sign in sign up
<<
Home , Aphelocoma, Corvidae, Cyanocorax, Cyanolyca, Perisoreus, Unicolored jay

The Auk 109(1):133-147, 1992

PHYLOGENY AND RATES OF MOLECULAR EVOLUTION IN THE JAYS ()

A. TOWNSEND PETERSON• Committeeon Evolutionary Biology, University of Chicago,Chicago, Illinois 60637, USA

ABSTRACT.--Iexamined hypotheses of Aphelocomajay phylogenyderived from allozyme data.Results from various algorithms differ in details,but the overallpatterns are consistent: ScrubJays (A. coerulescens)and UnicoloredJays (A. unicolor)were derivedindependently from differentpopulations of Gray-breastedJays (A. ultramarina).Within ScrubJays, the californicasubspecies group was derivedfrom the populationsof interior (woodhouseiigroup). One UnicoloredJay population and two ScrubJay populations, all strong- ly differentiated,are placedconsistently at the baseof the phylogeny,but phenotypic,bio- geographic,and theoretical evidence suggests that thesepopulations represent rapidly evolv- ing populationsderived from within populationsof theirrespective . Because analyses of ratesof molecularevolution demonstrate significant rate heterogeneity,I suggestthat the applicationof a molecularclock to date-splittingevents in the Aphelocomajays is not a valid approach.Received 18 February1991, accepted 3 July 1991.

THE THREEspecies of Aphelocomajays range erra Madre Occidental);and (3) ultramarinagroup throughout western and southern North Amer- (TransvolcanicBelt). Finally, the UnicoloredJay ica and northern (Fig. 1; Pi- (A. unicolor)consists of five allopatric popula- telka 1951). Scrub Jays (A. coerulescens)range tions, each a separatesubspecies, in southern from and Wyoming south to the Isth- and northern Central America (Pitelka musof Tehuantepec,with disjunctpopulations 1951). on Santa Cruz Island off the coast of southern The Aphelocomajays have been the subjectof and in peninsular . The 15 numerous comparativestudies. Evaluations of subspeciesform five groups,each characterized socialsystems in the genushave led to advances by unique combinationsof , morpho- in understanding ecologicalfactors important logical, and behavioral characters(Fig. 1; Pi- in the evolution of sociality(Woolfenden and telka 1951,Peterson 1991a): (1) woodhouseiigroup Fitzpatrick 1984, Fitzpatrick and Woolfenden (Wyoming and southeastern Oregon south 1986,Brown 1987,Peterson and Burt, in press). through and along both sides of Differential use in ScrubJays (Peterson RockyMountains, and then alonginterior slopes and Vargas1992) is correlatedwith geographic of and Sierra Madre Oc- variation in shape, suggestingthat beak cidental of northern Mexico to southern Chi- shapeshave respondedto natural selection(Pe- huahuan Desert and vicinity of Mexico City); terson, in prep.). Integration of phylogenetic (2) californicagroup (western Oregon, Califor- information into suchinvestigations will allow nia, and BajaCalifornia); (3) sumichrastigroup an important new dimensionof understanding (southern Mexico); (4) coerulescensgroup (pen- (Brooks and McLennan 1990). Hence, I have insular Florida); and (5) insularisgroup (Santa attemptedto estimatethe phylogeny of the dif- Cruz Island). The Gray-breastedJay (A. ultra- ferentiated forms in the . marina) ranges throughout the mountains of Ratesof molecularevolution.--Since the pub- northern and central Mexico and the south- lication of the influential paper of Zuckerkandl western United States.The sevensubspecies fall and Pauling (1962), the idea of a "molecular into three groupscharacterized by unique com- clock" has been controversial in molecular bi- binations of morphological and behavioral ologyand .The clockconcept is based characters(Fig. 1;Pitelka 1951 ): (1) potosina group on the assumptionof a uniform rate of molec- (SierraMadre Oriental); (2) wollweberigroup (Si- ular evolution in a group. If the uniform rate assumptionwere correct, the accumulationof geneticdifferentiation between sister taxa would •Present address:Department of Zoology, Field be time-invariant, and divergence times could Museum of Natural History, RooseveltRoad at Lake be estimatedfrom genetic distances. Shore Drive, Chicago,Illinois 60605, USA. The clock conceptis an important feature of 133 134 A. TOWNSENDPETERSON [Auk, Vol. 109 many sectors of molecular systematics.At- tempts have been made to calibrate clocksfor a variety of taxa and biochemicaldata sets(e.g. Sarich 1977, Wilson et al. 1977, Sibley and Ahlquist 1981). Several commonly used tree- building algorithmsdepend on the assumption of a clock (Felsenstein 1982, 1989). The molec- ular clock is now a common method of dating divergencetimes between taxa (e.g. Zink 1982, Nevo et al. 1987, Johnson and Marten 1988). Hence, a uniform rate of molecular evolution and the consequentexistence of a clock can be important assumptionsin molecular systemat- ics. However, the assumption that evolutionary rates are homogeneousoften goes untested. If molecular evolutionary rates differ among lin- eages,the clockbecomes unreliable. Equality of rates in sister lineages can be tested with the relative-rate test: in the three-taxon statement ((A, B), C), A and B should show similar degrees of differentiation from C (Wilson et al. 1977). Significant departure from equality indicates that A and B have evolved at different rates. Rateuniformity hasbeen testedin a number of taxa, with some studies documenting homo- geneity (e.g. Sibley and Ahlquist 1983, Bledsoe 1987), and others finding heterogeneity (e.g. Britten 1986, Sibley et al. 1987, Sheldon 1987, Springerand Kirsch 1989).A uniform rate and molecularclock are commonlyand uncritically assumedin systematicstudies, so a test of the assumptionof rate uniformity was desirablein my study.

