sign in sign up
<<
Home , Anthochaera, Aplonis, Black solitaire, Cedar waxwing, Corvinae, Corvoidea

THE RELATIONSHIPS OF THE (STURNIDAE: STURNINI) AND THE (STURNIDAE: MIMINI)

CHARLESG. SIBLEYAND JON E. AHLQUIST Departmentof Biologyand PeabodyMuseum of Natural History,Yale University, New Haven, Connecticut 06511 USA

ABSTRACT.--OldWorld starlingshave been thought to be related to crowsand their allies, to weaverbirds, or to troupials. New World mockingbirdsand thrashershave usually been placed near the thrushesand/or . DNA-DNA hybridization data indi- cated that starlingsand mockingbirdsare more closelyrelated to each other than either is to any other living . Some avian systematistsdoubted this conclusion.Therefore, a more extensiveDNA hybridizationstudy was conducted,and a successfulsearch was made for other evidence of the relationshipbetween starlingsand mockingbirds.The resultssup- port our original conclusionthat the two groupsdiverged from a commonancestor in the late Oligoceneor early , about 23-28 million yearsago, and that their relationship may be expressedin our passerineclassification, based on DNA comparisons,by placing them as sistertribes in the Sturnidae,Superfamily Turdoidea, Parvorder Muscicapae, Suborder Passeres.Their next nearest relatives are the members of the Turdidae, including the typical thrushes,erithacine chats,and muscicapineflycatchers. Received 15 March 1983, acceptedI November1983.

STARLINGS are confined to the Old World, dine thrushesinclude Turdus,, Hylocich- mockingbirdsand thrashersto the New World. la, Zootheraand .d) Cinclusis closerto Starlingsusually have been thought to be re- the thrushes,flycatchers, starlings and mocking- lated to crows and other members of the cot- than to the wrens." void assemblage,or to New World troupials A Distance Wagner tree (Farris 1972) based (Icterini) and/or Old World weaverbirds (Plo- on the DNA hybridization data showed ceinae).Mockingbirds and thrashershave been and Sturnusto be more closely related to each viewed as relatives of thrushes (Turdidae) and/ other than to any other taxon, and they clus- or wrens (Troglodytidae). tered with the thrushes,muscicapine flycatch- Evidenceof a relationship between starlings ers, erithacine chats, and . and mockingbirds had been reported at least The same pattern emerged from a more ex- twice (Beecher 1953, Stallcup 1961) before we tensivestudy of the passetines(Sibley and Ahl- published the results of a comparisonof their quist in pressa) in which the mimine-sturnine DNAs (Sibley and Ahlquist 1980). The 1980 relationship was based on the same data as in study was based on 153 DNA-DNA hybrids the 1980 paper, but many additional data per- among 18 oscinegenera used as radio-labeled taining to other passefinegroups were includ- single-copy"tracers." Among our conclusions ed. In the classificationderived from this study, was the following statement.(DNAs of the cit- we treated the tribes Sturnini and Mimini as ed genera were used in the 1980 study). sister groups in the family Sturnidae. The fol- "The thrushes,muscicapine flycatchers, mock- lowing is a partial outline of this classification ingbirds, starlings,and dippers are membersof to indicate the relative positions and the names a monophyleticassemblage. a) Within this group we have assignedto the taxa mentioned in the the muscicapineflycatchers (, Melae- present paper. Details for most groups have nornis,, Rhinomyias)are closely related to been omitted. the -like thrushes (Erithacus,Erythropygia, Passeriformes ,, , Pogonocichla, Myr- mecocichla,Copsychus). b) The starlings (, Suborder Oligomyodi, the suboscines ,, Ampeliceps, ) are Suborder Passeres, the oscines closestto the mockingbirdsand (Mi- Parvorder Corvi mus,Dumetella, , Oreoscoptes. c) The tur- SuperfamilyCorvoidea

