names are synonymous if theyreferto clades stemmingfromthe same ancestor.(5) Because of theirrelevance to synonymy, phylogeneticdefinitions also are relevantto priority, ofbothnames and authorship. In phylogenetic taxonomy,priorityis based not on first use of a name at a particularrank or rank-groupbut on firstuse of a name in association with a particular ancestor. [Definition; diagnosis; nomenclature;phylogeny; priority; synonymy;taxonomy.]

Such expressionsas that famous one of Linnaeus, and which we often meet with in a more or less concealed form,that the charactersdo not make the genus, but thatthe genus gives the characters, seem to imply thatsomethingmore is included in our classification, thanmere resemblance.I believe thatsomethingmore is included; and thatpropinquity of descent ... is the bond, hidden as it is by various degrees of modification, which is partially revealed to us by our classifications [Darwin, 1859: 413-414].

Despite the widespread belief thatmodern biological taxonomy is based on our the deknowledge of evolutionaryhistory, velopment of phylogenetic taxonomy remains incomplete.As the principleof common descent gained acceptance,it took on a superficialratherthan centralrole in taxonomy. Preexistingtaxonomiescame to be interpreted as the result of evolution, and evolutionaryprincipleswere developed to justifylong-standingtaxonomic practices, but the principle of descent did not bewhich taxonomic come a centraltenetfrom principles and methodswere derived (Stevens, 1984; de Queiroz, 1988).

In several importantrespects,the concept of evolution retains this superficial ratherthan centralrole. The definitionsof the names of taxa is one area in which the concept of evolution has yet to be granted a central role-a change that would have important taxonomic consequences. Althoughthe possibilityofphylogeneticdefinitionsof taxon names, thatis, definitions based on phylogenetic relationships,has been suggested (e.g., Wiley, 1979, 1989; Ghiselin, 1984; Ridley, 1986; Rowe, 1987; de Queiroz, 1988; Gauthier et al., 1988a), the significance and consequences of adopting this kind of definitionhave not received a detailed treatment. Here we exof taxon plore the history ofthe definitions names, the fundamental role that definitions based on ancestryplay in phylogenetic taxonomy,and some of the consequences of adopting such definitions. as we use this "Phylogenetictaxonomy," is concerned with the representation term, of phylogeneticrelationships,specifically, with issues related directlyto the naming

307

308

SYSTEMATIC ZOOLOGY

VOL.

39

how it is in taxonomy." In this light, it is suroftaxaand notwiththoseconcerning Consequent- prising that the way in which the names aredetermined. relationships ly, we assume the goal of representingof taxa are defined has changed littleover as well as some the last 2,000 years. It also exemplifiesthe relationships phylogenetic of the moreinclu- failure of the theoryof evolution to penof the basic principles sys- etratethe core of biological taxonomy.Alsive disciplinecalled "phylogenetic have been of definition tematics," for example,the principleof thoughotherforms taxon names have of suggested,traditionally, and the untenability synapomorphy taxa. Thelogicalbasesofthese been treatedas iftheyare definedby charparaphyletic have been consideredthor- acters possessed by those organisms conpropositions 1966, sidered to be membersof the taxa. Under 1965, elsewhere (e.g.,Hennig, oughly taxa have definitions, 1966; Nelson, 1971, 1974; such character-based 1975; Brundin, 1978;Eldredgeand been treated as if they are classes of orRosen,1974;Platnick, 1980;Wiley,1981; Ax, 1987; de ganisms (Ghiselin, 1969, 1974; Hull, 1976, Cracraft, Queiroz,1988).Becausewe are concerned 1978; de Queiroz, 1988; for explicit examof common ples see Gregg, 1950; Beckner, 1959; Buck with the namingof systems "classification,"and Hull, 1966, 1969; Muir, 1968; Suppe, we avoidtheterm descent, whichis bestrestricted to therecognition 1974, 1989) in that their charactershave 1974; Hennig, 1975; been viewed as intensions, that is, those of classes (Griffiths, properties connoted by the taxon names. Ax, 1987;de Queiroz,1988). taxono- Consequently, traditional definitions of of phylogenetic Our discussion my concernsthe so-called higher taxa, taxon names have been considered intenare sional definitions(e.g., Ruse, 1973). Such systematics, which, in phylogenetic entities (e.g., definitionshave their roots in the Aristonamed cladesormonophyletic Hennig,1966),thatis, biologicalsystems telian formof definition,which was sumfrom the process marized by Hull as follows: theirexistence deriving 1974;Henof commondescent(Griffiths, nig,1975;Ax,1987;de Queiroz,1988).The In Aristotle's view threethingscan be known about and itsname. any entity-its essence, itsdefinition, to taxaof the relevanceof our discussion The name names the essence. The definitiongives one's dependson whether speciescategory a complete and exhaustive description of the esofthe to a subgroup refers speciesconcept sence. Derivatively,the name is the name of the "clades," in which case the arcategory entityand the definitiona descriptionof it [Hull, hereapply to species, 1965:318]. presented guments to theclassofbiologicalsysor ifit refers Most modern taxonomists have abantheprofrom their existence tems deriving in whichcase these doned the notion ofessences (but see Kitts, of interbreeding, -cess do notapplyto species(see de 1983; Bernier, 1984), and many have rearguments we jected the notion that taxa are classes of Queirozand Donoghue,1988).Finally, Def- organisms. Nevertheless, elements of the names. oftaxon discuss thedefinitions initions are given to words, including Aristotelianform of definitionhave pernames,not to the thingsthatthe names sisted in modern biological taxonomy in that (Ghi- thatthe names of taxa continue to be treatis,thetaxathemselves represent, authors sel- ed as if they are defined by lists of organselin,1966a).Becauseprevious (see thesectiontitled"Defbetweentaxaand their ismalcharacters domdistinguished definitions inition" in numerous recent taxonomic treated names, theyeffectively papers). as iftheyapplied to both.
THE HISTORY OF DEFINITIONS OF TAXON NAMES

Aristotelian Definitions as as important otherscienceis definition


According to Hull (1965:315),
....

Cluster Character-Based Concepts: that Accommodated Evolution Definitions


in no In post-Darwinian times it has become accepted widely that biological taxonomy is, or should be, evolutionary. However,

1990

PHYLOGENETIC DEFINITIONS

309

even disassociated fromthe notion of essences, the Aristotelianformof definition proved to be incompatible with the principle of evolution (Hull, 1965). As Hull (1965:318) noted, "... what Aristotlewas advocating in modern termsis definition by properties connected conjunctively which are severally necessary and jointly sufficient [formembershipin the taxon]." But evolutionarychange may resultin loss in lineages withina taxon, or modification, of the very charactersthatsupposedly definethe name of thattaxon.Given thatthe modifiedlineages remain parts of the taxon, it mustbe concluded thatthe characters in question were not necessary and sufficient to define its name (Sober, 1988; see also Beatty,1982). Hull (1965) noted thatnames can be definedby kinds of (intensional) definitions other than conjunctive ones, that is, definitions in which each propertyin the entire set is necessary for membership(e.g., a bachelor is both unmarried and male). For example, names also can be defined disjunctively,that is, by sets of properties not all of which are necessary,withoutviolating the spiritofAristoteliandefinition. Thus, each propertyin a definitedisjuncby itself, tive definitionmay be sufficient and only some subset of the propertiesis necessary for group membership (e.g., a sibling is eithera brotheror a sister).Hull (1965:323; see also Beckner, 1959; Hull, 1974) argued, however, that "... neither of these types of definitionis appropriate fordefiningthe names of taxa and, hence, for delineating taxa. Whether from the viewpoint of phylogenetic or numerical taxonnames can be definedonly taxonomy,

tions. Because no characteror set of characters was necessary for the definitionof a taxon name, and any one of numerous sets was sufficient, evolutionary change could occur without invalidating the definition.Nevertheless,clusterconcepts,like Aristotelian definitions, defined taxon names in terms of characters. Although such definitions may have been consistent with evolution, they were not evolutionary,which is to say thattheymade no reference to common descent or any other evolutionaryphenomenon.