METHODS

Samplingdesign.--During 1986-1989, I studiedand collectedAphelocoma jays at 35 sites in the United Statesand Mexico.Tissue from the endangeredFlor- ida population of Scrub Jayswas salvagedfrom re- Fig. 1. Summaryof collectinglocalities and sub- cently dead individualsat nestsand roadsides.Sam- speciesgroups of three speciesof Aphelocomajays; ples of heart, liver, and pectoralmuscle from each Florida range indicatedin upper right-hand corner. individual were stored in cryotubesin liquid nitro- (A) ScrubJays: open circle, Santa Cruz Island(ACINS); gen. All tissuewas depositedin the FrozenTissue closedcircles; californica group (CO CAL); closedtri- Collection of the Field Museum of Natural History. angles,northern woodhouseii group (CO WOO); open In total,I analyzed615 Aphelocomajays (458 Scrub triangles, southern woodhouseiigroup (CO MEX); Jays,138 Gray-breasted Jays, and 19 UnicoloredJays). squares, sumichrastigroup; dotted circle, Florida Specimensof Scrub,Gray-breasted, and Unicolored (ACCO2).(B) Gray-breasted Jays: triangles, wollweberi jayswere collectedat 24, 8, and 3 sites,respectively, group (UL WOL); circles,potosina group (UL POT); in the United Statesand Mexico (Table 1, Fig. 1). I squares,ultramarina group (UL ULT). (C) Unicolored attemptedto obtainsamples of 20 individualsat each Jays:triangles, east slope populations (UN E); circle, site to maximize tree stability (Archie et al. 1989): west slope population (ANGUE). ScrubJay samples averaged 19.2 (range 13-27) indi- viduals; Gray-breastedJay samplesaveraged 17.3 (range7-22) individuals;and Unicolored samples January1992] Jay Phylogenyand EvolutionaryRates 135

TABLE1. Samplelocalities and samplesizes (complete locality data in Peterson1990).

