PRINCIPLES OF SYSTEMATIC” sues al aspects of ving tionships among them” the stmt sts ll ape aera nds ae ete meninges whoa a sound cass ce eso apes oat, the tro vl, ant Hi ‘ation uns a ahead anes hey are based on sound csi se Gaetan ue prea Inport in aes bole sae The trop of ths portance expe why even oy cei oes are dected tte ak of mproving the less ton of smal HAPTER, TAXONOMIC CHARACTERS ‘Taxonomic characters provide the evidence from which relationship be teen taxa is inferred. The more characters two taxa have in common andthe more similar they appear to the taxonomist, the more closely re Tated they are considered to be (Chapter 6). In taxonomic practice, how ver, iis actully more productive to look for diferenes between tax. ‘The definition of a taxonomic character i bused on this experience: A Laxonomic characteris any atribute by which a member ofa taxon dif fers or may der from a member of another taxon ‘Any difference between two individuals isa characte, but not all characters are useful for taxonomic purposes. Knowing where useful ‘characters ae 0 be found and estahishing their specifi value is perhaps the most important skill of systematist, requiring not only good the ‘oretcal knowledge but experience a well Defining a taxonomic character as “any attribute ofan organism’ is incorrect. Differences between phena (Chapter 4) are not taxonomic characters, Therefore, featires by which individuals ofthe same popula tion difer—that is, diferences between sexes, age classes, and other polymorphisms—are not taxonomic character. However, uliferences hexween corresponding phena of dillerent species ae taxonomic cha (CHARACTERS AND SIGNIFIERS ‘The word character isnot the exclusive property of system ters ina type alphabet ae called characters, andthe moral160. rants: sucrorasonowy person constitute that person's character, The word's meaning in syS- Tomatics ie derived ffom one ofits most common uses: A character isa Jropery. attribute, or feature that distinguishes one thing, individual, oF rou from another, Black feathers are a character of Australian swans, id white feathers area character of European swans. This use of the word predates Linnaeus ‘the late 19508, numerical taxonomists encountered a dificulty with the word. To be handled by a computer all characters must be assem= ‘bled into a able called data matrix (Chapter 11). Typically the columns pr the table ae for tana, and each ro of cells conan the numerical Codes for characters, The problem was what 10 call the Tow. ‘To solve tis problem, these workers unfortunately transfered the meaning of the term character from the difference between t8xa (6.8. Dac vessus white feathers in svans) 0 the structure that varies (@. feather colon i order to havea term to designate the rows uations. The datum in each cell, trtionally a characte ited a character stale. AS ‘explained by Sokal and Michener (95ka:1410: ‘Our se ofthe word hart” wil ee some elaboration. its com ea Sonam wage a carctr iv any eae of ne kind of orn ‘Barreca another Rind. Thus there abJomen of oe ee is aa rating i fom another Boe wih te aiomen Bac. 19 chs horace jo the werd in a second connotation ony hat a «fate ARS) ares tm one king of organi 1 ether. Now; 10 se the above CSc thomna color the characte, which oct ia two “sates” oF “ermtives red and Wack. ‘This proposal had the unfortunate result that the term character is row used in taxonomy in two Very different senses, In computer taxon ony its used inthe transferred sense of Sokal and Michener, while in ‘rafnary taxonomic discussions itis used inthe classical sense of the shord, This equivocation has caused considerable confusion inthe taXo~ homie erature. ‘Problems caused by the dual use ofthis word are frequently found in studtes done by numerical taxonomists. For example, in phenetcs {(Ghopters # and 11), some formulas used to measure difference between arn produce reslis that also reect how many actual characters are tived fn study. This makes it ifcult to compare a study using few chr- Stee with a stody using many. Pheneticsts sugested that each similar ip value be divided by the numberof characters” in each study to find 10 Meatsge istance. What they intended was to divide bythe numiber of “istint characters, but the correction factor actually employed in their enmulas and computer programs was the numberof variables. The result ‘rMeTER LON CHARACTERS 161 of this confison of ems was that the ney made and supposedly cox parable phenogras were nothing ofthe sorts Numerous ene o ts {ype ean be found inthe recent sumer erature "io end ths conuson Astlock (1985) proposed restoring the term character wo its tational mesnng and introducing ae ter, the signifier, for a feature that varies from one organism 10 another. The feather, then, would be the senile sgn bear, and “Tether white an ether ac wl eh shart AS ds Shes oration ofthe tational nang ofthe word vhracier woul be, the transferred meaning hasbeen so aly exis ster 30 year hal fur constants song aed against the option of the em Senter Ths character, atc in ear o numerical method fers to a variable feature, and character state is defined as an attribute b) which a member of a taxon differs from a member of another taxon, ze KINDS OF CHARACTERS. Almost ay att oan ris my be sf as nono har fete iri difer from the euvale feature in member of another taton, Howeter prone lasing work sl oly whch aegte atta rom many ssi smn vate conparon Muneuns prove th opr an ths uty te tconoms re fem care that nf ay abevednpeseied spesinen ‘morphological characters) (Table 7-1). d Trench gop of organs, wheter is, butts, se urchins, oF sna ieren axons ches es spr tenn txxonomit Become ui ith th chaste ht se mst ht the prt tao whch tatonomis plans ose, Mono. tar an hanks sal ge eetaled design te has tenth roth oh Aon ss, nce schon 98) dented peso ee dessin sa Mato of the taxonomic characters of that group, : ae Morphological Characters Features of external morpology vary according t kn vary according to kinds of animals “They rane from sch sper etre as cil features as plumage ad plage char ters of tits and mamas tugh scale counts of fishes Sn fp 0 the highly conservative and piylogenetialsignfant sites and sees he artiropd ot nal anatny oies many ao omic characters in practical al poop of high animal The exer to which sh characters are wed routinely vares fom group to up, senealy in inverse proportion othe abunance and useunen feat162 pants: wncnoraouony FINDS OF TAXONOHIG CHARACTERS 1 Merogel cece Gora ena ron, 3 Socal stuctres o-oo) © itera mrny ny) 4 Emeooar {Kandy tn er ote onaceess 2 Prone ros ao nce © Gene ty crs rans estns € Arar an eave tone 4 Dn neton {Oe mouse ctrecas «Seton cern 1 Couthip and cer agen teg mechs 8 One acho pao 5 asa onto 6 Geopan recrs «Geral boectnic bton pare 1 Sportage elton opera ets execute nthe perso fm i te Stoic oul prose so ut cleion, whe rept ans