Phylogenetic Definitions Taxonomists have long suspectedthatthe characters of taxa are merely manifestations of a more fundamentalphenomenon (see epigraph).Butdespitethefactthatmost taxonomists subsequent to Darwin have accepted evolution as thatunderlyingphenomenon, they are only beginning to realize that this concept provides the possibility of an entirely differentkind of definition. Viewing taxa as wholes, or individuals, led Ghiselin (1966b, 1974, 1981, 1984) and Hull (1976, who rejected his earlier treatmentof taxaas classes) to conclude thatthe names of taxa are proper names and hence can only be definedostensively,thatis, by pointing out the entityto which the name is given. A simple enumerationof included subtaxa (pointing out the members or does not qualifyas an evorepresentatives) forit makes lutionaryostensive definition, if no referenceto evolution. Furthermore, enumerationis viewed as listing a taxon's known contentwith an implicitstatement about common ancestry, it fails to capture covarying properties by sets of statistically in indefinitely defi- the intended meaning of the name. Simple longdisjunctive arranged nitions"(emphasis in original). The dis- enumeration makes no provision for the tinction between definite conjunctive discovery of previously unknown repredefinitions and indefinite disjunctive sentativesof a taxon. In order to make the was also made by Beckner(1959) definitionsof taxon names evolutionary, definitions and by Sneath (1962), who used the terms theymustbe rooted in the concept of com"monothetic" and "polythetic" to referto mon ancestry.Thus, an evolutionary oshereafter referred to as the classes of taxa based on these two class- tensive definition, a phylogenetic definition, consists of es of definitions. Indefinite disjunctive definitions, or pointing to a clade, that is, to an ancestor clusterconcepts,avoided the problem that and its descendants (Ghiselin, 1984). This evolution posed for Aristotelian defini- can be accomplished, verbally or on a

310

SYSTEMATIC ZOOLOGY

VOL.

39

a
FIG. 1.

Three possible ways of definingtaxon names phylogenetically.(a) Node-based definition.(b) Stembased definition.(c) Apomorphy-baseddefinition(bar indicates origin of the apomorphy). See text for explanations and examples.

branching diagram, in at least three differentways: 1. By definingthe name of a taxon as the clade stemmingfromthe mostrecentcommon ancestor of two other taxa, hereafter (Fig. definition referredto as a node-based la). For example, Gauthieret al. (1988a:34) defined"Lepidosauria" as "the mostrecent and squacommon ancestor of Sphenodon mates and all of its descendants." (For other examples see de Queiroz [1987],Estes et al. [1988], Gauthier et al. [1988a, b], and Rowe [1988].) 2. By definingthe name of a taxon as all those entities sharing a more recent common ancestor with one recognized taxon to as referred than with another,hereafter (Fig. lb). For example, definition a stem-based Gauthier et al. (1988a:27) defined "Lepiand squadosauromorpha" as "Sphenodon mates [Lepidosauria] and all saurians sharing a more recent common ancestor with them than ... with crocodiles and birds." (For other examples see Gauthier et al. [1988a, b].) 3. By definingthe name of a taxon as the ancestor to clade stemmingfromthe first possess a particularsynapomorphy,heredefto as an apomorphy-based afterreferred inition (Fig. lc). For example, "Tetrapoda" veras the first is usually definedimplicitly tebrate to possess digits (i.e., hands and feetratherthan fins)and all of its descendants. All three classes of phylogenetic defirooted in the concept of nitions are firmly common ancestry in that this concept is fundamental to the meaning of taxon names. Although the apomorphy-based to makesreference phylogeneticdefinition

fundamentallyfrom a character,it differs In intensionaldefinition. a character-based contrastwith the latter,the characterused in an apomorphy-basedphylogeneticdefinition is simply a means of specifyingan does not implythe ancestor.The definition presence of the characterin all organisms of the taxon, for these charactersmay be lost in some descendants of the specified ancestor.Thus, it is neither necessary nor sufficient for an organism to possess the characterin order to be considered part of the taxon; what is both necessaryand sufficientis being descended fromthe specified ancestor. The use of phylogeneticdefinitionsliberates biological taxonomy from a 2,000year-oldtraditionof basing the definitions of taxon names on characters.Because of the antiquityof this traditionand the imin taxonomy(Hull, portanceof definitions 1965), one might expect the adoption of to have far-reachphylogeneticdefinitions ing consequences for other taxonomic principles,practices,and conventions.
PROPERTIES OF PHYLOGENETIC DEFINITIONS

totheConcepts ofMonophyly, Relationship andPolyphyly Paraphyly,


Traditional definitionsof taxon names take the same form,that is, lists of characters,regardless of whether the taxa are monophyletic, paraphyletic, or polyphyletic. Although the definitions of these three classes of taxa are stated in termsof common ancestry(e.g., Hennig, 1966; Farris, 1974), the classes themselves are nevclasses erthelesscharacterizedby different of characters:synapomorphiesin the case

1990

PHYLOGENETIC DEFINITIONS

311

of monophyletictaxa, symplesiomorphies in the case of paraphyletictaxa, and convergences in the case of polyphyletictaxa (Hennig, 1965, 1966, 1975). For any given taxon,however, both the patternsof common ancestrythatwould place it in one of the three classes of taxa and the apomorphic, plesiomorphic,or homoplasticstatus of its characterscan be recognized only in the context of an accepted phylogeny. Consequently, under character-baseddefinitions of taxon names, nonmonophyletic taxa can be named accidentally. If phylogenetic relationships are unknown (the or hosymplesiomorphic, synapomorphic, moplastic status of the characters is ignored or uncertain) or if notions of phylogenetic relationships are in error (symplesiomorphies or homoplasies are mistakenforsynapomorphies),then parataxa can be named phyleticor polyphyletic accidentallyeven by thosewishing to avoid them. Under phylogenetic definitions,names to nonwill never be defined as referring monophyletic taxa as the result of mistakes; the associations must be deliberate, and this holds regardlessof potential mistakesabout phylogeneticrelationships(see below). Consequently,the verydefinitions of theirnames give warningofparaphyletic and polyphyletic taxa. The names of paraphyletictaxamustbe definedas monophyletic taxa fromwhich a part or parts have been removed,that is, as incomplete systemsof common ancestry.For example, "Reptilia," traditionally the name of a paraphyletic taxon, might be defined as "the most recent common ancestor of Mammalia and Aves and all of its descendants except Mammalia and Aves." On the otherhand, the names of polyphyletictaxa must be defined as sets of two or more separate systemsof common ancestry.For example, "Haemothermia," the name of a diphyletictaxon,mightbe definedas "the most recent common ancestor of monotremes and therians and its descendants (=Mammalia) andthe mostrecentcommon ancestorof paleognathsand neognathsand its descendants (=Aves)." The incomplete nature of paraphyletic taxa and the com-