Sample Group Sample Locality size Scrub Jays CO CAL ACCAL CA, Monterey Co., Bradley 24 ACCAU CA, Trinity Co., Hyampom 16 ACHY3 BCS,Sierra de la Laguna 16 ACIMM OR, Adair Village 15 ACOBc CA, Orange Co., Silverado Canyon 16 ACOBk BCN, Rosade Castilla (Ojos Negros) 20 ACOOC CA, San Ramon 16 ACSP1 CA, Alturas 22 ACSP3 CA, Bodfish 19 CO MEX ACCY1 COA, E1 Diamante Pass 21 ACCY2 SLP, Sierra de Bledos 21 ACGR! DUR, Villa Ocampo 21 ACGR3 JAL, Lagosde Moreno 20 CO WOO ACNE! NV, Toiyabe Mountains 14 ACNE4 NV, Mt. Charleston 27 ACNE6 AZ, Drake 25 ACTEX TX, EdwardsPlateau, Carta Valley 22 ACWO! CO, Front Range,Gardner ! 1 ACWO3 NM, Manzano Mountains 26 ACWO4 TX, Davis Mountains 21 CO SUM ACREM GRO, Xocomanatl/•n 13 ACSM2 OAX, Sierra de los Mijes, Zempoalt•petl 20 ACINS CA, Santa Cruz Island 14 ACCO2 FL, Archbold Biol. Station 18 Gray-breastedJays UL WOL AUAZ1 AZ, Pefia Blanca Lake 16 AUGRA JAL, Sierra de Bolafios 7 AUWO1 ZAC, Valparaiso 20 UL POT AUCO2 COA, Sierra E1 Carmen 22 AUPO2 SLP, Sierra de Bledos 22 AUPO3 HID, Jacala 20 UL ULT AUULT MOR, Huitzilac 14 AUCOL JAL, Sayula 17 Unicolored Jays ANGUE GRO, Sierra de Atoyac 7 UN E ANOAX OAX, Sierra de los Mijes, Zempoalt•petl 10 ANCON HID, Sierra de la Huasteca, Tlanchinol 2