23O The Auk 101:230-243. April 1984 April 1984] -MockingbirdRelationships 231

Family tallica builds a pendant, globular nest with a , crows, jays, side entrance,and Acridotheresginginianus digs , birds-of-paradise, a nest hole in an earthen bank. The color of Australian magpies, etc. starling eggsvaries from white to green or blue- Parvorder Muscicapae green; some are unmarked, and others have Superfamily Turdoidea reddish or brownish markings. Family Bombycillidae, Davisand Miller (1960)recognized 31 Family Cinclidae, dippers in 13 genera in the "family Mimidae," includ- Family Turdidae ing Donacobiusatricapillus, which Miller (1964) Subfamily Turdinae, typical noted as having the voice and habits of a thrushes (Troglodytidae).Clench et al. (1982) concluded Subfamily Muscicapinae that Donacobiusactually is a wren, which is con- Muscicapini, muscicapine sistent with several DNA comparisons.In Ta- flycatchers bles ! and 3, Donacobiusclusters with the Syl- Tribe Erithacini, chats vioidea, including the wrens, not with the Family Sturnidae mimines in the Turdoidea. Tribe Sturnini, starlings Most mockingbirdsand thrashersoccur from Tribe Mimini, mockingbirds southern Canada and the to the and thrashers West Indies and Central America. With the ex- Superfamily , Old World clusion of Donacobius,only Mirnus occurs in , wrens, titmice, nut- South America. hatches, , etc. All mimines build open, cup-shaped nests, Superfamily Fringilloidea, , sun- usually placed in bushesor trees. The Pearly- birds, weavers, fringillines, trou- eyed (Margaropsfuscatus) places its pials, etc. "bulky, cup-shaped nest . .. in a bush or tree, cavity of a tree, or on the side of cave or cliff" Although this classificationcontains depar- and the (Cinclocerthiaruficauda) builds tures from previous arrangements,most of the its nest "in a cavity of a tree or tree fern, or at changeswere received with approval or mild the baseof a palm frond" (Bond 1971:169). doubts,but the inclusion of the starlingsand colorsare blue or green with various amounts mockingbirdsin the same family elicited vig- of spotting or streaking. Thus, the and orousdissent from severalcolleagues. We have nests,although variable, are somewhatsimilar therefore made a new study, using different and do not opposea relationship between the taxafor the "tracers,"several additional species, two groups. and many morecomparisons. Our original con- clusions have been confirmed by these new data, and we have alsofound severalcongruent TAXONOMIC HISTORY morphological characters.We have been un- Opinionsabout the relationshipsof starlings able to find hard evidencethat precludesa close have ranged widely, but most of the classifi- relationship between starlings and mocking- birds. cationsof the past130 yr have placedstarlings closeto crows and their allies, including birds- In this paper we review the taxonomic his- of-paradise,Old World orioles,and .A tories of the two groups, note evidence from relationship to the corvoid groups was advo- other sources,and present the results of the new DNA hybridization study. catedby Sharpe(1890, 1891),Reichenow (1914), Stresemann (1934), Stonor (1938), Delacour and Vaurie (1957), and Bock (1963). A relationship DIVERSITY AND DISTRIBUTION to the corvoids and the New World troupials (Icterini) was indicated in the classificationsof Amadon (1962) recognized 111 speciesof Bonaparte(1850), Gray (1870),Sundevall (1872), starlings in 26 genera, approximately evenly Sclater (1880), the early A.O.U. Check-lists divided between and . Three species (1886, 1895, 1910), Shufeldt (1889), and Coues of Sturnusoccur in ,and Aplonisreaches (1896). Sharpe (1890: 1) declared that the star- and many southwest Pacific islands. lings are "undoubtedly allied" to the Corvidae Most starlingsnest in cavities,but Aplonisme- and repeated this emphatic opinion in his ex- 232 SlBLEYAND AHLQUIST [Auk, Vol. 101 tensive review of avian classification in 1891 starlings and corvids seemsto have been based (p. 85). The corvoidsand ploceineswere viewed on the similar shape of the bill in starlings, as starling relatives by Reichenow (1882), von some corvines, and Old World orioles and the Boetticher(1931), Amadon (1943, 1956), Mayr fact that some starlings and some corvines are and Amadon (1951),Mayr and Greenway(1956, black. It appears that the most influential 1962), Storer (1971), and Voous (1977). "character" has been tradition, with most au- Wallace (1874) associated starlings with thors following the opinions of previous au- weaverbirds, , and larks as thors. "Sturnoid Passeres."Parker (1875) reported The taxonomic history of the - similarities between the skulls of the Celebean thrashergroup hasbeen relatively simple. Most (Enodes)and that of a Song (Tur- authors have placed them with or near the dusphilomelos) but decided that the myna was thrushes; some have allied them to the wrens. more -like and possibly related to the They were considered to be thrushes by Bo- birds-of-paradiseand the drongos(Dicrurus). naparte(1850), Gray (1869),Coues (1896), Ridg- Stejneger (1885) placed starlings with cor- way (1907), Beecher (1953, 1978), Morioka vids and meliphagids, which may have influ- (1967), and Gulledge (1975). Of these authors, enced Wetmore (1930, 1951, 1960), who listed only Beecher (1953) compared mockingbirds starlings near meliphagids, although far from and thrusheswith starlings, and he found sev- Corvidae. Wetmore (1960) also placed New eral similarities, as noted below. Zealand wattlebirds (), shrike- Baird (1858), following Cabanis (1850), re- (), and (Laniidae) near star- moved the Miminae from the Turdidae and lings. His influence on the A.O.U. Check-lists placed them with the babblers, wrens, and of 1931, 1957,and 1983 is apparent.In the 1983 (Chatnaea) in the "Liotrichidae" (= list, the Sturnidae are between the Laniidae and Timaliidae). He noted that "the Miminae ... Meliphagidae, far from the Corvidae. have a Thrush-like appearance, which has Amadon (1943, 1956) reviewed the genera of caused them to be placed by most authors starlings and, as possible relatives (1956: 9), among the Turdidae."He concluded, however, suggestedthe Oriolidae, Vangidae, and Dicru- that "it is very difficult to draw the line be- ridae, with the Prionopidae, Cracticidae, Par- tween this sub-family and the wrens; the chief adisaeidae, and Corvidae as less likely rela- difference lies in the larger size and bristled tives. He also noted that the might gape." In 1864 Baird reversed himself and rein- be related to the starlings, as von Boetticher stated the mimines as a subfamily of the Tur- (1931) had suggested.Mayr and Amadon (1951) didae. assignedthe Sturnidae to a group of "Weav- Stejneger(1883) was the principal proponent erbirds, Starlings and Associated Families," of the view that mockingbirds are closely re- which included the Ploceidae, Oriolidae, and lated to wrens. This opinion was adopted by Dicruridae, followed by the Corvidae and their the A.O.U. Check-list Committees of 1886 and allies. This arrangement was alsoused by Mayr 1895, which treated the mockingbirds as a and Greenway (1956, 1962). subfamilyof the Troglodytidae,and the Check- Berndt and Meise (1962) placed starlingsbe- list Committee of 1910, which listed the wrens tween the Ploceidae and Fringillidae, and Bock and mockingbirds as adjacent families. Lucas (1963) suggested that birds-of-paradise and (1888) forged a compromisewhen he suggested may have been derived from star- that "the Miminae hold a somewhat interme- lings. Voous (1977: 382) placed the Sturnidae diate position between the Wrens and the between the Corvidae and Passeridae and not- Thrushes"and thus provided the basisfor one ed that "the Ploceidae may probably be their of the most frequent arrangementsin lists, viz., closest relatives." Troglodytidae, Mimidae, Turdidae. This pat- Thus, there has been no consensus about the tern was followed by Wetmore (1930, 1951, relationships of starlings, right up to the pres- 1960), Stresemann(1934), Mayr and Amadon ent. Among the groups that have been sug- (1951), Mayr and Greenway (1956, 1960), Ama- gestedas their closestrelatives are membersof don (1957), Delacour and Vaurie (1957), Berndt both parvordersand three of the six superfam- and Meise (1962), Storer (1971), and Voous ilies that we recognizein our DNA-based clas- (1977). The A.O.U. Check-list Committee of sification. The alleged relationship between 1983 placed the "Mimidae" between the poly- April 1984] Starling-MockingbirdRelationships 233 phyletic family "Muscicapidae" (which in- ).Lanyon noted that "The only character cludes the thrushes) and the "Prunellidae." complex that varies in any significant way Thus,as for starlings,there hasbeen no con- amongthese four samplesis the size of the pes- sensus about the relationships of mocking- sulus... and the positional relationship of the birds, although an alliance to the thrusheshas ventral ends of the A3 elements to the pessu- been a reasonably consistent suggestion. lus. In Dumetella, Catharus,and Sturnus,the ven- tral ends of the left and right A3's do not meet MORPHOLOGICAL EVIDENCE but rather are separatedby the comparatively wide pessulus,which is continuous with the Beecher (1953) compared the jaw muscula- fused A4's. But in Dendrocittathe pessulusis so ture, the ectethmoidplate, and other characters narrow that the ventral ends of the A3's nearly of the head region in a quest for evidence of touch each other medially. I can separate all phylogeneticrelationships among the oscines. eight of the corvid specimenson the basis of In starlings,he found the jaw-musclepattern this character complex. I would not be able, to be "similar to that of the Turdinae in com- with certainty,to distinguishbetween turdids, plex M3b..." and the palate to resemble that , and sturnids, however." Lanyon noted of the Turdinae. In mockingbirds,the jaw-mus- that Warner (1972: 385) also observed these dif- cle pattern and the bill, palate, and tongue are ferencesbetween corvoids and the starlingsand similar to those of the thrushes. Starlings, thrushes. mockingbirds, and thrushes have a double ect- Although these morphologicalcharacters do ethmoid foramen, thus differing from the cor- not prove that mockingbirds and starlings are void groups, in which it is single. Beecher(p. each other's closest living relatives, they do 282) noted that "the double ectethmoid fora- suggestthat it is unlikely that starlingsare more men and the muscle differences suggestthat closelyrelated to corvoidsthan to mimines and the supposedaffinity of the Miminae to the turdines. babblers and wrens is the result of conver- gence. The Miminae may stem from the Tur- SEROLOGICAL EVIDENCE dinae .... "Thus, Beecher was the first to sug- gesta fairly closerelationship among thrushes, Stallcup(1961) used the precipitin technique mockingbirds,and starlings. to comparethe saline-solubletissue proteins of In suboscinesand mostof the corvoidgroups 17 genera representing 15 of the families of (crows, birds-of-paradise, Old World orioles, oscinesrecognized in the 1957 A.O.U. Check- drongos, woodswallows, etc.), there is one list. Among the taxain this study were the Cor- pneumatic fossa in the head of the humerus. vidae, representedby the Blue ( In most of the other oscines(= our parvorder cristata);Troglodytidae, (Thry- Muscicapae),there are two fossae,as described othorusludovicianus); Mimidae, Northern Mock- and reviewed by Bock (1962). Starlings have ingbird (Mimuspolyglottos) and Brown Thrash- two complete fossaeor, in Eulabes(= ), er (Toxostomarufum); Turdidae, the beginning of a second fossa. Sturnushas (Turdusmigratorius); and Sturnidae, European two fossae,as in mockingbirds and thrushes. Starling (Sturnusvulgaris). This characteris not consistentin its develop- Antisera were prepared against tissue ex- ment among non-corvoid oscines,but the dou- tractsfrom these speciesand comparisonswere ble condition in starlingsat least suggeststhat made between them and representativesof 11 they are not closely related to corvoids. or 12 other oscine families. The results were W. E. Lanyon (pers. comm.) found syringeal presentedin two tables and a series of dia- characters that unite mimines, sturnines, and grams,and Stallcup offered the following in- turdines and distinguishthem from corvines. terpretations of the precipitin measurements:

Lanyon examinedthe stained cartilaginoussy- 1. "Members of several families are more like ringeal elements in these groups, as follows: Cyanoclttathan is Sturnus.... "(p. 51). Mimines (Dumetella, Mimus, Cinclocerthia,Mar- 2. "Thryothorusdoes not resemble closely the garops);Turdines (Catharus,Turdus, Copsychus); members tested of the families Mimidae, Tur- Sturnines ( Sturnus,Aplonis, Scissirostrum, Cinny- didae, and Regulidae." (p. 52). ricinclus);and Corvines (,, 3. "On the basisof the serologicaldata, there is ,Cyanocitta, , , Calocitta, some justification for assuming relationship 234 S•BLE¾AND AHLQUIST [Auk, Vol. 101

between the mimids and thrushes,although The thermal dissociationcurve of a DNA-DNA hy- the serological resemblanceof ... these two brid is a plot of the different ratesof nucleotidesub- families is not as great as that between Tox- stitution that have occurred in different sequences ostomaand Sturnus."(p. 52). (Grula et al. 1982). Thosehybrid duplexescomposed 4. "The serologicaldata present no clear idea as of homologous sequencesthat have diverged most to the speciesmost like ... Turdus.... The rapidly, and hence are most different, will contain speciesthat show greatest serological corre- the largestnumber of mismatchedbase pairs and will spondence to Turdus are Sturnus, Dendroica, therefore melt at the lower temperatures.The DNA , and Mimus. It would seem, therefore... hybrids that do not melt until exposedto the highest that the turdids and mimids are related. There temperaturesare composedof sequencesthat have are other species... that show greater sero- evolved slowly and contain few, or no, mismatched logical correspondenceto Turdus than does bases.Figure 1 shows the relationship between tem- Mimus. Sturnusis a notable example." (pp. 52- perature and percentageof counts eluted, which 53). measuresthe percentageof basepairs that have melt- [Note: Stallcup was considering only the ed. question of the relationship between mi- To obtain a measureof genealogicaldivergence be- mines and turdines in these comments. Thus, tween taxa we calculate the T5oH statistic, which is he did not attempt to explain the reactions the temperature in degrees Celsius at which 50% of between the antiserum against Turduswith the DNA duplexes in a given hybrid have melted Dendroica and Parus. In other tests Dendroica into single strands.Thus, the delta TsoHmeasures the was most like Spiza and ,and Parus averageamount of divergence,which is a product of was most like Cyanocittaand Turdus.The de- the averagerate of divergence.The delta T•oHis the grees of relationship indicated in these com- average (or median) difference between the two ge- parisons with Parus were lower (82%, 79%) nomes composing a DNA-DNA hybrid. In the cal- than those between Turdus and Sturnus(92%, culation of delta T•r/, it is assumed that all of the 86%).] nucleotide sequencesin the genomes of the two 5. "Serologically,Sturnus seems to be most like speciesbeing compared have homologs in the other Toxostoma."(p. 54). The Sturnus-Toxostomare- species,that all sequencespotentially can hybridize actionswere 92% and 91% and, thus, among with their homologs,and that all degreesof diver- the higher values in the experiments. gence can be detected. The percentageof hybridiza- tion declines, however, as the amount of divergence There are some internal discrepancies in increases,and the thermal-stability curve is progres- Stallcup'sdata that are characteristicof serolog- sively truncated by the effects of the experimental ical comparisonsin which several rabbits are conditions. Figure 1 illustrates this by the intercepts used to produce antisera. As Stallcup noted between the 60øCtemperature and the percentageof (1961: 49), this explains the unequal reciprocals counts eluted values. For the curves in Fig. 1 this is in some of the tests.Stallcup's data clearly re- not a problem, becauseall of the curvescross the 50% hybridization level, and the delta TsoHvalues can be veal a closerelationship between starlings and calculated easily. For more divergent taxa, whose mimines, however, with the thrushes nearby melting curves are entirely above the 50% level, it is and the corvids and wrens at a considerable necessaryto extrapolatethe most nearly linear por- distance away. tion of the curve to its intercept with the 50% level. This is done by fitting a cumulative distribution METHODS function to the data to find the intercept. Thus, the T5oHvalues incorporate the percentageof hybridiza- Our proceduresare based primarily on those of tion and the measureof thermal stability in a single Britten and Kohne (1968), Kohne (1970), and Britten number. et al. (1974). A brief description of our methods was To convert the delta TsoHvalues into a phylogeny published in TheAuk (Sibley and Ahlquist 1982) and requires a procedure to obtain a hierarchical cluster- more completeversions in Sibley and Ahlquist (1981, ing of taxa.We usethe "averagelinkage," unweight- 1983). ed-pair-group method, which begins by clustering The two strands of a DNA-DNA duplex molecule the dosest pair or pairs of taxa. The next step links are held together by hydrogen bonds between the the taxa that have the smallest average distance to complementarybase pairs in homologousnucleotide any existing cluster. This procedurecontinues until sequences.The DNA hybridization technique deter- all taxa are linked. The underlying idea is that the mines the similarity between the two basesequences only reasonfor "closeness"is true homologyof char- by measuringthe temperature required to melt the acters,while "distance"may resuIt from the failure hydrogen bonds and thereby to convert double- to identify existinghomologies. The DNA hybridiza- stranded DNA into single-stranded DNA. tion data are especially compatible with this view, April 1984] Starling-MockingbirdRelationships 235

compensatefor all sourcesof experimental error by using five or more speciesand/or replicatesfor each pairwise comparisonbetween taxa so that an average 90 delta T5oH,its standard error (SE), and standard de- viation (SD) can be calculated for each branch node

80 in a phylogeny. Becauseof the incremental summa- tion of values in the averagelinkage procedure,the older nodes in the phylogenies are often the aver- agesof 10, 20, or more delta T•oHmeasurements.