posite nature of polyphyletictaxa are thus revealed directlyby the manner in which their names are defined. In contrast,definitions of the names of monophyletictaxa contain no phrases excluding certain descendants of the specified common ancestor from the taxon, nor do they require conjunctions linking separate systems of common ancestry. The relevance of these considerationsto phylogenetic taxonomy is that because names can only be associated with nonmonophyletic taxa deliberately, such associationsare easily avoided. Furthermore, phylogeneticdefinitions can be worded so that they necessarily refer onlyto monophyleticentities. This does not mean that mistakes cannot be made; however, such mistakes will be seen for what they are, that is, mistakes about phylogenetic relationships, diagnostic characters,or included taxa. They will never resultin the conclusion that a name originally defined as referringto a monophyletic taxon actually refers to a paraphyletic or polyphyletic group. For example, Gauthier et al. (1988a) defined the name "Lepidosauromorpha" as all saurians sharinga more recentcommon ancestor with lepidosaurs than with archosaurs. They considered Lepidosauromorpha to include lepidosaurs, kuehneosaurs, Paliguana*,Saurosternon*, Palaeas illustrated agama*, and Younginiformes, 2a (an asteriskaftera taxon name in-Figure indicates that evidence for monophyly is lacking). Now suppose that the true phylogenetic relationshipsof Younginiformes are as illustratedin Figure 2b; this taxon is the sister group of the clade stemming fromthe most recent common ancestor of Lepidosauria and Archosauria. Under traditional definitionsof taxon names, one might conclude that Lepidosauromorpha is paraphyletic, because itnow appears that Archosauria is descended from the most recent common ancestor of the taxa originally considered to be lepidosauromorphs (Fig. 2b). Butaccordingto the phylogenetic definitionas stated,this is not so. The definition of "Lepidosauromorpha" does not contain a phrase excluding any descen-

312
A Pa S P K

SYSTEMATIC ZOOLOGY
A Pa S Pi K A Pa S Pi

VOL.
K

39
L

bC

and monophyly.(a) PhylogeneticrelationshipswithinSauria, with bubble FIG.2. Phylogeneticdefinitions encircling Lepidosauromorpha as its name was defined by Gauthier et al. (1988a). (b) Hypothetical change (Y), which seems to implyparaphylyofLepidosauromorpha in ideas about therelationshipsofYounginiformes when the contentratherthan the definitionis considered. (c) The actual limits and monophylyof Lepidosauromorpha under the revised relationships according to the definitionas stated. See text for additional Pi, Paliguana*; A, Archosauria;Y, Younginiformes;Pa, Palaeagama*;S, Saurosternon*; discussion. Abbreviations: K, Kuehneosauridae; L, Lepidosauria.

dants of the specified common ancestors; this name mustreferto a monotherefore, phyletic taxon. And, because it is defined with respectto Archosauria,one mustconclude notthatLepidosauromorpha is paraphyletic, but that Younginiformesis not part of it (Fig. 2c). By granting the principle of descent a central role in biological taxonomy,phylogenetic definitions reorient questions away fromthe monophyletic,paraphyletic, or polyphyleticstatus of taxa to their content and diagnoses. The concepts of monophyly, paraphyly, and polyphyly were formulatedin an era when the evolutionary world view had been accepted but was grantedthe role of littlemore than of the order interpretation an after-the-fact already manifestin taxa (see de Queiroz, 1988). Largely as a resultof Hennig's writings (e.g., 1965, 1966, 1975), the role of the concept of evolution in biological taxonomy began to change -fromthat of a suto that of a central perficialinterpretation tenet (de Queiroz and Donoghue, 1988, 1990). But Hennig did not take this perspective to its logical conclusion, that is, oftaxon adopting phylogeneticdefinitions in such names and statingthese definitions a way that all taxa necessarily would be monophyletic. And because taxon names have continued to be definedwith respect to charactersratherthan ancestry,it was meaningfulnot only forHennig, but also for systematicsup to the present time, to inquire about whether these names referredto monophyleticentities.

will definitions The use ofphylogenetic a new era in biological initiate effectively will be, in one In thisera there taxonomy. for no existing taxa(namedentities), sense, thenameshavenotyetbeentiedexplicitly defiphylogenetic through to the entities Once thesenamesare tied to the nitions. as complete entitiesby theirdefinitions all taxon systemsof commonancestry, to namescan be wordedso thattheyrefer Conseentities bydefinition. monophyletic and quently,the conceptsof paraphyly and insuperfluous willbecome polyphyly in a historical context. primarily teresting no longerwill be question The significant paraphya taxon is monophyletic, whether (as in theexinstead or polyphyletic; letic, above) the ample of Lepidosauromorpha, aboutany taxonwill questions significant charand itsdiagnostic concern itscontent acters (see below). the of granting As one manifestation of descentthe role of a central principle definiphylogenetic tenetin taxonomy, thatcommon ancestry tionsacknowledge to the existenceof the is fundamental thispropbeingnamed.Accepting things ositionprovidesthe basis forstable and meaningsof unambiguousphylogenetic of changing taxonnames in the context both ideas aboutthedetailsofphylogeny, and in terms of in terms of relationships charthepreciselevel at whichparticular This staactersexistas synapomorphies. thefact that, derives from and clarity bility ancestry common forany two organisms, theprecise whereas atsome levelis certain,