averaged6.3 (range2-10) individuals.To avoid prob- largerinfraspecific divisions, I dividedthe speciesin lemswith nonindependenceof individualgenotypes a manner equivalentto the subspeciesgroups of Pi- when relatedindividuals (e.g. parents and offspring, telka (1951),except that the woodhouseiigroup was multiplesiblings) were collected, I includedonly one dividedinto northern(United Statespopulations) and memberof each potentially related pair of individ- southern(Mexican populations)parts (Fig. I, Table uals. 1). For conveniencein referring to populationsam- To provide outgroupsfor phylogeneticanalyses, I ples, I assignedeach a five-letter code,in which the included70 individualsof 17 other corvidspecies. To first letter indicatesthe genus,the secondindicates maximizedetection of plesiomorphicalleles for po- the species,and the remainingthree indicatesub- tentially closely related genera (Hardy 1969, Hope speciesand (if necessary)the samplefrom that sub- 1989),I includedall materialavailable (up to 20 in- species(e.g. ACNE2 is the secondsample of A__phelo-dividuals) of eachof the followingspecies: coma coerulescensnevadae; Table I). For reference to cristata,C. stelleri, mirabilis, C. cucullata,C. 136 A. TOWNSENDPETERSON [Auk,Vol. 109 viridicyana,and Gymnorhinuscyanocephalus. For less in the scleroticring; Curtisand Miller 1938)not pre- closelyrelated corvid genera,I included one or two sent in Gymnorhinus(Peterson, unpubl. data). While individualsof eachspecies available: Calocitta formosa, Cyanocittahas a number of unique features,Aphelo- C. colliei, violaceus, C. cyanomelas,C. yncas, comais not sowell defined.Because it seemsunlikely Corvusbrachyrhynchos, C. cryptoleucus, pica, P. nut- that Cyanocittawas derived from within Aphelocoma, tallii, Nucifragacolumbiana, and Perisoreuscanadensis. I assumefor the purposesof this paper that Aphelo- Electrophoresis.--Ihomogenized approximately 0.4 g comais monophyletic. each of heart, muscle,and liver from each sample I performedphenetic analyses on matricesof mod- in a 1 ram disodium EDTA/100 ram Trizma base/0.2 ified Rogers'genetic distances (as modified by Wright ram NAD, NADP, and ATP buffer.Homogenates were 1978),which were calculatedusing programs devel- centrifuged at 12,000 rpm for 30-45 min. Superna- opedby ScottM. Lanyon.Allele-frequency data and tants were drawn into capillary tubes,and stored at genetic-distancematrices are available in Peterson -76øC for later use. Gels for electrophoresiswere (1990) or on request from the author. Exploratory madeof 10-12%starch in the appropriatebuffer. After analysesbased on other genetic distancemeasures electrophoresis,gels were sliced, and stained differ- either produced similar results (e.g. Cavalli-Sforza entiallyfollowing Shaw and Prasad(1970) and Harris chorddistance) or were difficultto interpret(e.g. Nei's and Hopkinson (1978). Loci studied included ACON geneticdistance), so I report only resultsfrom Rogers' (enzymenumber 4.2.1.3;Harris and Hopkinson 1978), distances.Two different clusteringmethods, the un- ACP (3.1.3.2), ADA (3.5.4.4), ADH (1.1.1.1), AK (2 loci, weighted pair-group method (UPGMA) and single- 2.7.4.3), CK (2 loci, 2.7.3.2), ES (2 loci, 3.1.1.1), FUM linkagecluster analysis, were appliedto the genetic (2 loci, 4.2.1.2), GDA (3.5.4.3), GOT (2.6.1.1), GPD distancematrix using NTSYS-pc(version 1.40; Rohlf (1.1.1.8), GPI (5.3.1.9), IDH (1.1.1.42), LDH (2 loci, 1988). 1.1.1.27), MDH (2 loci, 1.1.1.37), MPI (5.3.1.8), PEP One potentialproblem with strictlyphenetic meth- (leucyl-alanine,2 loci; leucyl-aminopeptide,1 locus; ods is the assumptionof equal evolutionaryrates in 3.4.11), PGD (1.1.1.44), PGM (2.7.5.1), SOD (1.15.1.1), eachlineage (Felsenstein 1982). A firststep in relaxing and SORDH (1.1.1.14).Details of electrophoreticcon- this assumptionis the Fitch-Margoliashalgorithm, ditions are given in Peterson (1990). Loci showing which allowsfor somevariation in evolutionaryrates difficult-to-interpret patterns of variation were ex- (Fitch and Margoliash 1967, Felsenstein1989). I used cluded. Stained gels were digitized with an optical Steller's Jays (Cyanocittastelleri, CYSTE) as the out- imaging system, and visual images were stored in group,because it is sufficientlyclose phylogenetically computer files; gels were later discarded. to be informative, and the sample size (20) was suf- To assurecorrect assignment of homologies,I in- ficiently large to permit accurateestimation of gene cludedreference individuals on eachgel. Allelesfrom frequencies.Because of limited computermemory, different populationswere not consideredequivalent the entire dataset could not be analyzedat one time, until found indistinguishablein side-by-sidecom- soI analyzedsubsets of the datain two ways.I either parisonson the same gel. I recheckedhornology as- assumedspecies to be monophyleticand analyzed signmentsfor each locus under different buffer con- eachseparately, or I usedrepresentative populations ditions and/or at maximum migration distances(an from each of the subspeciesgroups. All trees were averageof two recheckruns per variable locus;Pe- checkedfor stability using a iackknife manipulation terson 1990). (Lanyon 1985), in which the data were reanalyzed Tree-buildinganalyses.--I employed a variety of sequentiallyomitting eachingroup taxonand the re- methodsto estimatephylogenetic relationships of the sulting treescombined by strictconsensus using pro- Aphelocomajays. I beganwith simplemethods with gramsdeveloped by ScottM. Lanyonto complement restrictiveassumptions; then, with successivelymore those of PHYLIP (Felsenstein 1989). complex approaches,I attempted to relax these as- Cladisticanalysis of electrophoreticdata is difficult sumptions.I initially analyzedrelationships of all 35 both logicallyand methodologically,chiefly because population samples,but later combined individual of problemswith the treatmentof polymorphicchar- samplesand investigatedrelationships of the groups acters(Buth 1984, Swofford and Berlocher 1987, Swof- of samplesdescribed above. ford and Olsen 1990). Computerpackages such as Although an explicit test of the monophylyof the PAUP deal with the problemby selectingthe most- Aphelocomajays has yet to be conducted,several lines parsimoniouscharacter state from among the possi- of evidenceindicate that the group is monophyletic. bilities and ignoring potentially conflictinginfor- Several characters of plumage, vocalizations, and mation in other allelespresent in the population(D. morphologyare characteristicof Aphelocomajays (Pi- Swofford,. comm.). Because loci in my studywere telka 1951). Hope (1989) found that Aphelocoma,Cy- highly polymorphic(Peterson 1990), coding loci as anocitta,and Gymnorhinusform a well-supportedgroup characterswould not be very informative;instead, I based on phylogeneticanalysis of qualitative and chose to code alleles as characters. mensuralosteological characters. Aphelocoma and Cy- Using the heuristics/subtree-pruningand regraft- anocittashare a synapomorphy(a new bony element ing branch-swappingoption of PAUP (version 3.0; January1992] JayPhylogeny and Evolutionary Rates 137

TAI•I•E2. Matrix of characterstates representing presence and absenceof 38 phylogeneticallyinformative allelesat 29 loci in 35 Aphelocomajay populationsand a compositeoutgroup.