T•E UNIFORM AVERAGE RATE OF •60 _ DNA

The time dimension of the DNA hybridization data •o T50H derives from the observed evidence that the same averagerate of sequenceevolution occursin all lin- eagesof birds. We also have some evidence that the same average rate occurs in mammals (Sibley and Ahlquist in press b). This should not be surprising, D. becausethe Uniform AverageRate (UAR) is nothing 30. more (or less) than the inevitable statistical result of averaging over billions of nucleotides and millions of .Different DNA sequencesevolve at many different rates, as the melting curves demonstrate, but the averagerate is the same in all lineages,be- causethe genomeis so large comparedwith the range in the rates of different sequences.Sibley and Ahl- quist (1983:270) provide additional detailsabout the UAR. Our DNA hybridization data fit the definitions o DegreesCelsius of Farris (1981: 6, 13) for "metric" and "ultrametric" 60 65 70 75 30 85 9•0 95 distance measures, i.e. the data never violate the "tri- Fig. 1. Cumulative thermal dissociationcurves of angle inequality" and they are "clocklike." hybrids involving the radio-labeled DNAs of the Becauseall avian genomes evolve at the same av- Long-billed Thrasher (Toxostomalongirostre) and the erage rate, the delta T•oH values are measures of rel- Red-shoulderedGlossy Starling (Lamprotornisnitens). ative time. They may be used, therefore, to recon- Solid circles are the average data points of three ho- struct the branching pattern of a phylogeny. To mologous hybrids of Toxostomalongirostre; solid tri- convert the delta values into absolutetime, it is nec- angles are the average data points of three homolo- essaryto calibrate them against an external dating gous hybrids of Lamprotornisnitens. Open circles are source, viz., fossils or geological events that have the average data points of 20 hybrids between Tox- causedphyletic dichotomies.Using this procedure, ostomaand various starlings;open triangles are the we have obtained an estimate of the calibration such average data points of 16 hybrids between Lampro- that delta T5oH1.0 = about4.3-4.6 million years(my) tornis and various mimines. Smooth curves identified (Sibleyand Ahlquist1981, 1983, in pressb; Sibleyet by letters are averagesof hybrids between both la- al. 1982;Diamond 1983).These constants of propor- beled Toxostomaand labeled Lamprotornisand other tionality are tentative and subjectto correction,but passerinegroups as follows: (A) Turdidae, n = 18; (B) we will use them to provide approximate dates for Cinclus,n = 2; (C) Bombycillidae, n = 6; (D) Super- divergence nodes. families Sylvioideaand Fringilloidea, n = 30; and (E) Parvorder Corvi, n = 15. RESULTS AND DISCUSSION

Tables1-5 and Figs.! and 2 presentthe data from nine experimentalsets in which the Long- because there is no known reason, other than nu- billed Thrasher (T0x0st0malongirostre), the cleotide-sequencehornology, for the thermal stabil- ities of DNA-DNA hybrids. (Mimus polyglottos),the The experimental error in our data has been mea- Red-shoulderedGlossy Starling ( sured, and a single delta T5oHvalue should be as- nitens),and the EuropeanStarling (Sturnusvul- sumedto have a possibleerror of + 1.0 (Sibley and garis)were the radio-labeledtaxa. In Table 1 the Ahlquist 1983: 270). We have found it possibleto delta T5oH values between the Long-billed 236 $IBLEYAND AHLQUIST [Auk,Vol. 101

TAI•LE1. DNA-DNA hybridizationvalues between the radioiodine-labeledsingle-copy DNA of the Long- billedThrasher (Toxostoma longirostre) and the DNAsof otheroscine birds. Under "Group index" M = Mimini, S = Sturnini, T = Turdidae,B = Bombycillidae,C = Cinclidae,Sy = Sylvioidea,F = Fringil- loidea,and Co = .Numbers in parentheses,e.g. (2) = morethan one hybrid averaged.

Commonname Scientificname Delta T5oH Groupindex Long-billedThrasher Toxostomalongirostre 0.0 M BrownThrasher Toxostomarufum 0.9 M Curve-billed Thrasher Toxostoma curvirostre 1.0 M CrissalThrasher Toxostomadorsale (2) 1.3 M Toxostoma redivivum 1.3 M SageThrasher Oreoscoptesmontanus (2) 2.0 M Long-tailedMockingbird Mimuslongicaudatus 2.0 M Northern Mockingbird Mimuspolyglottos (3) 2.1 M TropicalMockingbird Mimusgilvus 2.1 M Gray Dumetellacarolinensis (2) 2.6 M Red-shoulderedGlossy Starling Lamprotornisnitens (3) 5.0 S GreaterGlossy Starling Lamprotornisaustralis 5.2 S SingingStarling Aploniscantoroides 5.4 S MicronesianStarling Aplonisopaca 5.6 S Philippine GlossyStarling Aplonispanayensis (2) 5.7 S Creatophoracinerea (2) 5.7 S Violet Starling Cinnyricinclusleucogaster 5.7 S Hill Myna Graculareligiosa (2) 5.8 S SpotlessStarling Sturnusunicolor (2) 5.9 S EuropeanStarling Sturnusvulgaris 6.0 S Red-wingedStarling Onychognathusmorio 6.0 S African Spreobicolor 6.1 S Acridotherestristis 6.2 S Golden-crestedMyna Ampelicepscoronatus 6.4 S Rufous-tailed Monticola saxatilis 8.2 T Catharusfuscescens 8.3 T White-throatedJungle Flycatcher Rhinomyiasumbratilis 8.3 T Townsend'sSolitaire Myadestestownsendi 8.7 T Black Entomodestescoracinus 8.8 T Western Sialia mexicana 8.8 T Orange-breastedFlycatcher Ficeduladumetoria 8.8 T American Robin Turdusmigratorius 9.0 T CommonNightingale Erithacusmegarhynchos 9.3 T Palm Chat Dulus dominicus 9.9 B American Cinclusmexicanus 10.0 C Phainopeplanitens 10.3 B Cedar Bombycillacedrorum 10.5 B Black-cappedChickadee Parusatricapillus 10.5 Sy Yellow-ventedBulbul Pycnonotusgoiavier 10.5 Sy Passerdomesticus 10.5 F Ploceuscucullatus 10.6 F Black-tailedGnatcatcher Polioptilamelanura 10.6 Sy Red-breastedNuthatch Sittacanadensis 10.7 Sy OrpheanWarbler Sylviahortensis 10.7 Sy A uriparusfiaviceps 10.9 Sy CarolinaWren Thryothorusludovicianus 10.9 Sy Red-billedLeiothrix Leiothrixlutea 11.0 Sy Bushtit Psaltriparusminimus 11.0 Sy SwampSparrow Melospizageorgiana 11.0 F BankSwallow Ripariariparia 11.1 Sy Green-headed verticalis 11.2 F Horned Eremophilaalpestris 11.2 F Black-cappedWren Donacobiusatricapillus 11.3 Sy DuskyMunia Lonchurafuscans 11.3 F MusicianWren Cyphorhinusaradus 11.4 Sy White-eyedVireo Vireogriseus 12.1 Co White-bellied Oriole Oriolussagittatus 12.1 Co excubitor 12.2 Co Shining Monarch alecto 12.7 Co Red-shoulderedCuckoo-shrike Campephagaphoenicea 12.9 Co Corvus mellori 13.5 Co April 1984] Starling-MockingbirdRelationships 237

TABLE2. DNA-DNA hybridizationvalues between the radioiodine-labeledsingle-copy DNA of the North- ern Mockingbird(M#nus polyglottos) and the DNAs of other oscinepasserine birds. Under "Group index" M = Mimini, S = Sturnini, T = Turdidae, C = Cinclidae, B = Bombycillidae,Sy = Sylvioidea, Co = Cor- voidea.