1990

PHYLOGENETIC DEFINITIONS

313

content ofa taxonand itsdiagnostic char- temology. This distinction is obscuredby acters are not. traditionaldefinitions of taxon names, because traditional definitions take the Metataxa same form as descriptions and diagnoses, is,listsofcharacters The relevanceof phylogenetic defini- that (Ghiselin, 1966a). tions to metataxa-taxabased on plesio- Froman essentialist's perspective, characmorphy but forwhich evidenceof para- ter-based intensional definitions makeonphyly, and monophyly, is lacking tological (as well as epistemological) stateinthat (Donoghue,1985;Gauthier etal., 1988a)- ments they areviewed as descriptions is similar to thatconcerning paraphyletic ofa taxon's essence, and theessenceis that it is pos- from and polyphyletic taxa.Although whichthetaxon derives itsexistence. sible to definethe name of a metataxon When the essentialist are underpinnings phylogenetically, there wouldbe little rea- removed,however,character-based defison fordoingso. Metataxa requireuncer- nitionsconfuse ontology and epistemolotainty aboutrelationships of common de- gy.The taxonseemsto exist solelybecause scent, and, therefore, the phylogenetic itsorganisms havecharacters bywhichwe of theirnamesrequireambig- recognize themas belonging definitions to thetaxon. uous statements abouttheserelationships. Adopting phylogenetic definitions clarthedistinction thename ifies between Usinga phylogenetic definition, definition and of a metataxon mustbe defined as a com- diagnosis.Here the definitions of taxon monancestor and all,ormaybe onlysome, names are ontological statements in that refer of its descendants. tomonophyletic Like the conceptsof they entities (clades), toexist and polyphyly, the metataxon whicharepresumed underthecenparaphyly ofcommon descent conceptis theproduct of an era in which traltenet independent phylogeneticrelationships had become of our abilityto recognizethem.Monotaxa, that is,namedclades,arenot but characters, phyletic important considerations, ratherthan phylogeneticrelationships, concepts(contraL0vtrup, 1986) but real in the things-systems continuedto be grantedprimacy derivingtheirexistence thevery from common ancestry relationships definitions oftaxon names. Indeed, and meta- amongtheir factthatparaphyly, parts polyphyly, (Griffiths, 1974;Hennig, taxaareimportant is evidence that 1975;de Queiroz,1988).Phylogenetic concepts defas if initions are of littleuse, however, taxonnames continueto be treated unless rather than the criticalrelationships theyare defined by characters can be deterIf the use mined.Thisis one roleofcharacters. by phylogenetic relationships. And of phylogenetic becomespre- afterthe relationships definitions have been estabthenthemetataxon like lishedand thetaxonnamesdefined, dominant, concept, charand polyphyly, willbe- actersalso permitthe determination thoseofparaphyly of come superfluous that is, except in a historical which entitiesfit a definition, a given specimenis a represencontext. whether tativeof the taxonin question(see Hull, to Characters Relationship as symptoms are used to diag1976).Just of phylogenetic has an unseen Some properties defi- nose a disease(whichoften nitionsare clarified their underlying are used cause), so characters by considering to characters. It is sometimes to diagnosetaxa(Ghiselin, relationship 1984). said thatcharacters are diagnostic Thus, in the previous example, the (or deoftaxarather thandefining scriptive) (e.g., changein ideas abouttherelationships of does notchangethe defGhiselin, 1966a,b, 1984,1985;Hull, 1976; Younginiformes Wiley, 1979;Beatty, 1982;Ridley, 1986;So- inition of "Lepidosauromorpha" or its status. It does, however, betweendefi- monophyletic ber, 1988). The difference nitionand diagnosisis that -betweenan imply changes in ideas not only about itself and the evidenceforits exis- included taxa but also about diagnostic entity and epis- characters. One possibility is that thechartence,thatis,betweenontology

314

SYSTEMATIC ZOOLOGY

VOL.

39

a closer to support formerly thought acters to Taxon Limits and Relationship to Lepiof Younginiformes relationship Origination Times diactually Archosauria dosauriathan to Byprecisely specifying theclade (ancesagnose a moreinclusivetaxonof which phylogenetic Archosauriais a part. Perhaps derived tor)to whicha namerefers, about the to be absent definitions clarify ambiguities thought formerly characters under tobe present limitsof taxa thatare perpetuated aredetermined in archosaurs definitions. character-based In thiscase the char- traditional, form. in a modified ofnew thediscovery outto be irrelevant One oftheseconcerns in question turn acters form of defifitthe def- taxa. Under the traditional whichentities to determining consideredprefofthename"Lepidosauromorpha." nition,it was generally inition ofa taxon name tobasethedefinition sharing erable as all saurians Thisnameis defined because presumably ancestor withLepi- on severalcharacters, common a morerecent better constitutes even if havingmorecharacters dosauria than with Archosauria, forthattaxon.However,the disof support to be diagnostic thought the characters of organisms possessingsome but apply to a moreinclu- covery thistaxonactually sive clade. notall ofthedefining ofa taxon characters the conclusionthatcer- posed problems. Aretheseorganisms part Alternatively, of ofthattaxon? are diagnostic Areall ofthecharacters detain derivedcharacters or only some of them,and if the Lepidosauromorpha may remain un- fining, the whichones?Thus,forexample, changed,but now it is concluded that latter, do notpossessthesechar- discovery led to a debate of monotremes younginiforms and the about whether are extinct, Younginiforms theseanimalswere mamacters. in questionmayhave involved mals or representatives of some separate characters partsof the body. Per- taxon(Gould, 1985).That thisdebate ocpoorlypreserved orfurther preparation curredimpliesthatthe limitsand hence hapsnewspecimens has revealed that the charactersonce the definingcharacters of "Mammalia" absent. were called into question.Similarprobare in fact thought to be present as di- lemscontinue of viewingcharacters A strength tothepresent day,especialagnostic ratherthan definingof taxon ly in paleontology whereorganisms with for new combinations problems namesis that itavoidscertain are disof characters caused coveredregularly form of definition the traditional (see Rowe,1988). in no way diHow can definitions Phylogenetic and homoplasy. byplesiomorphy of such intermeby the minishthe significance be defined forexample, Tetrapoda, attention (e.g., diateforms, buttheyfocus away presenceof limbsif some tetrapods issues to phylois touse from purelydefinitional One possibility snakes)lackthem? in whichnone ofthe de- geneticones. For example,in the case of cluster concepts is necessaryfor taxon Archaeopteryx, finingcharacters the interesting issue is not is tocon- whether these fossilstrulycan be called Another possibility membership. it is theirphylogenetic relasidertheabsenceoflimbstobe a modified birds;rather, Given thatthese relationships "limbspresent" (e.g., tionships. form ofthecharacter thenthe question and have been determined, is simpler 1979).Thesolution Platnick, is a birddepends Archaeopteryx underphylogenetic ofwhether morestraightforward on how "bird" is defined(e.g., Here the absenceof limbsin entirely definitions. the name Thulborn, 1984;Gauthier, 1986). Thus,if snakesin no way compromises is thesister ofall other group as the clade stemming Archaeopteryx defined Tetrapoda, and "birds" limbedancestor. Because knownfeathered vertebrates a particular from of this is defined as the mostrecent common ansnakesare amongthe descendants whether of paleognaths and neognaths and they cestor ancestor, theyare tetrapods thenArchaeopteryx is not have limbsor not.The absenceoflimbsin its descendants, to es- a bird.But,if"bird"is defined as thefirst snakesonly makesit moredifficult and itsdescendants feathered vertebrate or tablish theirtetrapod ancestry.