an She nna seed alsa ae {ia aul rset, Or the whe sets femal ana ery tach orien of hier re en Akan eyo for dnctinfton he sg eve Fon com tence pened bd hess ma Feestnce hy son rey fet hon int anal are gat ices have Ben made in een deta Desrptns have bene mre setae and heer sana 12a Cuca mlsescons an lover vere reve ‘Sandie of cua ven sch sen ones forms enmtdc.The development now sner meetin hme as revealed wealth of characters even among protozoans, particularly cl iates. Scanning electron microscopy has added enormously to our knowl. ‘edge ofthe morphology of insects, arachnids, and other small organisms ‘New organs and strictures are steadily added to those which show. taxonomically important differences. The spermatozoa of many ta instance, have a highly peculiar an specie morphology and may scr¥e 8 useful indicators of relationship. Hard Parts and the Work of Animals It woul be senseless 10 worry bout whether the items referred to as hard parts (shells, external shee tons, ec.) are morphological, physiological, or behavioral characters ‘Much of the classification of invertebrates is based om characters of exoskeletons and shells Similaly, among the protozoan, texts, shells, thecal plates, eysts, and other hard pars are vital in the classification of foraminierans,radolrians,testaceaus rhizopods,Rlaglltes, nd er organisms. The orientation types of calcite crystals im the skeletons of ‘echinoderms agree well with thei clasiiestion in families and orders ‘aup 1962) ‘The description of many taxa of dinosaurs has been based on fossil tracks. Inthe clasifcation of gall insects, the gall sometimes yields 00d a clue to relationship as do the insects themselves. The form of the ‘mines isan important taxonomic character in mining insets, andi even sheds light on their history since these mines are sometimes well pre served in fossil leaves (Figure 7-1). However, since 1930 it has not been Permissible to base the name of new species exclusively “on the work of ‘an animal” (Articles 126 (8) and 2 (i) ofthe International Code of Z00- logical Nomenclature] Coloration Color pattern and other aspects of coloration are among the most easily recognized and thus most convenient characters in r= tin groups of animals. Every species of bird can be recognized by i coloration except fora few genera with sibling species (es. Collocola Empidonas). The same is teve of certain ret fishes and battrtes, Even Where coloration is not completely diagnostic, it often helps nairow down drastically the number of species to be considered. In groups in which subspecies ae recognized routinely, suchas mammal, bids, bl. terfes, and some wasps. colot agin plays an important role: many subspecies are identified entirely onthe basis of coloration, ‘The quality of colo is nt easy to describe in words. In orignal revi- sionary work, therefore, iis preferable not to ely on descriptions but fo base one’s judgment on the comparison of specimens if possible, There are, however, various ways in which greater precision can be given tO ‘color determination and description (Chapter 12).SO party mand erin tego tera SRE seer apes meteor slaraaltaahansa nesters ane a Genltalic Structures For reasons that are not yet fully understood (aay 1963:108; Eberhard 1985), the genitalia of many animals, patio feny arthropods, not only show a great deal of stractral detail bat are also highly speciesspeciic (Figure 7-2). Because they are three Simensional strctures, genitalia have to be carefully prepared to be sinely comparable. In many groups of insects and spiders genital ructares ere more important for spocies diagnosis than is any other Character, However, even here it has been found that a single species ay have a good Jel of variation or that two related species may have Omran: TateNeMe RATERS 185 ure 72 ‘pel iromaecs ng Ovo ete ra edo ct ae indistinguishable strictures. In most vertebrates the genitalia are sot, but the gonopodium of some fishes, the hemipenis of snakes, and the ‘haculum of mammals may supply good taxonomic information, ‘Other Characters Morphological characters of adult specimens are still used more frequently than are any others, bu they are supplemented. oan increasing extent by other characters, as listed and discussed below (GBlir 1962; Muneoe 1960), This is particularly true for “ical” spe cies, genera, and families in which the evidence from morphology has been equivocal or contradictory. The increasing uilization of new char- sctrs is justified because (1) morphology refccs only part of the gen0- type and may not reflect genetic relationship accurately, (2) morphology incertain taxa doesnot supply sufiient characters, and (3) any charac ter may be misleading because of special adaptations, The introduction of new kinds of taxonomic characters has been one feature of the so-called new systematics. These characters (molecules, chromosomes, behavior, .) supplement but donot displace the use of morphological characters, ‘A particularly important reason forthe utilization of new characters that they serve asa check on the conventional morphological characters, When discrepancies occur between a classification tat fs based on mar- phology and one that isnot, sll other sets of taxonomic characters must, bbe used. Fortunately, he newer characters usally confim classics166 penta: icnoraKencun tions based on morphological characters. It seems that general morphol fy usually reflects large part ofthe genotype and thus generally per ‘its reliable eonclsions on relationships, Larval Stages and Embryology Various immature or larval stages, ‘embryos, and sometimes even ests may provide taxonomic information. ‘The various sibling species ofthe Anopheles macallpennis complex (mo laria mosquitoes) were discovered because of dilerenee in exe stv ture The classification ofthe Aleyroddae (whites) i based primarily ‘on the pupae. Discovery ofthe leptocephaustike larval stage of the fish ‘onder Heteromi confimed the previously suspected relationship with the ls Tn groups with complete metamorphosis. entirely diferent sets of characters evolve in larvae and adults, and conclusions drawn from the ‘haraters of one stage form a very useful check on conclusions drawn feom those of another. This was shovtn for fogs (Anura) by Orton (1957) find Inger (1967). A careful comptison of larval and adult characters of tigger wasps (Sphecidae) permitted Evans (195) to show that some fe tures ofthe adule structure (emarginate eyes, los of wing veins) had been fvervalved. A proper evaluation of lvl characters ll to an improved {lassiicaton. Sif findings have been made in other groups of insects {Wan Emden 1957. Larval and adult character are visible manifestations bt the same renonpe. There can be diferent identification schemes for Itevae and adults, but there can be only one classifeation for a given ‘group of organist, and this classification must be based