Sample 1 11 21 31 OUTGR 7110010110 0100011000 0000001000 00100000 ACCAL 0000010000 0001110000 0100111000 00000110 ACCAU 0000000000 1000110100 0000010000 00000100 ACCO2 0100000000 0000010000 0010000000 01000000 ACCY1 0100000000 0000010001 0000011001 00011111 ACCY2 0100000000 0010010001 0100011000 10000100 ACGR1 0000110100 1010010001 0100011000 10000110 ACGR3 0000000000 0000010000 0101011000 10000110 ACHY3 0000010000 0000110000 0110011000 00110010 ACIMM 0000000000 0000110000 0000011000 00000010 ACINS 0000000000 0000010000 0000000000 00000000 ACNE1 1000010110 0000010000 0010011110 00000110 ACNE4 0000010100 0000010101 0110011111 00000110 ACNE6 0000000010 0000010000 0010011000 00000110 ACOBc 1000000001 0000110000 1010011000 00000111 ACOBk 0010000000 0000110000 0000010000 00001010 ACOOC 0000100000 1000010001 0000111000 00000110 ACREM 0000000010 0000010000 0000011000 01000000 ACSM2 0000100010 0000010010 0000011000 01101000 ACSP1 1000000000 1000110100 0000011010 00000111 ACSP3 0000000000 0000110000 0000011000 00001110 ACTEX 0000000000 0000010000 0000011000 00000010 ACWO1 1000000000 0000010001 0001001010 00000110 ACWO3 0000000001 0010010001 0010011010 10000110 ACWO4 1000000000 0010010101 0000001110 00000110 AUPO3 0000001000 0000010010 0000011100 00000000 AUGRA 0000001000 0000011000 0000001000 00000000 AUCOL 0000000000 0100001000 0000001000 00000000 AUULT 0000000000 0000001000 0000011000 00100000 AUPO2 0011000000 0000010010 1000011100 00000000 AUWO1 0000001000 0000011010 0000001000 00000000 AUAZ1 0000000000 0000001000 0000000000 00000000 AUCO2 1001000000 0000011010 0000011100 00000000 ANCON 0000000000 0001001000 0000000000 00000000 ANOAX 0000000000 0100001000 0000011000 00000000 ANGUE 0000000000 0000001000 0000000000 00000000

Swofford 1989), I analyzed a binary coding of 102 phasize apparently informative alleles at loci with alleles from 29 loci (Table 2). Of these alleles, 38 were more alleles.Hence, succeedinganalyses were based potentially informative about relationshipsamong on unweighted presence-and-absencedata for alleles. ingroup taxa.This approachcan have the undesirable I useda compositeoutgroup representing allele pres- consequenceof hypotheticalancestors lacking alleles encesand absencesin all outgroup taxa to root re- at particular loci (Buth 1984, Swofford and Berlocher sulting trees.All data setswere testedfor significant 1987).However, this problemcan be ignoredif pop- phylogeneticinformation with the randomizationtests ulation samplesare large enoughthat the presence of Archie (1989). To assessthe stability of nodes on or absenceof one allele in the samplecan be assumed the resulting trees,a jackknifemanipulation (Lanyon to be independent of the presenceof other alleles.In 1985) was conducted,in which the analyseswere re- other words, the alleles under study were indeed peatedsequentially omitting eachof the ingroup taxa. lacking in thoseproblematic hypothetical ancestors, In each repetition, the heuristic search was allowed and someother allele is assumedto have been present to run to completion, with up to 2,500 trees stored at that particular locus. for branch-swapping,and the resulting 35 setsof 13 If the assumptionof independenceof alleles at a to 2,500 equally-parsimonioustrees were combined locus were unreasonable,weighting alleles so that using a majority-rule consensus. eachlocus contributes equally to the analysiswould One problem plaguing cladisticanalyses of elec- be desirable.In this particulardata set, however, pre- trophoretic data is failure to detect alleles actually liminary runs indicated that weighting servesonly present in populations because of inadequate sam- to overemphasize loci with few alleles and deem- pling (Swoffordand Berlocher1987). This nondetec- 138 A. TOWNSENDPETERSON [Auk, Vol. 109

T^SLE3. Matrix of characterstates representing presence and absenceof 33 phylogeneticallyinformative allelesat 29 loci in eight subspeciesgroups of Aphelocomajays and a compositeoutgroup.