1990

PHYLOGENETIC DEFINITIONS

315

as all archosaurs sharinga more recent E commonancestor with extantbirdsthan C\J = 0 withcrocodylians, thenArchaeopteryx is a ? : @ o H bird.And althoughit maybe moretraditional,convenient, or usefulto adoptone ofthevarious possible definitions of"bird" and hencesettlethe issue of whether Archaeopteryx is or is not partof thistaxon, Mammalia the choice is a definitional rather than a biologicalissue. Because phylogenetic definitions preciselyspecify the limits FIG. 3. Ambiguitiesin the boundaries of taxa unof taxa,theyalso definitions of taxon names. If difclarify questions abouttimes oforigin and der character-based ferent authorsdefinethe name "Mammalia" with reconsequently about temporaldurations, spect to different sets of the characters(indicated by bothofwhicharesusceptible to confusion bars in diagram),then "Mammalia" will referto difundertraditional definitions (Rowe,1988). ferent clades, some moreinclusive thanothers.ThereUnder character-based dis- fore,the three fossilswill be mammals according to definitions, in theages assignedto a taxon some definitionsbut not according to others. crepancies can result, even in theabsenceofconflicting data, from inconspicuous differences in the definition of the name. What ap- paper.Theseare (1) thetimeat whichcerpearsto be thesametaxonas judgedbyits tainlineagesrepresented byextant organnameand many, perhapsmost, of its"de- ismslastshareda common ancestor (Henfining" characters, mayin fact be twodif- nig's * group, Jefferies's crown group, ferent taxain terms ofthecommon monophylum ances- Lauterbach's sensustricto; a torsimpliedby slightly different sets of specialcase of thenode-based definition), characters. Forexample, Mammalia ofsome (2) the timeat whichthe stemlineage of authorsis based solely on the character thatclade separated from its sistergroup articulation totalgroup,Lauterbach's dentary-squamosal between (Jefferies's pancranium and mandible, a special case of the stemwhereasotherau- monophylum; thors basea taxon ofthesamenameon this baseddefinition), and (3) thetime atwhich and additional characters of the extant (Rowe, 1988). the "typical"characters repGiventhatsomeof theadditionalcharac- resentatives of the clade arose in its stem ters arosebefore orafter thedentary-squa- lineage(a specialcase of the apomorphymosalarticulation, as indeeditapearsthey based definition). A givenname could be did (Rowe, 1988),then the name "Mam- definedin any of these threeways,but malia" does not alwaysrefer to the same each would referto a different ancestor clade(Fig.3). Becausethese to a moreor less inclusive different clades and, therefore, form a nestedseries, their differ. clade. This fact has importantconseagesmust estimates of the age of Mam- quencesforvarioustime-related Therefore, phenommalia maydiffer of the origination given the same hypoth- ena, such as estimates esized relationships of taxa and the caliand known fossils timesand durations of molecular becausecertain simply fossils clocks. will be con- bration sideredmammals to some defi- Phylogenetic definitions according avoidsuchamnitions of "Mammalia"butnotothers. biguities thelimits concerning oftaxa, beThe general problemdescribedabove cause theancestors in questionare clearly was recognized This is not to say thatdifferent byHennig(1965,1981;see specified. also Jefferies, 1979;Lauterbach, cannot use thesamenameto refer 1989).In- authors different ofa taxon's todifferent deed,three clades.However, meanings phylogenetic timeof originrecognized by Hennigrep- definitions of the same name as used by resent specialcases of the threeclassesof different authorsshould reveal whether definitions in this that phylogenetic identified namerefers tothesameorto different
CM X ) Cu) U) U) CD c CD 0

316
cu

SYSTEMATIC ZOOLOGY
0

VOL.

39

mata"is defined as the mostrecentcommon ancestor of Lacertilia, Serpentes, and ~~~00. 0 CL~~~ and its descendants;(2) Amphisbaenia, (0~~~E E g "Lacertilia" is defined as the mostrecent ~~~~~~~~~~~~~~o common ancestor of Iguania, Gekkota, and Anguimorpha, and its Scincomorpha, descendants; and (3) the relationships amongthesetaxaare as shown in Figure 4 (Esteset al., 1988).Without considering phylogenetic thedefinitions relationships, of the two names seem to indicatethat toa more refers inclusive "Squamata" clade thandoes "Lacertilia." Thisis contradicted bythephylogenetic relationships. Thedefinitionsof both "Squamata"and "Lacertilia"refer to the same ancestor-theone at thebase of thediagram (Fig. 4). Theredefinitions. phylogenetic under FIG.4. Synonymy fore, given the definitions statedand the the from as thecladestemming "Lacertilia" (defined these namesare illustrated, of Iguania,Gekkota, relationships ancestor common mostrecent and "Squamata" synonymous. That the two names previand Anguimorpha) Scincomorpha, recent ouslymayhave been viewed as referring from themost as thecladestemming (defined and Am- to different Serpentes, of Lacertilia, ancestor common clades is accountedforby a (circle)if the to thesameancestor refer phisbaenia) mistake the contentof Lacertilia, about among these taxa are as illustrated. relationships and Amphisbaenia thatSerpentes and "Squamata"are synony- namely, "Lacertilia" Therefore, mous. notto be partsof it. werethought Of course, can definitions phylogenetic issuesofsynonymy onlyin theconand iftheydo not,thissuggests clarify ancestors, phy- textof relevantknowledgeabout phyloconcerning possible disagreements be- genetic relationships.The situation is relationships (see Synonymy, logenetic greatly simplified by,but it is not depenlow). denton,there beinga singleaccepted phySynonymy logeny.Thus,in Figure5a, the name deas thecladestemming themost from The issue of taxonomic synonymy, fined common ancestor ofA and D is synto thesame recent namesrefer whether different defini- onymouswith the name definedas the is clarified by phylogenetic entity, from themost recent comand enumer- cladestemming tions.Undercharacter-based is a mat- monancestor ofB and E, even though the often synonymy ativedefinitions, of C are controversial. of relationships Synof degree. When the definitions teralso wouldobtain ifthedefinitions of onymy names are identicalin terms different or bothnamesincluded characters and included taxa, then syn- of one, theother, to C as one of the descendants onymyseems obvious,but how closely reference ancestor. ifthey are notiden- ofthe specified must theycorrespond Certain differences among competing are to be used in tical?And whatcriteria render hypotheses questions def- phylogenetic thisissue?Undertraditional deciding unresolvable. Suppose, using initionsof taxon names,the answersto ofsynonymy In the previousexample,thattwo different arenotstraightforward. thesequestions provide names are definedunder the alternative definitions contrast, phylogenetic hypotheses (Fig. 5b, c): one criterion phylogenetic and unambiguous a nonarbitrary from the mostreif forthe clade stemming namesare synonymous forsynonymy: ancestor ofA, B,C, D, and E to clades stemming centcommon and onlyiftheyrefer and theother for thecladestemming from the sameancestor. from common ancestor ofA, B, For example,suppose that (1) "Squa- themostrecent
0 E 0. 0

1990

PHYLOGENETIC DEFINITIONS
A B D E

317

a
A B D E A B D E

FIG. 5. Synonymyin cases of uncertainphylogeneticrelationships.(a) The name of the node-based taxon .stemmingfromthe most recent common ancestor of A and D is synonymouswith the name of the nodebased taxon stemmingfrom the most recent common ancestor of B and E, despite uncertaintyabout the relationshipsof C. (b, c) Names of node-based taxa definedwith respectto the mostrecentcommon ancestors ofthe relationships of A, B, C, D, and E and ofA, B, and C are synonymousunder one of the possible resolutions. of C (b), but not the other (c). Circles indicate ancestorsspecifiedby the definitions.