on the proper tveighting of both adult and larval characters. Sometimes iis the adults that have acquired specialized adaptations, and sometimes iti the lat ‘ae: only 2 biological analysis can lead to the proper evaluation, Ta ‘groups such as the sponges (Prifera), which show extremely few mor- Dhological characters, consideration of the embryology has been a great Felpin classification (Levi 1986) Genetic Characters ‘There has been much confusion about the mean ing ofthe term genetic characters. Because, breuy speaking all cha oles except nongenetic modifications of the phenotype are genetic Characters the term is virally Meaningless, Isolating mechanisms and ‘monogenic characters have sometimes been designated as genetic char- acters, but this restition i unjustified. Sterity. On the whole, only closely related species ean hybridize sucessfully n groups of related forms the presence or absence of ross Stenlity provides important information, bu it must be employed care~ fully. Ducks, fr instance, show much crossferiity not only of species ‘OmTER?: TaxeNome KARACrERS 167 but even of related genera, yt crosses of two such closely related spe cies as the wood duck (Aix sponsa) and. the mandarin duck ‘elericnlata) are completely sterile a & result of chromosomal rear Fangement. In birds there is much fetity among closely eelated species, but itis hard to generalize about insects. Geographically remote popula: tions of the same species may be nearly intersteile in some insect groups, while in other groups some congeneric species ae fully Fertile. A few exceptions notwithstanding, the relative degree of fertility in species ‘crosses ison the whole a very sensitive measure of relationship. Dubois (W982) suggested that any two species whose genotypes are sills ‘tly similar to permit the production of hybrids (stele or not} ought to be inched in a single genus. Conversely, the inability to produce hy- bids does not. ofcourse, constitute proot of generic difference Chromosomes Botanists have made use of the abundant information provided by chromosome patterns far longer than have zoologist. In provements in cytological fechnigues over the past several decades now permit chromosomal stidies even in such diffeultsroups ay mammals, Birds, lepidopterans, and other groups with many small chromosomes tare hander to work with than are the Orthoptera with their few, lage ‘chromosomes and those groups of Diptera in which giant salivary chro ‘maxomes are especialy suitable for eyitaxonomie studies. White (1973, 1978) provided useful summaries, but they have to be continuously up dated because of recent advances. A series, Animal Crogenetles, cited by John (1974 et seq.), includes several fascicles which review recent ‘Work on invertebrate (especially insect) and vertebrate groups. "The taxonomic character most frequently used by early eytogen ctcists was chromosome number. Since then rch aay of additonal ‘chromosomal characteristics has been discovered which can he used 19 ‘determine degrees of relationship. Chromasomes may have single arm ‘owing to the messy terminal position of the centromere (acrocentic ‘thromosomes) orto arms (of equal or more a less unequal length) 0: ing to.a more central poston ofthe ceriromere (metaceniri chromo somes); allematively, they may have a difse centromere (holoeattic ‘hromosoetes), a condition widespread among the Hemiptera and afew bother orders ofinsects, The number of chromosome arms (nombre fonda Inental) in a group soften more constant than isthe rato of aroveateie {to metacentrie chromosomes; this due tothe frequency of Robertson jan fusion and fssions, in which two acrocenric rod chromosomes are fused into one metacentri one or the reverse. Such fusions and Rssions may be reversed, resulting in homoplas. ‘Species and higher taxa of animals may difer by an extraordinary umber of chromosomal phenomena, ineliding paracentric inversions,168. penta: kcrorasonow pericentric inversions, translocations, Robertsonian changes, sex [eterminng mechanisms, the presence of diflse or concentrated hetero- ‘Chromatin and the presence of supermumerary chromosomes (White 1973). The fine structure of ehyomosomes can be brought out BY ePPro- priate staning methoxls that show G bands, C bands, or Q bands. Not Bly do these banding patterns reveal a wealth of deal, they ae also highly conservative ia some groups. convincingly docamenting relaton- Ship Baker. Quinsich, nd Hood 1987), Blekham and Carr 1983) have ‘demonstrated the importance of chromosomal information forthe deters ‘nination of url phylogeny. Robertsonian changes sions and fsions) fre not necessarily unigue events in phylogeny. Even this chromosomal Phenomenon is subject to homoplasy “The phylogeny of horses, aves, and zebras (genus Eqs) has been realy cared by means of & comparison of their chromosomes. The Scndrogram of the karyotypes agrees remarkably well with the dendrogram based on morphology, What no one understands, however, fe'why there are such drastic numerical diferences among closest rela- fives ia this axon. The chromosomal number 2n) ranges from 66 in [Baus praewalsiy to 32 in E, zebra. E. przewalsyt has 40 acrocentic find 24 metacenre atosomes (ai apa of sex chromosomes), while oionus onager has only 8 acrocenrics and 46 metacentrics. Fusions lind translocations must have Been rampant within this gents. By Con- tena inthe camel family the (wo old world species and the four South ‘Rimrican species (lama, ef.) have 74 very similar chromosomes, do: ‘menting great chromosomal conservatism. ln namtrous orders nd fa- ties of animals, for instance, bas and turtles he study of chromosomes fas led ta considerable improvement in our understanding of relation ‘hips Te greatest value of chromosomal information is tha it can serve Estrcheck on morphological, molecular, and other types of information. Chromosomes re particularly useful on two different levels. At the lover level the) aid inthe comparison of closely related species. Sibling [pecicn are often far more diferent in chromosomes than they ar i eX {ermal morphology (Mayr 1963, Chapter 3). In higher taxa, chromosomal patterns may be of extreme importance in establishing phylti lines Most chromosomal changes are unigue events: once the new pattem is ‘ota it choractersie ofall descendants ofthe ancestal popu- Tation, Changes in sex determination; earangemens of chromosomes land eniromeres through fusions, fission, oF translocations: and the as- {Gulsion of supernumerates offen supply unequivocal cles to relation Ship. For instance, the simlaity in the spermatogenesis of Mallophaga fan that of Anoplira (tue lice) strongly supports the belie in close re- Tavonshipof these taxa The sant salivary gland chromosomes of certain Curren: ratcnome unerens 169 larval dipterans permit the constuction of very precise dendograms of related species Polyps are in animals compared