THE AUK A QUARTERLY JOURNAL OF ORNITHOLOGY Vo•,. 76 JULY, 1959 No. 3 THE STATUS OF AVIAN SYSTEMATICS AND ITS UNSOLVED PROBLEMS* BY ERWIN STRESEMANN BIOLOGISTSall over the world have devoted the year 1958 to the mernoriesof Darwin and Wallace. We ornithologistsalso should call to our attentionthe powerful impetuswhich we have receivedfrom the intellectualwork of thesegreat men, to whom we owe a better understandingof the origin of the great diversityamong birds. Two factorsare responsible,according to Darwin's theory, for the ever greater perfection of the living world: variation and selection. I suggestthat thesesame factors are of major importancein the interplay of ideas and concepts. Permanent advancesin our scientificunder- standingwe owe likewiseto variation and selection. Variation is the consequenceof the individuality of thosewho work on the sanhetopic. Zoologistsdiffer in their philosophicalbackground, in the extent of their knowledge,in their thoroughness,and in their gift of combination. This variability is certainlydisplayed to a high degreeby the avian systematists.What Max Fiirbringer has written about them, will forever remain true: "At various times a few fortunate individualshave existedwho were gifted with such an acute insight that it revealedto them intuitively, one might almostsay instinctively, this or that systematicrelation among related forms without the necessityof laboriousinvestigation." At the other extremethere have been sonhepoor devils who did wrong whatever they did and who were completelylost without methodology. Survivalof the fittest will decidewhich of the many competing theorieswill prevail. Only one can finally survive. Eachrevisor at- temptsto shortenthe struggleby actingas a selectivefactor. When * This paper was delivered in essenceat the Seventy-fifthAnniversary Meeting of the American Ornithologists'Union on October 1õ, 1958, in New York, during the Symposium on Contemporary Problems in Ornithology. Participation of Dr. Stresemannwas made possibleby a grant from the National ScienceFoundation, which the American Ornithologists'Union gratefully acknowledges. [" Auk 270 STRESEMANN,Status o[ .4vian Systematics I.Vol. 76 he hasto synthesizea modern system of birdshe is forcedto choosea singleone among many conflicting theories, often without having the opportunityof examiningthoroughly the argumentsof the different authors. I myselfhave painfully experienced the feelingsof a tax- onomiccompiler, for I developed(1934) an eclecticsystem of bird classificationsome 25 yearsago. While doingso, I made a few mis- takes,as I now realize. Has not the experienceof othersbeen the same?On the wholeall the arian systemspresented in the standard worksin thiscentury are similar to eachother, since they are all based on Fiirbringerand Gadow. My systemof 1934 doesnot differ in essencefrom thosewhich Wetmore (1951) and Mayr and Amadon (1951)have recommended. The most obvious differences between the several recent classifi- cationsare in the delimitation of the higher categories. Wetmore arrangesthem in 27 orders,Mayr and Amadonin 28 orders,while I recognized49 orders(1933-34: 738-853). In otherwords: I preferred to leavethe questionof phylogeneticrelationship open in many more casesthan the authorsof the other two systems.-Today I would recooo-nize51 orders (seeAppendix). An answerhas tentativelybeen given now to someof the questions of relationshipwhich I had consideredas unsolved;others which seemedto be solvedat that time have meanwhilebeen reopenedfor discussion. The outsider who reads some of the recent critical dis- cussionsmay easilyget the impressionthat our conventionalsystem is full of errors. It seemsto me, however,that one must apply to many of the newer proposalsthe sameevaluation which wasmade by Alfred Newton in 1893,who wrote: "Someof the later attemptsto systematicarrangement are in my opinion amongthe most fallacious, and a gooddeal worsethan thosethey are intentedto supersede." A considerablepart of theseobjections are due to effortsto create an avian systemwhich specifieswith great precisionthe degreeof phylogeneticrelationship of all groups. The constructionof phylo- genetictrees has openedthe door to a wave of uninhibitedspecu- lation. Everybodymay form his own opinion on the phylogenyof the higher categoriesof birds, because,as far as birds are concerned, thereis virtually no paleontologicaldocumentation which has revealed suchimportant information on the phylogenyas has been the casewith the otherclasses of vertebrates.The investigatorof arian phylogeny mustrely on indirectclues, which are nearlyalways ambiguous. It is for this reasonthat Seebohm(1890) recommendedto the systematists to ignore phylogeneticendeavors by the following sentence:"The classificationof