Sample 1 11 21 31 OUTGR 7110010110 1001100000 0000000100 000 CO CAL 1011100011 0011010111 1010100111 111 CO WOO 1000101110 0101010101 1111111000 110 CO MEX 0101101001 0101000101 0110011011 CO SUM 0001000100 0001001000 0010000101 000 UL POT 1010010000 0001101010 0011000000 000 ULWOL 0000010000 0001101000 0000000000 000 UL ULT 0000000000 1000100000 0010000100 000 UNICE 0000000000 1010100000 0010000000 000

tion causesextra homoplasyin the resultingtrees in is part of an unresolvedtrichotomy in sometrees, and the form of "reversals,"which areactually undetected the user-treeoption of PHYLIP requires dichotomous synapomorphicalleles. In my study, several popu- branchingat internal nodes.Branch lengths from the lation sampleswere available from most of the sub- replicateanalyses were comparedusing Mann-Whit- speciesgroups, which reducedconsiderably the prob- ney U-tests. ability of nondetectionof alleles.I conducteda cladistic analysisof relationshipsamong regionswithin the three species(i.e. subspeciesgroups; Fig. 1, Table 1) RESULTS by combining all alleles found in populationsof each region into a subspeciesgroup (Table 3). I found 33 Pheneticmethods.--The two phenetic analyses alleles to be phylogenetically informative about re- differed in the placement of a few groups and lationshipsof ingroup taxa (Table 3). Resultsbased in the degree of resolution(Fig. 2). However, on other phylogeny-estimationalgorithms, such as both established one cluster that consisted of the polymorphism parsimony of Felsenstein(1979), populationsof the californicagroup of ScrubJays, yielded similar results,and are not presentedbelow. and another of populationsof the woodhouseii Ratetests.--I tested for evolutionaryrate uniformity and sumichrastigroups of Scrub Jays (Pitelka by focusingon three populationspreviously identi- fied asdistinctive in studiesof geneticdifferentiation 1951).The only exceptionsare the inclusionof (Peterson1990): Scrub Jays from Florida; ScrubJays the populationsof the southerntip of BajaCal- from Santa Cruz Island; and Unicolored Jays from ifornia and northwestern California (ACHY3 Guerrero. The hypothesistested was that historical and ACCAU) as lineagesbasal to both of the independenceof lineagescould accountfor zonesof clustersin the single-linkagemethod, and the extremegenetic differentiation among populations of inclusion of four populationsof Gray-breasted Aphelocomajays. Using three-taxonstatements from Jaysat variouspoints within theseclusters. The the phylogeneticresults, I conductedrelative-rate tests remainingpopulations of Gray-breastedJays and that comparedgenetic distances in all representatives the two eastslope populations of the two taxaconnected by the interior nodeof these (ANCON and ANOAX) form a sister cluster to treesto all representativesof the third taxon.Because at leasttwo of the groupsare representedby numer- the two just described.Scrub Jay populations from Florida and Santa Cruz Island, and the ous population samples,replicate comparisonswere possible. GuerreroUnicolored Jay population fall outside To visualize patterns of nonuniformity of rates, I all of these clusters. also estimated branch lengths and their variability Fitch-Margoliashanalyses.--Analyses of the re- for two subspecies-groupphylogenies with a resam- lationshipsof the 24 populationsamples of Scrub pling technique. The resampling manipulation was Jays(Fig. 3A) showedbasal branches leading to an attemptto estimatehow robustbranch-length es- the Santa Cruz Island and Florida populations timates are to choice of populationsfor sampling. I (ACINS and ACCO2), and then branches to used the user-treeoption of the FITCH programin populations from the interior western United PHYLIP to estimate branch lengths in 30 replicate States, northern Mexico, and southern Mexico analyses.Each replicate included ACINS, ANGUE, ACCO2, and randomly selectedpopulation samples (woodhouseiiand sumichrastigroups). Derived representingeach of the remaining eight subspecies from within the latter group are the populations groups(Table 1). The wollweberigroup of Gray-breast- of the californicagroup. A jackknifemanipula- ed Jayswas excludedfrom some analysesbecause it tion of this data set preservesthe same basic January1992] JayPhylogeny and Evolutionary Rates 139