wi plans, but there ae imerouy ott ways in wich an ineease or decrease in chromosome number may occ, Curiously, tbe mot fequntevolitionry trend i many groupe fom hgh olow’ chromosome number trough chrom soma nn, Even ough cron reeset the sete tena soo re thatthe amount of chromosomal change reflects the ant fe angle nie ay how cone Fangement indeed, many species ate polymorphic for various types of Cheon vagrants However, cc nov ich onigerabie degre of genic chang eno reflected oi nl gh te feted in the chromosomal pater, atin some Hawaian Drosophila (Cason 1987) Physiological Characters This group of characters is hard o define. Al structures are the product of growth that of psiloc! processes, and are thus ulna physological characters. Also, physiology is reputed by enzymes and fther macromolecules and thus isnot separable from biochemical char- acters. Physiological characters generally include growth constants, tem peratue tolerances, and the various other processes studied by the com parative physiologist. Species diflerences are abundant’ in. these jaracters, but since they are nt present in preserved material and their ‘ay eres pel hey earl sed he ‘onomist, Some representative cases are discussed by Maye (196: 40-65), land fuller treatments are given in textbooks of comparative physiology (Prosser and Brown 1973; Schmidt-Nielsen 1979). : Molecular Characters “The major molecular inventions were made by the caries living ies ving organ- isms, Even the most primitive prokaryotes have on the whole the same kinds of macromoleetles and metabolic processes that are found inthe highest animals and plants. Stl there is enormous specificity at every taxonomic evel nd ths spec Bing ieesiney ext by ‘Serology provided th ealest widely used method of comparing pro- teins. This method is based on the principle thatthe proteins of one of- ‘anism will show a stronger antibody reaction tothe proteins ofa closely related organism than they will to those of & more distanily related oF‘ganism, Unfortunatly, this method fas various technical dificlies. ‘Though used for more than 60 years thas not contributed 2s much to & Clavifcation of otherwise ambiguous eases as had been hoped. A sum inary of some ofthe achievements ofthis method can be found in Leone (1968) Asa result of improvements in these techniques, there as re ‘ently been revival of interest in the quantitative study of antigenic e- Selions, An extension ofthis work is the method of microcomplement fixation (Champion et al 1974). The study of bloodgroup genes {immanozenetics) has shed light on relationships among species of i eons (rwin 1947 and later authors of bis school) and has been used in the study of primates. ‘Mich recent work has been devoted to developing a taxonomy of spe- cific chemical components and mcromolecues. The erly technique of {fusion chromatography used capillary action to separate molecules on filter paper. In this method a sample (ga body fuk or macerated tis. sie) i spotted onthe paper, which i dipped ins solvent. Movement of the solvent caries different molecules at diferent rates, so tha diferent Stmples produce diferent patterns of spots thal represent various com ponents of the samples, Difusion chromatography has largely been re Placed by electrophoresis, which uses an electrical potential Instead of Eapillary action and hydrolized starch or polyarylamide instead of fiter paper. Improved resolution has been obtained by processing samples Second time at 0 degrees tothe st run t produce a two-dimensional Separation, Most component ae colorless and must be stained to show Giferences or must be viewed under ultraviolet light, which may be ab Sorbed or cause thorescence. Improvements are constantly being intro ‘oced, and the newest iterature must be consulted forthe Intsttech- higues and instrumentation (Brewer 197; Harris and Hopkinson 1970) "A lage number of new molecular techniques have been inroduced since the early days of serology and elestrophoresis. They make use of proteins, nuclear DNA, mitochondrial DNA, and ribosomal nuclei ae- Hs. A study of the restriction site mapping of ribosomal DNA showed the value ofthis method in establishing telatonshis in the genus Rana (its and Davis 1986) comparison of 188 ribosomal RNA permitted a provisional reclassification of the invertebrate phyla (Field et al. 1988), but subsequent findings led 19 considerable modifications. Ribosomal RNA tas been particulary useful for an understanding of the relationship ofthe lower eukaryoes and the various branches of the prokaryores. Tihe molecular method that provides th sYeaest amount of informs: tion isthe sequencing of proteins and auceic acids. The number of dit ferences in the bose pais of two nuclei acids reflects the number of mu tational steps that Separate them (except fOr tansposons and other insertions), By comparing the sequence of base pirs of two homologous DNAs or RNAs, ene can determine the number of mutational differences (Goodman 1982}. The same technique can be applied to amino acid re placements in proteins, but this provides much less information, The Complete amino acid sequence is now know for many macromolecules (Dayhotl 1973, 1976, 1979), and the degree of relationship of two species can be inferred from the anount of diference. The cose relationship of hhumans and chimpanzees, for instance, i indicated by the Tact tha their hemoglobins and fibrinopeptides are stil identical (King and Wilson 1975), Different macromolecules change at diferent rates—histones very slowly, evtochrome C quite slowly, globins at an average rate, and fibrinopeptides very rapidly. ‘A given macromolecule usually changes ina particular phyetic lineage fata rather constant rate. This observation led Zackerkandl and Paling (4962) t0 propose the idea ofa molecular clock. The clock is calibrated by estimating the age of certain branching points between higher taxa with the help of the fossil recor or biogeographic evidence and then cal culating from these dates and the known difference in the amino acd the average numberof amino acid replacements per 10 mili years. There tilting time frame permis the constriction of dendrogram that depict. the inferred phylogeny (Fitch and Margoiash 1967, ‘Use of molecular evolution permits inferences on the probable time of, the splitting of phylei ines and sometimes even on the Eeosraphy of the splitting event, For instance, the immunological distance of the hemo- Ivmph proteins of the Haviaiian drosophilds from those oftheir Nocth “American relatives is consistent with «colonization of the islnds about 40 milion years ago, that is, before any ofthe currently existing islands had emerged (the older islands are now submerged sea mounts) (ever- ley and Wilson 1985) This confirms the previous resonsriction ofthe history of the Hawaiian biota by Zimmermann (1948) “The assumption of completely constant