to one of the two hy- dosauria, and C. According and itsdescendants (a node-based (Fig. definition ofthecrowngroup).Neverthepothesesthe namesare synonymous they less, the implicit meaningof "Amniota," to the alternative 5b), but according one of these as thisnameis most used,is the commonly are not (Fig. 5c). Accepting from thefirst to vertebrate hypothesesand rejectingthe other re- cladestemming but if the possessan egg with an amnion(an aposolves the issue of synonymy, are consideredunresolved morphy-based birelationships definition). In theextant of Amniotaunder these areretained, theissue ota, the content andbothhypotheses definitions is identical;howcannot be resolved. This alternative of synonymy are also considorganisms situation should be viewed as a strength ever,if extinct def- ered,thenthesetwodefinitions mayrefer ofphylogenetic than a weakness rather one more inclusive than can todifferent clades, thatsynonymy initions. It emphasizes In other of therel- theother. theamniote egg words, be resolvedonlyin thecontext mayhave arisenpriorto the mostrecent information. evantphylogenetic Although the implicit phylogenetic commonancestorof extantamniotes.A clear stem-based ofthe"total"group definition oftaxonnamesis generally meaning are consid- (Jefferies, 1979)is also possible;forexamwhen only extantorganisms de(1982, 1983) implicitly often arisewhenconsid- ple, Gardiner ambiguities ered, and and Rosen, fined"Amniota"as Recentamniotes ering fossils(e.g., Patterson that are morecloselyrelated 1979; Hennig, 1981; Lau- all organisms 1977; Jefferies, is no dis- to themthan to extantamphibians.Rethere 1989).Forexample, terbach, of"Amniota" ofwhichdefinition about the content of Amniota gardless agreement illusthedifferent are consid- is favored, possibilities when only extantorganisms itis usually def- trate a general thephylogenetic point:although ered.In thiscontext, a seriesofancestors as be stated as the possibleto identify of"Amniota" inition might associated withthename most recentcommonancestorof Mam- beingimplicitly taxon(e.g.,Fig.4), thepreand Lepi- ofa traditional malia, Chelonia, Archosauria,

318

SYSTEMATIC ZOOLOGY

VOL.

39

cise ancestorin this series oftenwill be Theproblem ofchanging thecladeswith ambiguous.Furthermore, althoughthere whichnamesare associatedis aggravated existsan awarenessof these distinctions, by trying to achieve a phylogenetic taxtraditional practice hasbeentousethesame onomy within thetraditional Linneansysname forboth crownand "total" clades tem.This resultsfrom to elimiattempts (e.g.,Jefferies, 1979;Hennig,1981;Lauter- nate paraphyletic taxa, which are often bach,1989).Giventheambiguity resulting accompanied by changesin ranksassocifrom associating thesamenamewithclades atedwithparticular namesin orderto satstemming from different ancestors of mandatory in a lin- isfythe convention categothetaxaAgamidae* eage,itseemspreferable togivea different ries.Forexample, and nameto thecrownclade on theone hand Chamaeleonidaehave traditionally been therank offamily, assigned and the "total"clade on the other. and ithasbeen As can be seen from the preceding dis- suggestedthatAgamidae*may be parawith respect to Chamaeleonidae cussion,therenecessarily will be ambi- phyletic guitiesabout synonymy duringthe tran- (Estes et al., 1988; Frostand Etheridge, sitionto a system areas illustrated based on phylogenetic 1989).Iftherelationships definitions. Associationsof names with in Figure6, then"Agamidae"is paraphyancestors have heretofore been leticaccording to itstraditional particular definition. rather thanexplicit, one might interpret implicit and theforms Nevertheless, this ofdefinition in use often an implicit currently render nameas having association with such associations ambiguous.This situa- ancestor A, the mostrecentcommonanofitsincludedsubtaxa tionshouldencourage the redefinition of cestor (u-w). "Chataxon in explicit names phylogenetic terms; maeleonidae,"on the otherhand, is imhowever, due consideration shouldbe giv- plicitly associated withancestor C. en to diversecriteria, includingimplicit Such being the case, uniting"Agamiassociationswith ancestors, current and dae" and Chamaeleonidaeinto a single historical wouldeliminate a paraphyletic taxusage,alternative names,prior- family ity(see below),and utility to the greatest on. If thisis done, thenaccording to tranumber of biologists. ditional,rank-basednotions of priority (e.g., ICZN, 1985),the name of the clade Priority from A will be "Chaancestor stemming of phylogenetic def- maeleonidae,"because this is the older The consequences initions forsynonymy bearon another is- namein the family group(Frost and Eth1989).Giventhedesirability resolvedin a eridge, ofstawhichis often sue, priority, of names,this meanings phylogenetically unsatisfactory wayunder ble phylogenetic of In the actionhas theundesirable consequence traditional taxonomic conventions. of a name from traditional system (e.g.,ICZN, 1985),pri- changingthe association inclusive one.Thename is tied to categorical ranks.This sit- onecladetoa more ority uationcan lead to drastic changesin the "Chamaeleonidae,"which was formerly withthe clade stemming ofnames(i.e.,theancestors with associated from meanings whichthenamesareimplicitly associated) ancestorC, is now associatedwith that fromancestorA, the ancestor as the resultof simple changesin rank. stemming severaltaxaof the same withwhich"Agamidae"was formerly asThus,combining rank into a single one requiresthatthe sociated. It might be arguedthatthe association oldestnameis thevalidone for that taxon. can resultin changingthe of "Agamidae"withancestor A is unjusThis practice withwhichthenameis implicitly tified-thatit is preferable to consider ancestor associated tothat ofa more inclusive clade. "Agamidae,"the name of a paraphyletic a singletaxonof a taxon,to have had no implicitphylogeSimilarly, partitioning This does not,however, rank intoseveral same neticdefinition. taxaatthat particular in restricting the origi- solve the problemof the change in the rankoften results "Chamaeleonidae" nal taxon'snameto a less inclusiveclade. cladewithwhich, is as-

1990

PHYLOGENETIC DEFINITIONS

319

Acrodonta sociated.Furthermore, the name "Acrodonta"(Cope, 1864)already has been used "Agamidae" Chamaeleonidae forthe clade stemming fromancestor A v w x z y U_ (Esteset al., 1988).Acrodonta, however, is considered tobe assigneda rank abovethe familylevel in traditionaltaxonomies. Therefore, "Acrodonta" wouldbe replaced by"Chamaeleonidae," simply becausethe former is nota nameof the family group. Fromthe preceding example,it can be seen thatthetraditional practice of determiningpriority according to ranksis obfrom fuscatory the perspective of phylo- FIG. 6. Changes in the clade with which a name genetic taxonomy. Theassociations ofnames is associated under the traditional,rank-basedcriteIf "Agamidae" and Chamaeleonidae of priority. withparticular cladeschangeas theresult rion are combined into a single taxon assigned the rank of arbitrary decisions about categorical of family,the name of this familyis "Chamaeleoniand namesappliedtomonophyletic dae" because this is the older name in the family ranks, entities are rejected simplybecause they group. The result is a change in association of the "Chamaeleonidae" from the clade stemming are notoftheappropriate rank. Giventhat name fromancestorC to that stemmingfromancestor A. a rank-based conceptof priority confuses the associations of nameswith particular clades,a phylogenetic conceptof priority 1985), which deals with priority of aushouldnotbe based on ranks. rather is thorship thanof the nameitself, In phylogenetic taxonomy, the concept another rulethatsettles isnomenclatural ofpriority shouldnotbe basedon first use sues of priority in a phylogenetically unofa nameata givenrank orgroup ofranks, satisfactory way. According to the princibut on first use of a name in association ple of coordination, all namesin a given witha particular clade (ancestor). famgroup(e.g.,genusgroup, Thus,to rank-based continue withthe same example, to the first author "Agam- ily group)are credited idae" and "Chamaeleonidae"would re- to use anynamein thegroupbased on the main associatedwith the same ancestors same nomenclatural type.Thus,although (A and C, respectively). Iftherelationships Cope (1886)coinedthename"Iguaninae" are as illustrated (Fig. 6), then Chamae- fora particular taxonofsquamate reptiles, leonidaewould be considered a subgroup hisinsight is credited toBell (1825)merely of Agamidaerather thana separatetaxon becauseBell coined"Iguanidae"and both of the same rank.The namesare notsyn- names belong to the familygroup (de and the former onymous, of Queiroz, 1987). This rule would be un"paraphyly" Agamidaeis seen as a mistakeabout its problematical if it applied only to situacontent. because the name tionsinwhichthedifferences Alternatively, in thenames "Acrodonta" has already beenused for the involved only differences in categorical moreinclusive a name ranks and their associated suffixes, for clade,"Agamidae," associatedwith a paraphyletic example, formerly ifthefamily Iguanidaehad been group,mightbe. abandoned in favorof loweredin its entirety to subfamily rank. "Acrodonta" etal., 1988).Ifthename The rule applies, however,even if the (Estes "Acrodonta" did notexist, be pref- names referto different itmight forexentities, erableto coin a new name forthisclade ample,when"Iguanidae"refers to a more rather thanusing"Agamidae." Anyone of inclusive taxon than does "Iguaninae." these alternatives is reasonable from a phy- Fromthe viewpointof conveying an aulogenetic perspective, because none of thor's association ofa name witha particthemchanges the associations of names ular taxon, this rule confuses matters, withparticular clades (ancestors). whetherin phylogenetic or in Linnean The principleof coordination It reflects an overemphasis on (ICZN, taxonomy.