Common name Scientific name Delta T50H Group index Northern Mockingbird M#nus polyglottos 0.0 M SageThrasher Oreoscoptesmontanus 1.9 M Curve-billed Thrasher Toxostoma curvirostre 2.5 M Dutnetellacarolinensis 2.7 M European Starling Sturnusvulgaris 6.2 S Natal Chat Cossyphanatalensis 8.0 T Northern BlackFlycatcher Melaenornisedolioides 8.2 T Eurasian Robin Erithacus rubecula 8.9 T American Robin Turdusmigratorius 9.0 T coracinus 9.3 T Cinclusmexicanus 9.7 C Botnbycillacedrorutn 10.0 B Phainopepla Phainopeplanitens 10.4 B Palm Chat Dulus dominicus 10.4 B Bewick'sWren Thryomanesbewickii 11.4 Sy MusicianWren Cyphorhinusaradus 11.6 Sy Garden Sylviaborin 11.6 Sy Troglodytesaedon 11.7 Sy Black-cappedWren Donacobiusatricapillus 11.8 Sy Asian ParadiseFlycatcher Terpsiphoneparadisi 12.1 Co AmericanCrow Corvusbrachyrhynchos 13.1 Co

Thrasher and the nine other mimines (14 hy- Donacobiusand the wrens are the same genea- brids)range from 0.9 to 2.6,indicating that these logical distancefrom the mimines. speciesare closely related. The 14 species(20 Becauseof the uniform average rate of DNA hybrids)of starlingshave delta T5oHvalues from evolution, the delta T•oH values between all 5.0 to 6.4, and the average is 5.7 + 0.1 SE, sturnidsand the membersof other monophy- +0.4 SD. In Tables 1 and 2 the 21 hybrids be- letic groupsshould be equal, within the limits tween Toxostomaplus Mimus vs. the starlings of experimental error. Therefore, using the range from 5.0 to 6.4 and have an averagedelta clusteringmethod of "averagelinkage," we can T•oH of 5.7 + 0.1 SE, +0.4 SD, as summarized combinethe delta T5oHvalues for hybrids be- in Table 5. tween the members of the Sturnidae and mem- Tables 3 and 4 present the delta T5oHvalues bers of the next nearest cluster, the Turdidae. between the two labeled starlings and the These 28 values range from 8.0 to 9.8, and the mimines. In Table 3 there are 15 hybrids be- average is 8.9 + 0.1, +0.4 (see Table 5). From tween Lamprotornisand the 10 speciesof mi- other data sets, we have 10 DNA hybrids in mines,and in Table 4 there are three hybrids which the labeled taxa are turdids (Turdus, between Sturnusand two speciesof mimines. Myadestes,, Erithacus). These 10 val- These 18 hybrids range from 5.3 to 6.4 and av- ues range from 8.7 to 10.1 and average 9.4 + erage 5.7 +__0.1 SE, +0.4 SD. The reciprocals 0.1, +0.5. The reciprocaldiscrepancy is 0.5, and (5.7:5.7) are identical. Thus, the average delta the average for all 38 sturnid x turdid values T5oHof 5.7 measures the branch node between is 9.0 + 0.1, +0.5. The divergence between the the Sturnini and the Mimini and indicates that Sturnidae and the Turdidae thus occurred about their divergence began about 23-28 million 38-42 mybp, in the late .In our study yearsbefore the present (mybp), in the late Oli- of the (in press a) we obtained an goceneor early Miocene. average of 9.5 for this node, based on the 10 Note that the delta T5oHvalues for Donacobius turdid x sturnid values noted above plus the atricapillusin Tables 1-3 are 11.3, 11.8, and 10.9, data in our 1980 study. which average 11.3. The six delta values for The dippers (Cinclus)represent the next old- typical wrens in thesetables average 11.4;thus er branch from the Sturnidae.By combiningall 238 $IBLEYAND AHLQUIST [Auk, Vol. 101

TABLE3. DNA-DNA hybridization values between the radioiodine-labeledsingle-copy DNA of the Red- shoulderedGlossy Starling (Lamprotornisnitens) and the DNAs of other oscine passerinebirds. Under "Group index" S = Sturnini, M = Mimini, T = Turdidae, C = Cinclidae, B = Bombycillidae,F = Fringilloi- dea, Sy = Sylvioidea, and Co = Corvoidea. Numbers in parentheses,e.g. (2) = more than one hybrid av- eraged.

Common name Scientific name Delta TsoH Group index Red-shoulderedGlossy Starling Lamprotornisnitens 0.0 S African Pied Starling Spreobicolor 1.0 S Red-wingedStarling Onychognathusmorio 2.2 S Violet Starling Cinnyricinclusleucogaster 2.5 S Common Myna Acridotherestristis 3.6 S Wattled Starling Creatophoracinerea (2) 3.7 S SpotlessStarling Sturnusunicolor (3) 3.7 S EuropeanStarling Sturnusvulgaris 3.8 S Golden-crestedMyna Ampelicepscoronatus 4.6 S Aploniscantoroides (3) 4.8 S Aplonisopaca 4.9 S Hill Myna Graculareligiosa 4.9 S Philippine GlossyStarling Aplonispanayensis 5.1 S Long-billed Thrasher Toxostomalongirostre (3) 5.3 M SageThrasher Oreoscoptesmontanus 5.4 M Toxostomadorsale (2) 5.5 M Northern Mockingbird Mimus polyglottos(2) 5.5 M Gray Catbird Dumetellacarolinensis (2) 5.6 M Curve-billed Thrasher Toxostoma curvirostre 5.7 M California Thrasher Toxostoma redivivum 5.7 M Long-tailed Mockingbird Mimus longicaudatus 5.9 M Toxostomarufum 6.0 M Mimus gilvus 6.1 M Townsend'sSolitaire Myadestestownsendi 8.4 T Rufous-tailed Rock-Thrush Monticola saxatilis 8.6 T White-throatedJungle Flycatcher Rhinomyiasumbratilis 8.9 T Veery Catharusfuscescens 9.1 T Erithacusmegarhynchos 9.1 T Sunda Whistling-Thrush Myiophoneusglaucinus 9.4 T Orange-breastedFlycatcher Ficeduladumetoria 9.4 T Sialia mexicana 9.6 T American Robin Turdusmigratorius 9.8 T Palm Chat Dulus dominicus 9.9 B American Dipper Cinclusmexicanus 10.2 C Cedar Waxwing Bombycillacedrorum 10.6 B Phainopepla Phainopeplanitens 10.6 B Village Weaver Ploceuscucullatus 10.6 F Yellow-vented Pycnonotusgoiavier 10.8 Sy Black-cappedChickadee Parusatricapillus 10.9 Sy Black-cappedWren Donacobiusatricapillus 10.9 Sy Red-billed Leiothrixlutea 11.0 Sy Orphean Warbler Sylviahortensis 11.2 Sy House Sparrow Passerdomesticus 11.4 F Eremophilaalpestris 11.4 F Melospizageorgiana 11.6 F Carolina Wren Thryothorusludovicianus 11.6 Sy Green-headed Sunbird Nectarinia verticalis 11.8 F Bank Ripariariparia 12.0 Sy Red-breastedNuthatch Sitta canadensis 12.2 Sy Lonchurafuscans 12.2 F Little Raven Corvus mellori 12.3 Co Northern Shrike Lanius excubitor 12.3 Co White-eyed Vireogriseus 12.3 Co carunculata 12.4 Co White-bellied Oriole Oriolussagittatus 12.5 Co White-browed Artamussuperciliosus 12.8 Co African Dicrurusadsimilis 13.1 Co Red-shoulderedCuckoo-shrike Campephagaphoenicea 13.2 Co Shining Monarch Myiagraalecto 13.2 Co April 1984] Starling-MockingbirdRelationships 239