molecular clock is, of course, unrealistic (Britten 1986). The same molecule may evolve at gute diferent rate in diferent phyletic ies. The less affected by selection certain molecular changes are, the more likely itis that they wll eur a ‘constant rate; this seems to be the case with many enzyme genes, 35 revealed hy electrophoresis (Kimura 1985). Regulatory genes, introns, transposons, and gene families may experience considerable deviations from 4 molecular clock eonsfaney. The molecular clock thus is only a first approximation, and is results must be fine-tuned through the use of additional methods ‘Chronologies can also sometimes be established through an analysis of| particularly stable molecular changes. For instance, in the echinoderm Phylum a certain gene regulatory mechanism (alpha subtype histone mar ternal mRNA} originated in the late Trissi and is now found in all the72 pane, uscnoraxonouy ‘orders of echinoid tat originated after that date (about 200 milion years go) (Raf etal. 198. In doing such calculation one must aways re ‘Momiber not ony that different molecules may evolve at ferent rates but also that within lineage the rate may change drastiealy. For in- Stance neleotide substitutions for seven mammalian proteins seem to hve ccourred i the primates at a distinelly slower rate than as the {ise in most other organisms (Fitch 1976). Its now known that several sesccules change at much as ive times a fast in some phyletic lineages 5 they do in others The eal of the phenetcists has always been to determine the total ‘overall similarity of to taxa. This What the technique of DNA hybrid- ination attempls to achieve atthe molecular level (Sibley and Ahlguist Toss, 1945). Phe method is fairy complex. The DNA is freed of al ro- teinand RNA. an then the purified DNA is sheared into fragments with in average length of 500 base pairs, The DNA is then melted at about Tore to thatthe (wo strands of double-stranded DNA separate, Then the repeated sequences are removed so that only the pieces of a sngle- Cops BNA are let, When the DNA of two species is mixed fone DNA is abel with 3 fadioactive isotope) and the mixture is gradually cooled, Falogots single-stranded pieees of the (wo DNAS wil pair while the ommatebing pieces will remain in solution. This indieates what percent ge of the DNA pairs and what percentage has become sufficient dit {event during evolution fo ar no Tonger. A dela (A) value can be deter fnined which reflects the dogree of diflerence. (This is a greatly [bbrevated necount ofa considerably more elaborate procedure.) Sibley At ist poslulated the rate of DNA evolution (.., nucleotide substi tion) to be the same inal lineages of rts and sated that took about Tullon years of divergence for the genomes of 1Wo species fo become ‘iferent in 1 percent of theie Base pairs On the basis ofthese assump- tions he constructed a dendrogram ofall avian families. The discovery of Considerable inequalities of rates hus necessitated certain moifeations {Briton 1986). Since Sibley’s dendrogram is based ona phenetc method, {ti really a pbenogram, even though iis quite diferent from Sokal ‘overall morphological siilarity. Mitochondrial DNA js of considerable felp inthe comparison of populations ofa single species. for such DNA {sinherte directly from the maternal paren and i not subjeet to sexual Fecombinatio, as is nuclear DNA (Avise and Lansman 1983). Diferent molecular techniques ate useful at different taxonomic ev cls The study of mitochondrial DNA and that of elctrophoretically dis Covered enzyne differences are most useful for the comparison of pop- tations and. closely related species (Barrowelough_ 1983; Patton and Yang 1977, Selander and Whitam 1983), while protein and nucleic acid ‘ourren?: taxcncme ouAnAcERs 173 sequencing, immunological methods, and DNA hybridization are most, informative inthe stdy of italy elated species and higher ak, ha isin the construction of phylogenies. One ofthe fist molecular tee, ha of Fitch and Manolo (1967), asso closet he tonal Phylogenet tre ofthe vertebrates that it made taxonomists are of {he importance of molecular methods, Unfortunatly the methods Tht are most precise and informative, such as protein and DNA sequenci are also the most time-consuming. i e = “ihe thoroughly documented demonstration by Davis (196), based on morphological characters, tha he gan panda abe nd is ot Felted Tore raccoons was dubied hy sf scents il Sach and Bren Confirmed the val of Davis conclusion by wsing several molecular tnathods (Mayr 1986), In similar mane the findings of comparative Snatomists have been confmed again and again in recent years DY mo lecular methods, However, there are exceptions. The cheetah (ACOn) tas always been considered by morphoogits Bethe most aberrant af llth cats, but molecular ties ave vealed that tact ia rather ‘ose relative ofthe longer aro, wth other genera of the ca aly being far more distant (Cairn 0 Bren 1589) "The stay of structural characters bas revealed tat each organ o 0 ‘ean system may have is own specie rate of evotonary change (nc Saicevouton) (Chapter 6. Much evince iniats that this principe anual ald for molecular characters. A comparison ofthe man Peng (lomo) wth the chimpanzee (Pan) shows, for instance, that there as teen lite evolutionary divergence in the hemoglobins and some other Ipolecules since they branched rom each ther eventhough the hominid Tine has since entered an emily new aapive zone (Figure 7-3). When fone uses taxonomi characters to daw inferences about lssifeation ‘ne mist always balance the potently coniting information derived ftom diferent character domains, One ust slo understand the subtle Aierence between evolutionary phenomena atthe molecular level and the ongnismic Ive (Mayr 964; Spun 1969). ‘Molecular taxonomy permis the analysis ofa whol ne set of char scters that are seemingly independent of more trains characters {Barrowclough 198), While molecular characters ae subject the same problems of convergence, paral, and reversal tht any character set Fay have its highly unlikely that morpolopicl and molecular harac- {er will be afected inthe sae manner. Where the molecular characters Srv supenr isin the extrrdnary deta hey can supply. The genome af $tnanmal onsss of abou ifn nucleotide base pai. In many eases there the morphologic evidence is ambiguous, molecular methods ave produced touly unequivocal conclusions. "The tui of evolving174 pants. nacroraKonoN [ADAPTIVE AND STRUCTURAL FUNCTIONAL ZONES F revo | sae br} Kur JX] | | | Bey oa we ee ae See, eer ey aca ad auctor 28% (anor Simpson 188. molecules superimposed on the background of casa txonomy is rae eet decropancies and inaccuracies which wil lead 1 i Provements i clsieaton (His 1987) arene Methods ae ot m panacea 1 alot every case (partic sane ital relavs) where taxa were classified with the a aiimerem molecular mets, erent lsiestions were