320

SYSTEMATIC ZOOLOGY

VOL.

39

on earlierdrafts of this ranks. As in thecaseofpriority, viewerfortheircomments categorical was supported bya TiltonPostauthorship in phylogenetic taxonomy paper.Thisresearch from theCalifornia of Fellowship Academy use of a doctoral should not be based on the first Sciences to K.deQ.and National ScienceFoundation but grant group, rank-based namein a particular BSR87-09455 to J.G. orexplicit ofa name association on implicit clade (ancestor). witha particular
CONCLUSION REFERENCES

Since Hennig's(e.g., 1965,1966) influ- atization of organisms on the basis of their phyential writings and theirpopularization, logenesis. JohnWileyand Sons, Chichester.340 pp. J. 1982; Classes and cladists. Syst.Zool., 31: has enjoyedcon- BEATTY, systematics phylogenetic 25-34. progress. BECKNER, and madegreat siderable success M. 1959. The biological way of thought. taxonomy, that Columbia Univ. Press, New York. [Reprinted by In contrast, phylogenetic withrep- Univ. California Press, Berkeley,1968. 200 pp.] branch ofsystematics concerned by BELL,T. 1825. On a new genus of Iguanidae. Zool. uncovered the relationships resenting J.,2:204-207. has laggedbehind. BERNIER, analysis, phylogenetic R. 1984. The species as an individual: Facof phylogenetic taxon- ing essentialism.Syst. Zool., 33:460-469. The development in partby theim- BRUNDIN, omyhas been hindered L. 1966. Transantarctic relationships and their significance, as evidenced by chironomid thatitis tobe assumption plicitor explicit Linnean midges, with a monograph of the subfamiliesPoachieved withinthe traditional and Aphroteniinae and the austral This assumptionis one ex- donominae framework. Heptagyliae. K. Sven. Vetenskapsakad. Handl., 11: in which 1-471. ampleofa generalphenomenon R.,AND D. L. HULL. 1966. The logical structure are takenfor BUCK, practices existing taxonomic of the Linnaean hierarchy.Syst. Zool., 15:97-111. worldview granted and theevolutionary R., AND D. L. HULL. 1969. Reply to Gregg. in- BUCK, is overlaidupon themas a superficial Syst. Zool., 18:354-357. But if the DarwinianRevo- COPE, E. D. 1864. On the charactersof the higher terpretation. lutionis ever to occur in biologicaltax- groups ofReptiliaSquamata-And especially ofthe 1978;de Queiroz,1988; Diploglossa. Proc. Acad. Nat. Sci. Phila., 16:224onomy(Patterson, 231. O'Hara, 1988),thenthe role of the prin- COPE, E. D. 1886. On the species of Iguaninae. Proc. ciple of descent must change. It must Am. Philos. Soc., 23:261-271. interpreta-DARWIN, change froman after-the-fact C. 1859. On the origin of species by means tenet from whichtheprin- of naturalselection. JohnMurray,London. 513 pp. tiontoa central [Reprinted by Harvard Univ. Press, Cambridge, are deciples and methodsof taxonomy 1974.] duced. The principle of phylogenetic DE Massachusetts, K. 1987. Phylogenetic systematicsof QUEIROZ, this change. Pre- iguanine lizards. A comparativeosteological study. definitions exemplifies tobe defined Univ. Calif. Publ. Zool., 118:1-203. taxawereconsidered viously, and the Darwinian after- DE QUEIROZ, K. 1988. Systematics and only interpreted by characters of evolution.Under Revolution. Philos. Sci., 55:238-259. the-fact as products QUEIROZ, K., AND M. J. DONOGHUE. 1988. Phythe principle of phylogenetic definitions,DE logenetic systematics and the species problem.Claenterdirectly distics,4:317-338. considerations evolutionary as DE QUEIROZ, K., AND M. J. DONOGHUE. 1990. Phyoftheir names.Just intothedefinitions or Nelson's version of cladisthe logenetic systematics theconcept ofsynapomorphy granted tics. Cladistics,6:61-75. role DONOGHUE, ofcommon descent a central principle M. J. 1985. A critique of the biological in systematic analysis,so the conceptof species concept and recommendationsfor a phythis con- logenetic alternative.Bryologist,88:172-181. definitions grants phylogenetic ELDREDGE, N., ANDJ.CRACRAFT. 1980. Phylogenetic role in taxonomy. cepta central
ACKNOWLEDGMENTS We thankD. Cannatella, M. Donoghue, R. Estes,R. Etheridge,S. Fink,D. Frost,M. Ghiselin,A. Graybeal, D. Hillis, D. Hull, P. Mabee, R. O'Hara, C. Patterson, E. Sober, E. Wiley, A. Wyss, and an anonymous repatternsand the evolutionary process. Columbia Univ. Press, New York. 349 pp. ESTES, R., K. DE QUEIROZ, AND J. A. GAUTHIER.1988. Phylogeneticrelationshipswithin Squamata. Pages 119-281 in Phylogeneticrelationshipsof the lizard families(R. Estes and G. K. Pregill, eds.). Stanford Univ. Press, Stanford,California.