TABLE4. DNA-DNA hybridization values between the radioiodine-labeledsingle-copy DNA of the Euro- pean Starling (Sturnusvulgaris) and the DNAs of other oscinepasserine birds. Under "Group index" S = Sturnini, M = Mimini, T = Turdidae, C = Cinclidae, B = Bombycillidae,F = Fringilloidea, $y = Sylvioidea, and Co = Corvoidea.The number in parentheses(2) = two DNA hybrids were averaged.

Common name Scientific name Delta T5oH Group index EuropeanStarling Sturnusvulgaris 0.0 S African Pied Starling Spreobicolor 3.9 S Singing Starling Aploniscantoroides 5.9 S Gray Catbird Dumetellacarolinensis (2) 6.3 M Northern Mockingbird Mimuspolyglottos 6.4 M Eurasian Robin Erithacus rubecula 8.6 T American Robin Turdusmigratorius 8.7 T Large-billed Blue Flycatcher Niltava caerulata 8.8 T Northern Black Flycatcher Melaenornisedolioides 8.8 T -eating Chat Myrmecocichlaformicivora 9.1 T American Dipper Cinclusmexicanus 9.7 C Cedar Waxwing Bombycillacedrorum 10.7 B Ploceuscapensis 10.8 F Red-backed Mannikin bicolor 10.9 F Pale White-eye Zosteropspallida 11.4 Sy Sylviaborin 11.5 Sy Short-tailed Babbler Trichastomamalaccense 11.7 Sy Black-napedOriole Orioluschinensis 11.9 Co Black Cracticusquoyi 12.0 Co Black-faced Woodswallow Artamus cinereus 12.0 Co Spot-winged Monarch Monarchaguttula 12.0 Co Corvusbrachyrhynchos 12.1 Co RufousShrike-thrush Colluricinclamegarhyncha 12.3 Co Brown Whistler Pachycephalasimplex 12.3 Co Willie Rhipiduraleucophrys 12.6 Co Common Wood-shrike Tephrodornispondicerianus 12.7 Co Red-eyed Vireo Vireo olivaceus 13.0 Co Asian ParadiseFlycatcher Terpsiphoneparadisi 13.3 Co Red-backed Shrike Lanius collurio 13.5 Co Large -shrike Coracinanovaehollandiae 13.9 Co Lauterbach'sBowerbird Chlamyderalauterbachi 14.0 Co

available delta values for the Sturnidae and parvorders Muscicapae and Corvi (12.7) di- Turdidae, to and from Cinclus, we have 15 mea- verged about 60 mybp. surements with an average of 9.8 + 0.1, +0.5 Figure 1 presentsa series of thermal disso- for the divergence of the Cinclidae from the ciation curves between the tracer DNAs of Tox- sturnid-turdid lineage. This branch probably ostomaand Lamprotornisand other oscine taxa. occurred about 42-45 mybp, in the Eocene. It is apparent from these melting curves that The oldest branch in the Turdoidea led to the Toxostornaand Larnprotornisare the mostclosely living members of the Bombycillidae, which related taxa in this set, that the other turdoids includeswaxwings (Bornbycilla),silky flycatch- are progressivelymore distant, that the Syl- ers (Phainopepla,Phainoptila, Ptilogonys), and the vioidea and Fringilloidea are still more distant, Palm Chat (Dulus dominicus).We have 52 delta and that the Corvi are the most distant from T5oHvalues between the bombycillids and the the Sturnidae. other turdoids, which average 10.7 + 0.1, Figure 2 is basedon the data in Tables 1-5 +0.4. This indicates that the ancestor of the and the additional data noted in the text. All Bombycillidae branched from the lineage lead- DNA-DNA distance measurements are be- ing to the other turdoids about 46-50 mybp, in tween a radio-labeled species and the unla- the mid-Eocene. beled species, not between the unlabeled The sylvioid-fringilloid (11.3) branched species.Thus, the delta values between Larn- from the turdoid clade about 50 mybp, and the protornisand Spreo,Onychognathus, and other 240 SIBLEYAND AHLQUIST [Auk, Vol. 101

o

lO

20 , .5.1[5.7 [n2.( 39 30>--

9.0 n 38 _40

10_ 9.8n15 10.7n52 I 11.3 n42' I _50 12_ 12.7n31 _ 60

14_ I

Fig. 2. Diagramof relationshipsof and amongthe Sturnidae.Branch lengths at nodesare taken from the delta TsoHvalues given in Tables1-5. Solid circlesindicate radio-labeled taxa. MYA = millionsof yearsago.