existingbirds is the studyof a horizontalsection of July•xo•oa Sx•aA•, Statusof AvianSystematics 271 the greatbird massof the world,and oughtto form a differentand distinctsystem confined to the horizonof the presenttime." To be honest, we must admit that we mix the horizontal and the vertical systemto this very day. It is a concessionto the "horizontalsection" if, for instance,in a much usedcontemporary system the "true birds" are divided into two superorders-theImpennes or Penguinsand the Neognathaeor Typical birds--becausethe author has surelynot doubtedthe closerelationship between penguins and Procellariiformes and has surelyknown Simpson's(1946) important conclusions.Yet, in the samesystem, an attempt is made to group the "typical birds" accordingto the vertical principle; that is, not accordingto their apparent similarity or differencebut accordingto their presumable phylogeneticrelationship. For this reasonthe author follows Mc- Dowell (1948)by droppingthe former distinctionbetween Palaeogna- thae and Neognathae. Max Fiirbringer, with his incomparablepractical experience,has frankly admittedthat the decisionin questionsof relationshipis very often basedon rather subjectiveconsiderations. Some of the more recent revisorsdo not display a similarly wise modesty. Verheyen (1958b)makes the proud assertion:"in contradistinctionto the con- ventional classifications we will introduce rational classification," which is basedon the totality of as many individual charactersas possible. This total sum of charactersVerheyen calls the "morphological potential." After equatingthe morphologicalpotential for a given group of birds with I00, he comparesit with the morphological potentialof other groupsof birds, then calculatesthe percentageof agreement,and drawsfrom this his phylogeneticconclusions. In this methodarithmetic replaces the role of intuition and of a judgment trained in functionalinterpretation. ElsewhereVerheyen (1955) states:"A clear,precise and phylogeneticallysound classification must be basedon charactersthat are practicallyinvariable and which are essentiallyimmune to the adaptationand modificationimposed by thehabitat." To thisI answer:Taxonomy would be indeeda simple matter if such "practicallyinvariable characters"existed. However, the relentlessmodifying power of evolutiondoes not sparea single structuralelement; any changeof the well integratedmorphological total resultseither at onceor graduallyin correlatedchanges from which neither the skeleton,nor the muscles,nor the external features, northe behavior remain excepted. Let megive one example: the loss of the powerof flightcauses not only extensivechanges of muscles andbones of the anteriorlimb andof the shouldergirdle, but it also has effectson the pelvis and the posteriorlimb, and even on the [- Auk 272 SXRESEMANN,Status of Avian Systematics [Vol.76 feathersof the entire body, which becomegreatly simplifiedas shown by the struthionsbirds, by Apteryx,by the Dodo,and by someof the flightlessrails. Verheyenis not an isolatedexample. Other recent studentshave believedthey have found invariablecharacters that would showthem the way throughthe labyrinth of avian diversity. Lowe relied on the structureof the skull in his studiesof the Charadriiformes,but got badlyoff the track (Bock,1958). Beecher(1953) based a new classifica- tion of the songbirds largely on the jaw muscles. His methodsand paradoxicalconclusions have likewise receivedwell-merited criticism (Mayr, 1955). The easiestsolution was tried by certainauthors who comparedthe acceptedsystem of birds with the systemof their parasites,and who in all casesof conflictconsidered the parasitologicalevidence as more decisivethan the findingsof comparativeanatomy. This was done by Timmermann(1957). On the basisof their Mallophaga,he con- cludesthat Rostratuladoes not belongto the Limicolae (Charadrii) but with the rails,Phaethon not to the Steganopodes(Pelecaniformes) but with the Laro-Limicolae(Charadriiformes), and so forth. Such an exaggeratedevaluation of the parasitologicalevidence has had the effectthat ornithologistswill utilize parasitologicalinformation only with thegreatest caution. In thisconclusion I am in entireagreement with Ernst Mayr (1957). For instance,the fact that the flamingos (Phoenicopterus)are parasitizedby two generaof Mallophagawhich otherwiseoccur only on Anatidaecan by no meansbe consideredas proof for an origin of the flamingosfrom the Anatidae rather than from the Ciconiiformes.It appearsby far moreprobable, that the Mallophagahave been transferred rather recently from the waterfowl to the flamingos.This is not onlythe viewof the ornithologistMayr (1957),but alsoof the mallophaganspecialist, Dr. von Killer (1957). May I refer to still anothercase. Sometaxonomists have recently