Ax, P. 1987. The phylogeneticsystem.The system-

1990

PHYLOGENETIC DEFINITIONS

321

J.S. 1974. Formal definitionsof paraphyly FARRIs, and polyphyly.Syst. Zool., 23:548-554. D. R.,AND R. ETHERIDGE. 1989. A phylogenetic FROST, analysisand taxonomyofiguanian lizards (Reptilia: Squamata). Misc. Publ. Univ. Kansas Mus. Nat. Hist., 81:1-65. GARDiNER,B. G. 1982. Tetrapod classification. Zool. J.Linn. Soc., 74:207-232. B. G. 1983. Gnathostomevertebrae and GARDINER, the classificationof the Amphibia. Zool. J. Linn. Soc., 79:1-59. GAUTHIER,J. 1986. Saurischian monophylyand the origin of birds. Pages 1-55 in The origin of birds and the evolution of flight(K. Padian, ed.). Calif. Acad. Sci. Mem., 8. GAUTHIER, J., R. ESTES, AND K. DE QUEIROZ. 1988a. A phylogeneticanalysisof Lepidosauromorpha.Pages 15-98 in Phylogenetic relationships of the lizard families (R. Estes and G. K. Pregill, eds.). Stanford Univ. Press, Stanford,California. A. G. KLUGE,AND T. ROwE. 1988b. AmGAUTHIER,J.,
GHISELIN, M. T. 1966a. An application of the theory GHISELIN, M. T. 1966b. On psychologismin the logic GHISELIN, M. T. 1969. The triumphof the Darwinian

niotephylogeny and theimportance offossils. Cla4:105-209. distics,

of definitions to systematic principles. Syst. Zool., 15:127-130.

oftaxonomic controversies. Syst. Zool.,15:207-215. method. Univ.California Press, Berkeley. [Reprint1984.287pp.] ed byUniv.ChicagoPress, Chicago,

GHISELIN, M. T. 1974. A radical solution to the spe-

GHISELIN, M. T. 1981. Categories,life,and thinking. GHISELIN, M. T. 1984. "Definition,""character,"and GHISELIN, M. T. 1985. Narrow approaches to phyGOULD, S. J. 1985. To be a platypus. Nat. Hist., 94:

cies problem. Syst. Zool.,23:536-544. Behav.BrainSci.,4:269-313.

other equivocalterms. Syst. Zool.,33:104-110.

logeny:A reviewof nine books of cladism.Oxf. Surv.Evol.Biol.,1:209-222. 10-15.

GREGG, J.R. 1950. Taxonomy, languageand reality.

G. C. D. 1974. On thefoundations ofbiGRIFFITHS,


HENNIG, W.

Am.Nat.,74:419-435.

ActaBiotheor., 23:85-131. ologicalsystematics. Annu. 1965. Phylogenetic systematics. Rev.Entomol., 10:97-116. Univ. HENNIG, W. 1966. Phylogenetic systematics. IllinoisPress, Urbana.263 pp. W. 1975. "Cladistic ofcladistic clasHENNIG, analysis A replyto Ernst sification?": Mayr. Syst. Zool.,24: 244-256. and HENNIG, W. 1981. Insect phylogeny. John Wiley 514 pp. Sons,Chichester. ofessentialism on taxHULL,D. L. 1965. The effect Phiofstasis onomy-Twothousand years (I). Br.J. los. Sci.,15:314-326. ofbiological science. HULL,D. L. 1974. Philosophy 148 New Jersey. Prentice-Hall, EnglewoodCliffs,
PP. HULL, D. L. 1976. Are speciesreallyindividuals?

Syst. Zool.,25:174-191.

HULL,D. L. 1978. A matterof individuality.Philos. Sci., 45:335-360. ON ZOOLOGICAL COMMISSION ICZN (INTERNATIONAL NOMENCLATURE). 1985. Internationalcode of zoological nomenclature.Third edition. International Trust for Zoological Nomenclature, London. 338 PP. R. P. S. 1979. The origin of chordates-A JEFFERIES, methodological essay. Pages 443-477 in The origin of major invertebrategroups (M. R. House, ed.). Academic Press, London. D. B. 1983. Can baptismalone save a species? KITTs, Syst. Zool., 32:27-33. K. E. 1989. Das Pan-MonophylumLAUTERBACH, furdie Praxisder phylogenetischen Ein Hilfsmittel Zool. Anz., 223:139-156. Systematik. L0VTRUP, S. 1986. On the existence and definition of taxa. Riv. Biol., 79:265-268. MUIR,J.W. 1968. The definitionof taxa. Syst.Zool., 17:345. G. 1971. "Cladism" as a philosophy of clasNELSON, sification.Syst. Zool., 20:373-376. NELSON,G. 1974. Darwin-Hennig classification:A reply to ErnstMayr. Syst. Zool., 23:452-458. O'HARA,R. J. 1988. Homage to Clio, or, toward an historical biology.Syst. philosophyforevolutionary Zool., 37:142-155. in systematics. Syst. C. 1978. Verifiability PATTERSON, Zool., 27:218-222. C., AND D. E. ROSEN. 1977. Review of PATTERSON, and otherMesozoic teleostfishes ichthyodectiform and the theoryand practice of classifyingfossils. Bull. Am. Mus. Nat. Hist., 158:81-172. N. I. 1978. Gaps and prediction in clasPLATNICK, sification.Syst. Zool., 27:472-474. N. I. 1979. Philosophy and the transforPLATNICK, mation of cladistics. Syst. Zool., 28:537-546. RIDLEY, M. 1986. Evolution and classification.The reformation of cladism. Longman, London. 201 pp. D. E. 1974. Cladism or gradism?:A reply to ROSEN, ErnstMayr. Syst. Zool., 23:446-451. ROwE,T. 1987. Definitionand diagnosis in the phylogenetic system.Syst. Zool., 36:208-211. RoWE,T. 1988. Definition,diagnosis, and origin of Mammalia. J.Vertebr.Paleontol., 8:241-264. M. E. 1973. The philosophy ofbiology. HutchRUSE, inson Univ. Library,London. 231 pp. SNEATH, P. H. A. 1962. The constructionof taxonomic groups. Pages 289-332 in Microbial classification (G. C. Ainsworthand P. H. A. Sneath, eds.). 12thSymposium of the Society forGeneral Microbiology. Cambridge Univ. Press, Cambridge. E. 1988. Reconstructing the past. Parsimony, SOBER, evolution, and inference. MIT Press, Cambridge, Massachusetts.265 pp. STEVENS, P. 1984. Metaphors and typology in the 1690-1960,or developmentofbotanical systematics the art of puttingnew wine in old bottles. Taxon, 33:169-211. F. 1974. Some philosophical problems in biSUPPE, ological speciationand taxonomy.Pages 190-243 in of knowledge Conceptual basis of the classification

322

SYSTEMATIC ZOOLOGY

VOL. 39

(J.A. Wojciechowski, ed.). VerlagDokumentation, WILEY,E. 0. 1981. Phylogenetics. The theory and practice of phylogenetic Pullach/Miinchen. systematics. JohnWiley F. 1989. The semantic oftheories SUPPE, conception and Sons,New York.439 pp. and scientific realism. Univ.IllinoisPress, Urbana. WILEY, E. 0. 1989. Kinds, individuals, and theories. 475 pp. Pages 289-300inWhatthephilosophy of biology THULBORN, R. A. 1984. The avian relationships of is (M. Ruse, ed.).Kluwer Academic Publishers, DorArchaeopteryx, and theorigin Zool.J. Linn. ofbirds. drecht. Soc.,82:119-158. 1990; accepted25 June1990 WILEY, E. 0. 1979. An annotated Linnaeanhierar- Received13 February chy, with comments on natural taxaandcompeting systems. Syst. Zool.,28:308-337.

S-ar putea să vă placă și