unlabeled taxa do not indicate how these taxa as central Alaska (Kessel and Gibson 1978). are related to one another, exceptthat they are During the late and early Miocene, no more distant from one another than they however, there were emergent land connec- are from the tracer species.A delta value be- tions between Europe and , and tween two labeled taxa (e.g. the 5.7 averagebe- the climate in the northern hemisphere was tween the sturnines and mimines), however, is considerablymilder than today. Eldholm and a valid measurement of the genealogical dis- Thiede (1980) concludedthat "connectionsby tance between them. Note that the Sage shallow water and possiblyemerged, or locally Thrasher (Oreoscoptesmontanus) is closer to Mi- emerged,continent existedto the Middle Oli- musthan to the true thrashers (Toxostoma).Gul- gocene between Greenland and Svalbard [= ledge (1975) cameto the sameconclusion. Spitsbergen]and to the Plioceneon the Faeroe- We have suggestedthat the sturnine-mimine Iceland Ridge." Gradstein and Srivastava(1980) dichotomy occurred in the late Oligocene or noted that relatively warm Atlantic waters oc- early Miocene, about 23-28 mybp. This implies curred in the Labrador Sea in the Eocene and that their common ancestralpopulation was di- that "less warm Atlantic water influence pre- vided into two separatepopulations at that time. vailed in Oligocene and Miocene time but it The presentbreeding distributions of the Mim- was not until late Miocene time that there is ini are far south of the Bering Strait and the evidence of a cold Labrador Current hugging north Atlantic, and among the Sturnini only the Canadian shelves."Similarly, Wolfe (1980) Sturnusvulgaris breeds in northern Europeand found evidence of a mixed broad-leaved decid- Iceland (Voous 1960). In North America the in- uous forest at high latitudes in the northern troduced S. vulgarisrarely breedsas far north hemisphere during the Oligocene and Mio- April 1984] Starling-MockingbirdRelationships 241

cene,with no evidenceof glaciationuntil "late in the Neogene." It thus seemsprobable that the common ancestor of the starlings and mockingbirds was able to spread over the northern hemisphere in the Eocene and that +• the increasingly severe boreal climate in the late Oligocene-Miocenepushed the two pop- ulations southward and out of contact.

CONCLUSIONS

The DNA-DNA hybridization measurements indicate that the starlingsand mockingbirdsare each other's closestliving relatives. Compari- sons of their DNAs with those of all other liv- ing groupsof oscinepasserine birds show that there is no other group that is more closely related to one of them than they are to each other. The morphological charactersknown to be sharedby the Sturnini and the Mimini would not, alone, make it possible to recognize the closegenealogical relationship between the two groups, but, in conjunction with Stallcup's (1961) serological evidence, they support the DNA hybridization data. No other proposal about the relationships of the starlings and mockingbirdshas been based on evidence of equal or greater competence,and we suggest that additional congruencieswill be discov- ered, now that the DNA comparisons have made it possibleto reconstructthe phylogeny of these taxa.

ACKNOWLEDGMENTS

For assistancein the laboratory we thank L. Wal- lace,E. Nowicki, R. Gatter,and F. C. Sibley. For other help we thank A. Archer, M. Avedillo, R. Barker, A. H. Bledsoe,M. Bull, J. dupont, S. Furhiss,P. Garay- aide, A. Garza de Leon, P. Ginn, J. Gulledge, P. V. Kainady, J. Kennard, S. Koster, R. Liversidge, G. Mo- rel, J. O'Neill, T. O. Osborne, R. Parker, N. Procter, W. C. Russell,R. Semba,F. Sheldon, F. C. Sibley, J. Spendelow,N. and E. Wheelwright, and D. Wysham. The laboratorywork was supportedby Yale Univer- sity and the National ScienceFoundation (DEB-79- 26746).

LITERATURE CITED

AMERICAN ORNITHOLOGISTS' UNION. 1886. Code of nomenclature and check-list of North American birds. New York, Amer. Ornithol. Union. 1895. Check-list of North American birds, second ed. New York, Amer. Ornithol. Union. 242 $IBLEYAND AHLQUIST [Auk,Vol. 101

1910. Check-list of North American birds, . Abstracts, 100th meeting, Amer. Orni- third ed. New York, Amer. Ornithol. Union. thol. Union. 1931. Check-list of North American birds, Cours, E. 1896. Key to North Americanbirds, fourth fourth ed. Lancaster,Pennsylvania, Amer. Or- ed. Boston, Estes & Lauriat. nithol. Union. DAVIS,J., & A. H. MILLERß1960. Family Mimidae. 1957. Check-list of North American birds, Pp. 440-458 in Check- of the world, fifth ed. Baltimore,Maryland, Amer. Ornithol. vol. 9. (E. Mayr and J. C. Greenway,Jr., Eds.). Union. Cambridge,Massachusetts, Mus. Comp. Zool. 1983. Check-list of North American birds, DELACOUR,J., & C. VAURIE. 1957. A classification of sixth ed. Lawrence, Kansas, Amer. Ornithol. the oscines (Aves). Los Angeles County Mus., Union. Contrib. Sci. 16: 1-6. AMADON,D. 1943. The genera of starlings and their DIAMOND,J. M. 1983. by nucleotides. relationships. Amer. Mus. Novitates 1251: 1-21. Nature 305: 17-18. ß 1956. Remarkson the starlings,family Stur- ELDHOLM,O., & J. THIEDE. 1980. Cenozoic continen- nidae. Amer. Mus. Novitates 1803: 1-41. tal separationbetween Europe and Greenland. 1957. Remarks on the classification of the Palaeogeogr.,Palaeoclim., Palaeoecol.30: 243- perching birds [order Passeriformes].Proc. Zool. 259. Soc.Calcutta, Mookerjee Memor. Vol.: 259-268. FAe,•iS,J.S. 1972. Estimatingphylogenetic trees from ß 1962. Family Sturnidae.Pp. 75-121 in Check- distance matrices. Amer. Nat. 106: 645-668. list of birds of the world, vol. 15. (E. Mayr and 1981. Distancedata in phylogeneticanaly- J. C. Greenway,Jr., Eds.).Cambridge, Massachu- sis.Pp. 3-23 in Advancesin .(V. A. Funk setts,Mus. Comp. Zool. and D. R. Brooks, Eds.). New York, New York BAIRD,S.F. 1858. U.S. Pacificrailroad surveys.Vol. Botanical Garden. 9. Birds. Washington,D.C. GRAY,G. R. 1869. Hand-list of genera and species 1864. Review of American birds. Part 1. of birds. Part 1. London, Trustees, British Mus. Smithson. Misc. Collections. Washington, D.C., 1870. Hand-list of genera and speciesof Smithsonian Instit. birds. Part 2. London, Trustees, British Mus. BEECHER,W. J. 1953. A phylogeny of the oscines. GRADSTEIN,F. M., & S. P. SRIVASTAVA.1980. Aspects Auk 70: 270-333. of Cenozoicstratigraphy and paleoecologyof the ß 1978. Feeding adaptationsand evolution in Labrador Sea and Baffin Bay. Palaeogeogr.,Pa- the starlings. Bull. Chicago Acad. Sci. 11: 269- laeoclim., Palaeoecol. 30: 261-295. 298. GRULA,J. W., T. J. HALL, J. A. HUNT, T. D. GIUGNI, G. BERNDT,R., & W. MEISE. 1962. Naturgeschichte der J. GRAHAM,E. H. DAVIDSON,& R. J. BRITTIN. 1982. VOgel. Vol. 2. Stuttgart,Kosmos. Sea urchin sequencevariation and reduced in- BocK, W.J. 1962. The pneumatic fossa of the hu- terspecies differences of the less variable se- merus in the Passeres. Auk 79: 425-443. quences.Evolution 36: 665-676. 1963. Relationshipsbetween the birds of GULLEDGE,J.L. 1975. A study of phenetic and phy- paradise and the bower birds. Condor 65: 91- logenetic relationships among the mocking- 125. birds, thrashers, and their allies. Unpublished Ph.D. dissertation.New York, City Univ. New BOETTICHER,H. VON. 1931. Die verwandtschaftlich- York. systematischeStellung der WebervOgel(Ploce- KESSEL,g., & D. G. GIBSONß 1978. Status and distri- idae) zu den FinkenvtSgeln(Fringillidae) und bution of Alaska birds. Studies Avian Biol. Staren (Sturnidae). Senckenbergiana 13: 147-153. No. 1. BONAPARTE,C.L. 1850. Conspectusgenerum avium. KOHNE,D. E. 1970. Evolution of higher-organism Vol. 1. Lugduni Batavorum. DNA. Quart. Rev. Biophysics33: 327-375. BOND, J. 1971. Birds of the West Indies, third ed. LucAs, F. A. 1888. Notes on the osteologyof the Boston,Houghton-Mifflin. thrushes, Miminae, and wrens. Proc. U.S. Natl. BRITTIN,R. J., & D. E. KOHNE. 1968. Repeated se- Mus. 11: 173-180. quencesin DNA. Science 161: 529-540. MAYR, E., & D. AMADON. 1951. A classification of --, D. E. GRAHAM, & B. R. NEUFELD. 1974. Anal- Recent birds. Amer. Mus. Novitates 1496: 1-42. ysis of repeating DNA sequencesby reassocia- , & J. C. GREENWAY,JR. 1956. Sequence of tion. Pp. 363-418 in Methodsin enzymology,vol. passerinefamilies (Aves). Breviora58:1-11. 29 (L. Grossman and K. Moldave, Eds.). New ß & --. 1960. Check-list of birds of the York, Academic Press. world, vol. 9. Cambridge, Massachusetts,Mus. CA•3ANIS,J. 1850. Museum Heineanum. Part 1. Hal- Comp. Zool. berstadt, Gut St. Burchard. , & --. 1962. Check-list of birds of the CLENCH,M. H., J. L. GULLEDGE,& K. C. PARKES.1982. world, vol. 15. Cambridge, Massachusetts,Mus. The Black-cappedDonacobius is a wren, not a Comp. Zool. April 1984] Starling-MockingbirdRelationships 243