rived aereic i ot spring view of he universality of mosaic evolution rtm uot ihe ranches in pylogeny shat ae aed of aie co nnead of concentrating on such trouble spots, some ra es ave wased ine and alae resouees By CON Wiking toy uncontroversial ndings of xo a ea aponantaspet of sme of he meclar methods that shor data dteomine relationship bythe presence or absence of shared they em prone to clit the stance Gere of dierence) aaa ua The spevia numerical methods ee fr sh al ae donrbed tn Chaper Il These methods afe 30 new and calor any competing ons (at sill uncertain hich he Best thee iam ano wht cheumsances dstance determination pe or toa character aml craig acd for stable mater for taxonomies has ‘nna sens to esablsh molectar collections, that i, eo CueTEAT. Ta‘oNOMC CHARTERS 175 lectins of material suitable for molecular analysis. Such material con- sists of whole specimens, tissues, or body Suds preserved in such a way that they are suitable forthe study of proteins, nucleic acids, or other ‘molecular constituents o¢ for chromosome analysis. For information on methods forthe preservation and use of such material, see Chapter 13, Benavior Behavior is undoubtedly one ofthe most important sources of txonomic characters, Indeed, Behavioral characters are often clear superior to ‘morpologicel characters inthe study of closely related species, partic lary sibling species (May 1963), However, there are two major technical Gravbacks, Behivior cannot be stdied in preserved material, and i is intermittent even in living animal. Some types of behavior cevur only Clring the breeding season ates, oly during part ofthe 24-hour pe- Fied. ‘The comparative study ofthe Behavior of related species has be: ome an autonomous discipline known as comparative thology. It his leady made major contriutions (othe improvement of classifications Of birds, bees, wasps orthoplerans, foes, fishes, and other groups. ‘The reason for the importance of behavior is obvious. Behavioral characteristics are the most important isolating mechanisms jn, most an iimals, and new adaptations are offen iiited by changes in behavior. ‘The rapidly expanding literature on behavioral systematics has in part ‘been summarized in nomber of reviews neuding those of Mayr 1958), ‘Alexander (1962), and Wickler (1967) Exemplary studies include those (Of Evans (1957, 1966) on the digger wasps (Sphecidae), Speth (1952) 9 the genus Drosophila, Tinbergen (1989) and Moynihan (1959 on gulls (aridae), Eickwort snd Sakagami (1979) on halictid bees, and Gord and Delish (1978) on chalcid wasps. Flash pattems are species-specific in most genera of fees (Figure 7-1 and have been used by systematists to unmask sibling species (Barber 1981; Loyd 1983). However, in the ge- fnus Photuris males may mimic the signal of sympatic species of other fenera (Photinus, Pyractomena, tc.) When one encounters several {fash patterns in the repertory of single mimicking species (Lloyd 1985), fone must not coed that one has uiscovered several sibling species “A great technical advance in the study of behavior hasbeen the devel. lopment of devices for accurate sound recording andthe translation of ‘Sounds into graphic pattems [the sonagraph (Figure 7-)}- More than 40 Species of North American evickets were either discovered or redicov- {red by B. B. Fulton and his followers as result of a careful analysis. of their songs The classification of species in several avian genera (or i> Stance, Myiarchus, Empidonas, Tyrannus) kas been greatly helped by an ‘analysis of sound recordings. A comparison of the calls of frogs andtens iat tats io Nt resin ‘sk sca ety ara par Ssh. (Prom abe 887) as Prt), Hog an rh the eds no only ha let the discovery of previously unrecognized sibling Species but also has shed light oo the relationship between previously es {ished species. Important studies of comparative sound analysis ia anurans. were done by Haro, Bhir, Bogert. Litjohn, Main, and ‘Mecham. The acoustics of animal behavior has been summarized in a hhumber of recent volumes (Tavolga and Lanyon 1960; Busnel 1963; Sebeok 1977: Claridge 1988). The importance of bid song in avian sy tematics fs been reviewed by Payne (986) Th ation to courship and acoustic behavior, various other kinds of behavioral elements have taxonomic value. For example, the pattern of the webbing constracted by various mites and caterpillars may be used at ‘aris levels inthe classification, The tWo bee genera Andhidium and Diamam were slow to be reeosnized on morphological grounds, yet Aiknown species of Anthidwn construct nests fom cottony plant fibers ‘while those of Dianhidinn constrvet nests from resinous plant exuda- tions and sand or small pebbles “The use of extrancous materials inthe construction of nest or of t= valor papel cases provides chargeters at various levels in the classiea- tion of caidisworme and bagworms. The egg eases of praying mantis have a speciew-specifi form. ‘OueTER?: Taxonomic oramrERs 177 “EGh Wy: fc A MM vl tn a & ° — 1 rat vratstons of elaod spose ot porows. (Afr Tele 1054) Behavior patterns that are characteristic of higher taxa are far more rae, Examples ae the use of mud in nest building by barn swallows and rag martin (fiw, certain “comfort movements” (seratching, Stretching, bathing, et.) in bitds (McKinney 1963), and grooming move ments in inseets.Atenton upto now has been directed so strongly atthe178 pass, wxcnoraxonour comparison of closely related species that the diagnostic value of behav: for patterns a the level of higher taxa has gone largely unexplored. Hos ‘ver, Eberhard (1982) showed that web structure is diagnostic for certain fonera or even higher taxa of spiders this permitted the construction of 2 cladogram of the orb-weaving behavior ofthe araneoid spies. Behav jor cladograms have also been published for other groups, for instance, the nest architecture ofhalictne bees (Fickwort and Sakagam 1979) Ecological Charactrs eis now well established that every species has is own niche in natre, tending to differ from is nearest relatives in food preference: habitat se lection: breeding season; tolerance to vars physical factors; altitudinal distribution; resistance to predators, competitors, and pathogens: and other ecological factors, When two closely related species coexist inthe Sime general area, they avoid fatal competition by means of these Species specific niche characterstis (principle of competitive exclusion See Mayr 1963, Chapter 4). A number of sibling species were discovered {as a result of disrepancies in food preference (host specticity —for ex ample, the apple und blucberry maggot—or habitat preference (Mayr 1963, Chapter 3). Many aspects ofthe ie eyele, such as life span, fecun- ity and Tength or time of breeding season, may be diferent in closely related species (Mayr 196). "Niche specificity Is quite pronounced even in species that are not ticolny