MILLER,A. H. 1964. Mockingbird. Pp. 479-481 in A --, & . In pressb. The phylogenyof the new dictionary of birds. (A. L. Thomson, Ed.). hominoid primates, as indicated by DNA-DNA London, Thos. Nelson & Sons. hybridization. J. Mol. Evol. MORIOKA,H. 1967. and relationshipsof --, G. R. WILLIAMS,&: J. E. AHLQUIST. 1982. The thrushes, mimic thrushes, dippers, and wrens. relationships of the wrens (Acan- Unpublished Ph.D. dissertation. Urbana, Illi- thisittidae) as indicated by DNA-DNA hybrid- nois, Univ. Illinois. ization. Notornis 29: 113-130. PARKER,W.K. 1875. On aegithognathousbirds, part STALLCUPßW. B. 1961. Relationshipsof some fami- 1. Trans. Zool. Soc. London 9: 289-352. lies of the suborder Passeres() as in- REICHENOW,A. 1882. V6gel der ZoologischenGar- dicatedby comparisonsof tissueproteins. J. Grad. ten. LeipzigßL. A. Kittier. Res. Center, So. Methodist Univ. 29: 43-65. 1914. Die V6gel, vol. 2. StuttgartßF. Enke. STEJNEGER,L. 1883. Remarks on the systematicar- RIDGWAY, R. 1907. Birds of North and Middle rangementof the American Turdidae. Proc. U.S. America, part 4. Washington, D.C., Gov't. Print- Natl. Mus. 6: 449-483. ing Office. 1885. Standardnatural history. Boston,Cas- SCLATER,P. L. 1880. Remarks on the present state sino & Co. of the systemaavium. Ibis 22: 340-350; 399-411. STONOR,C. R. 1938. Some features of the variation SHARPE,R.B. 1890. Catalogueof birds in the British of the birds of paradise.Proc. Zool. Soc.London Museum, vol. 13. London, Trustees, British Mus. 109B: 417-481. ß 1891. A review of recentattempts to classify STORER,R.W. 1971. Classificationof birds. Pp. 1-18 birds. Pp. 1-90 in Proc. 2nd Intern. Ornithol. in Avian biology, vol. 1. (D. S. Farner, J. L. King, Congr., Budapest. and K. C. Parkes, Eds.). New York, Academic SHUFELDT, R. W. 1889. Contributions to the com- Press. parative osteologyof North American Passeres. STRESEMANN,E. 1934. Aves. In Handbuch der Zool- J. Morphol. 3: 81-114. ogie, vol. 7, part 2. Berlin, Waiter de Gruyter. SIBLEY,C. G., &: J. E. AHLQUIST. 1980. The relation- SUNDEVALL,C.J. 1872. Methodi naturalis avium dis- shipsof the "primitive insecteaters" (Aves: Pas- ponendarum tentamen. Stockholm, Samson & seriformes) as indicated by DNA-DNA hybrid- Wallin. (English transl. F. Nicholson. 1889. Lon- ization. Pp. 1215-1220 in Proc. 17th Intern. donß R. H. Porter). Ornithol. Congr. (R. N6hring, Ed.). Berlinß Voous, K.H. 1960. Atlas of Europeanbirds; English Deutsche Ornithol. Gesellsch. ed. London, Thos. Nelson & Sons. , & --. 1981. The phylogeny and rela- 1977. List of Holarctic species.- tionshipsof the ratite birdsas indicatedby DNA- ines. Ibis 119: 223-250; 376-406. DNA hybridization. Pp. 301-335 in Evolution to- WALLACEßA.g. 1874. On the arrangement of the day, Proc.2nd Intern. Congr. Syst.Evol. Biol. (G. families constitutingthe Order Passeres.Ibis 22: G. E. Scudderand J. L. Reveal, Eds.). Pittsburgh, 406-416. Pennsylvaniaß Hunt Inst. Botanic. Document., WARNER,R.W. 1972. The anatomy of the in Carnegie-Mellon Univ. passerinebirds. J. Zool. Soc. London 168: 381- ß & -- 1982. The relationships of the 393. Hawaiian honey creepers (Drepaninini) as in- WETMORE,A. 1930. A systematicclassification for dicated by DNA-DNA hybridization. Auk 99: the birds of the world. Proc. U.S. Natl. Mus. 76: 130-140. 1-8. ß& --. 1983. The phylogeny and classi- 1951. A revised classification for the birds fication of birds based on the data of DNA-DNA of the world. Smithson. Misc. Coil. 117: 1-22. hybridization. Current Ornithol. 1: 245-292. 1960. A classification for the birds of the ß & --. In press a. The phylogeny and world. Smithson. Misc. Coil. 139: 1-37. classificationof the passefine birds, based on WOLFE,J. A. 1980. Tertiary climatesand floristic re- comparisonsof the geneticmaterial, DNA. Proc. lationships at high latitudes in the northern 18th Intern. Ornithol. Congr., Moscow, Nauka hemisphere. Palaeogeogr., Palaeoclim., & Pa- Publ. laeoecol. 30: 313-323.