substrate-specifie, such 3s bids, mammals, and mollusks. Kohn {1950} found that every species ofthe genus Comet inthe Hawaiian Is lands fers ecologically from related species. Two sibling species (ebracus and chaldaens) feed on nered polychaetes, but of 19 ebraeus, 136 contained nereid species o and none Contained species b, while of 106 Chaldacus, 5 conained species a und 98 contained species b. The larvae Of Drosoplila muller! and D.aldrch live simultaneously in the decaying Dullp of the fruits of the cactus Opuntia lindheimeri. However, the 10 Species der markedly in their preference for certain yeasts and bacteria (Wagner 194), che specialization is even more pronounced in animals that are substrate-secific, including host specie plant feeders among insects {and mites and hostspecifi parasites. Some of this information can be found in resent books on cocvoation (Nitecki 1982; Futuyma and Slatkin 1983). There is als the excelent treatment by Detier (1947), who em phaszed the sensory aspects of host selection. Many new species of in Sects were discovered by comparing populations ofthe “same species ‘curring on aifferent plant host, Sone enthusiasts, however, catied this principle too far and made any occurrence on a different host an ex- Coupren?: raxonome OMMCTERS 179 cause forthe naming of a new species. In the North American Cryphalini (ark beetles) alone, $3 ofthe species described by Hopkins turned out 10 bbe synonyms, Downey (1962) and Kohn and Onians (1962) have given {seful summaries of some of the relevant Iiterature. Host specificity of external parasites has been discussed by Clay (1949), Hopkins (949), and Holland (1968) Ecological differences also occur among populations of the same spe cies, Indeed, most widespread species exhibit differences inthe ecology ‘of local populations, particulary of peripherally isolated populations (Maye 1968311, 359, ‘Again, a5 vith behavioral character, diferences atthe species level, have been stuied far more extensively than have ecolopeal dferences between higher taxa, many of which are simply taken for granted. The fact that whales occupy a different adaptive zone than do basis too ob- vious to be mentioned, yet a close stay shows that even most genera, When well founded, cou’ definably diferent niches or adaptive zones. {Ezck 1947) showed, for instance, that each genus of Galipagos inches is characterized by its utilization of the envionment. Geospiza i aground finch (Chief food, seeds), Camarhynchus isa tree finch (chief food, in secis)s and Certhidea is warbler finch chet food, small insects) (Grant 1986). Parasites and Symbionts_In several instances sibling species were dis- covered because their parasites were diferent. A previously undis- ‘timinated species of Octopus Was discovered in California because it had its own set of mesozoan parasites. A new species of termites wis discovered because ifs nests contained a different set of termitophile Saphylinid beetles than di the ness ofa previously known species. Par- "sites ae also important in contributing to our knowledge ofthe relation- ‘Ship of higher taxa, Parasites evolve together wit their hosts and are in Some cases more conservative Thao the hosts. Unfortunately, they shift to new hosts more frequently than is sometimes admitted, and so evi- idence based on parasites must be evaluated very carefully (Baer 1957. For instance, umingos (Phoenicopter) exhibit characteristics which they share with both storks and geese. Teir bir ie (Mallophga are of the same genera as those which occur on gese. At fis this might SOg- ‘esta close relationship, but it actually indicates merely a comparatively Fecent transfer of the lice from geese to flamingos. Ifthe bid ice had been derived from 4 common ancestor, we would expect related but slightly diferent parasites inthe two otders of birds. Neither anatomic for molecular evidence indicates a close relationship of flamingos and feese, Human beings (Homo) and the African apes (Pan and Gorilla) Share more external and internal parasites with each other than the Aiscan apes do with the orang (Pongo). This suenthens the case fora close SiksEatnip between Homo and Pan (esaished on other round) fhe fal that inacalllar symbionts supply important taxonomic charcters was uncovered by Buchner (19a) and his school. For i SRae the most pimtve tes ofthe eoccis (Stengel ct.) have wenStastont: tut once a coved taxon has aquired one, this symbion (sath ah ts higly specie saps) wl be Found inthe derived sta pes of coved Repeat, nna ax a cocci cou be raaikelBecnse the) fad hetroseneos complements of symbionts Buchner 1960) The same tre of symbionts in other groups of i eae prtovcu faunas inte intestines o termite evolved toete “Sih ots and are pote weft fare of elton EAs of ambiguity in tere chsieation (Kirby 1950) eograhie Characters ach taxon, fom the species 10 the highest aon, has a seoeranhic Tat Since this range changes a8 te taxon evlves, a imate fe fan exis teetn steals and tngsoeap Mes eat Hreeeagaphore were systematsts.Palteras of sstrhution often po wee pecnle cues for clarifying a confused taxonomic picture and for TeiaPehxonomi nypaeses The dixtions of most taxa play Terie har permit conelsion about thelr past history Because EeO- sie rangeproviey sich wc axonomic character, evry Ta arto become lana with the eran the Held of ogee aa i lowing works shoud be mentioned: Digs (1974, Brown BE Gibvon 98, Cars C978, Datiagon (957, 1968), Mayr Uptoy Palo (979), Raven and Axclod (174, Simpson (98S), and svar 196) fess ‘Seorapeeharacters are important on 90 evel In icrotaxonoms snc apapanicallopai relationship of populations is often decisive Theses species stats Two sympatric ppulstions in repro eae eemation can never be conspecie. The mapping of populations, sees and allopatic species indispensable nthe deiitaion of lyse species and superspecies. Pome taxonomy a understanding of the eatonshi ier at is then hoiped by an amayis of thir istibutins. Two questions ae Siar portant) Whats the dneibtion a the nearest relative Ta sts eu ofthese axon? an (2) te taxon bas a une ‘ange, what ithe probable cause of he range diunstin “St Darin, we hn tt ther re pte to canes fF junction: ‘Gurren: Tatcnowe oumacrens 181 1 Primary isolation (establishment of a founder population through dispersal) 2 Secondary isolation, thats, the fracture ofa previously continuous range by a new distibuional barrier across the previously continuous range (siearance) A knowledge of the Farth’s history and the geological record often, permits one to decide whether the isolation is primary or secondary and, ‘more important, for how long time the two isolates have been sepa rated. The fet thatthe mockingbirdson the voleaie Galapagos Islands had flown across from the adjaeent South American continent was pe fectly clear to Darwin. The fat thatthe fauna and fora of the volcan Hawaiian Islands had gotten thereby tansoceanic colonization has also been long evident; the same conclusion was inevitable forall vole ‘oceanic islands in any ocean, Some continental island, however, fr in Stance, Madagascar and New Zealand, have a mixture of ol vicariant clements that are remnants of the biota of the continents oF plates to ‘hich thse islands had formerly’ been atached and of more recent ele ‘ments that reached them by tansoceanic colonization, Depending onthe dispersal focly ofa taxon, either primary or sec ‘ondary isolation i the more important cause of range dijonctions. Some roups, sch as earthovorms (except when passively transported) and pri mary freshwater shes, have very tow dispersal facies, In such groups almost any range disjunction is secondary, caused either by geological events (plate movements, mountain building) or by climatic vegetational Shifts such as range disjunction caused by the Pleistocene ice caps or by Post-Peistocene habitat shifts ‘Groups with poor dspersl facies are particularly useful in deter- mining the age of taxa, If such a group is found in South America and ‘Australia But not on the northern continents, itis probubly an old ‘Gondwana element that goes back tothe time when Australia wus stil ‘connected with South America across Antarctica, If poorly dispersing fsroups are found in Africa and South America, they mresumably go back to the time before Afica separated from South America, Most faunss, however, area mixture of ol and more recent clement (Mase 1976552 564) and these elements have to be carefull discriminated to establish a reliable chronology. The New Zealand kiwi and mots, for instance, are among the remnants ofthe eld Gondwana fauna tht ruled when New Zealand was stil in contact with Antarctica, but most oral ofthe other birds of that island are descendants of more recent transoceenic imm ‘rants from Australis or Melanesia, “The facility with which certain groups can cross water gaps is dact- mented by the numerous colonizations ofthe Havaian Islands or of182 pans: uacnoraxoncur ven more remote Pie ishnds sich as the Marquess and the Kosa Ta he case ofesy diapers often io termine the source aa af colonization and this the nearest restive. Some ‘nooeopaphers tend founderestnats the diaper ay of mos orn ismehe ar um iguana others restricted the Ames, bar suring tet speci in Tengu and superspecies in FA tart stds foreland connesttons have come othe uexpected Snclusion these Herds could Rave reached these outposts oy BY longdivtence dispersal "fae question jae asked, How cana pater of dstrbuton Rep in determining the nearest ctv of higher axon? Here Basi le is SS Totowss in ease of daub, te neret relive fs the ame 0 anc reson or continent Tis wel llsrtd by the taxonomy af fumes: Astraan sods, Wen warbler, eather, shies, fhe. re crepe sn woo) swallows were found in Asta {Tus ony ala they wer led wh the eomesponding Eo Sine holctc eon, There were, however, ough puzing aspects of the scology and sion of these Avsalan ta to rae Job Coal they nua be convergent adpsve redaton finden AUS tralian stocks? That hin ses to Be the cne ws eventually shown BY Sey and Anis (58) wth he elp of DNA hyrization ne ncarestrenives however at vase found i the sae SEO raph Ten The nearest rte the lamasin not South America [hil ou tne ool cart The story ofthe marsepas s more come plot Toy te) at ested to South Ameria and Atala (xc Fert opossum ta uccsslyreivaded North Amerie) Foss mar ‘Spins however, hae been fund in North Ancien, Europe, Ari uaa" Gog ae now able to date he age of foi a of ai omer land connections ates the Fennyivanan rather acca). AS tars the marsupials are concerned hn est the wel-substaniated (Shelton that Norch America mas he rial sees tea, From hee Sheraton invaded rope and fom ere Alc and AS) ety the Tetary across north Atacand rie his ation soon Be ine ovine far mre sucesfl ration solonzed South America Shi rem tere expanded across Antarctic fo he Astalian eion, We ive foate ia tis now povable to poi easonaly accurate ates forthe tot ofthe vas pts and pate remnants "Apter sometimes srives In peripheral lated aces, suc te tars on New Zs an he nh an fem on Mulagnsca, 1 phenomenon wed by Mathew (915 and ober r03208- Mc sor oad pentazatom: Os he wl, however, the tne pre branes of tao ation are mot ikl tobe fad near the orginal source ara, with the expanding colonizations evel farther and farther away from the ancestral location, The theory of plate tectonics has shed considerable light on the evo: lutionary history of many animal taxa, particulary tat of poor dispers- ers. It has helped resolve a number of taxonomic puzzles, I 1922 Michaelsen made the unexpected discovery that the earthworm fauna of India more closely related to that of Africa and South America than it Isto that of the rest of Asia. He explained this onthe buss of Wegener's tory of continental drift, aeording 0 which the Indian plate was for. mer par of Gondwanalund and established contact with the Asian pate ‘only in the middle of the Tertiary. Many other biogeographic pases Ihave since then been explained inthe ight of plate tectonics. However, even today there are biogeographic puzzes which have not been salved by the better understanding of geographic history brought about by plate ‘Many genera of insects that are now found only inthe southem con tinents (southern Africa, South America, Australis) were once believed to be pat of the Gondwana fauna. Surprisingly, some ofthese genera have since been found as fossils in Oligocene Baltic amber, tts now clear thatthe southern distributions of these genera ate reits of for- tmerly much wider distributions (Ander 194), Such eases warm ks not (0 become too assertive in tying to reconstruct the connections tween, present distribution, the Earth's history. the fossil eeord, and taxonomic Felationshin. ‘THE WEIGHTING OF CHARACTERS ‘The more characters a taxonomist uses the more likely i is tht the in {formation provided by diferent characters wil be disconlant, According toone set of characters, species A and Bare more closely related to ea ‘other than they are to C, but according to another set, species A and C fare more closely related than either is to species By (Figure 11-2) Homoplasy and mosaic evolution ae responsible for such seeming con. traditions. There once Were great hopes of reslvig these conflicts with the help of numerical methods (Chapter II). but these hopes were only partially fulfilled. Unfortunately, the use of different phenetic methods (Presch 1979) often results in diferent phenograms, and the same prob- Jem is encountered withthe use of dfereat cladistic methods, Aware of these difficuies, numerical taxonomists are increasingly including ‘weighting in their methods (Funk and Wheeler 198), Evidently in these ‘cases different characters provide different information, ‘What the taxonomist mast ask is which ofthe discordant characters