• ACCAL A CCAL ACOOC ACOOC A COBC ACOBc A CIMM A CIMM ACOBk A COBk A CSP3 A CSP3 ACSP 1 A CSP 1 • ACCY1 A CCAU Single-linkage • ACGR 1 UPGMA ACHY3 ß ACCY2 ACCY1 A CGR3 A CGR I ACNE6 • ACCY2 ACW03 A CGR3 ACW01 A CNE6 ' ACNE4 ACW03 • ACNE 1 ACW01 ß ACTEX -- ACNE1 A CREM ACNE4 -F--- ACSM2 ACTEX • AUP03 • ACREM AUP02 A CSM2 • ACW04 AUP03 A CHY3 AUP02 A CCAU ACW04 AUGR• AUGRA AUC02 AUC02 AUCOL AUAZl AUAZ 1 ANCON ANCON AUUL T -- AUUL T z AUWOAUCOL 1 • AUWO 1 ANOAX ANOAX A CC02 ACCO2 ANGUE AClNS A ClNS ANGUE

0.3 0.2 0.1 0.0 0.3 0.2 0.1 0.0 Modified Rogers' Genetic Distance Fig.2. Pheneticrelationships among 35 populationsof Aphelocornajays based on (A) single-linkagecluster analysisand (B) UPGMA. See Table 1 for abbreviations. structure, with 8 of the 23 nodes lost due to subspeciesgroup in the genus yielded unex- instability of the trees (Fig. 3A). Removal of pectedresults. Inclusion or exclusionof the bas- populationswith very longbranches (Santa Cruz al-split populations ANGUE, ACINS, and Island, ACINS, and Florida, ACCO2), as sug- ACCO2 did not changethe placementof other gestedby Swoffordand Olsen(1990), preserves taxa, so these populationswere excluded from the samebranching pattern in the remaining analysis.In nine replications(all jackknifed) of populations. randomly chosenrepresentatives of each sub- The eight populationsof Gray-breastedJays speciesgroup (for three examples, see Fig. 4), divided into two main lineages,one consisting the southern Gray-breasted Jay population of the two southernpopulations and AUAZ1, (AUCOL or AUULT) clustered with the Uni- and the other of the easternpopulations plus colored Jay population (ANCON or ANOAX), AUGRA and AUWO1 (Fig. 3B). None of the and the eastern Gray-breastedJay population nodeswas unstableunder a jackknifemanip- (AUPO2, AUPO3, or AUCO2) clustered with ulation. Analysesof the three UnicoloredJay the various Scrub Jay populations. The western populationsindicated that the east-slopepop- Gray-breasted Jay population (AUAZ1, AU- ulations (ANOAX and ANCON) are sister taxa, GRA, or AUWO1) joined with the southern well differentiatedfrom the west-slopepopu- Gray-breastedJay-Unicolored Jay group, joined lation (ANGUE, Fig. 3C). with the Scrub Jay-eastern Gray-breasted Jay The analysis of representatives from each group, or separatedin a basal trichotomy. A1- 140 A. TOWNSENDPETERSON [Auk, Vol. 109

0.021 AUPO3 o.04• AUPO2 0.020 0 024 B ACOBk •,•'•.060 -AUGRA

•••o0 ,o5 C •.•ANOAX

•16•O2• CYSTE •ACIN• %8.d.= 0.74 CYSTE Fig. 3. Fitch-Margoliashanalysis of relationshipsamong populations within eachof three speciesof Aphelocomajays: (A) ScrubJays; (B) Gray-breastedJays; and (C) UnicoloredJays. Asterisks indicate nodes unstableto a jackknifemanipulation. Numbers indicate fitted branchlengths; unmarked branches sized proportionally.CYSTE is Cyanocittastelleri, the outgroup;see Table ! for other abbreviations. though the outgroup (CYSTE) is linked to the Thesetrees have severalinteresting features. network in a fairly consistentposition, it could First, they identify several pairs and trios of be arguedthat inaccuraterooting couldproduce populationsas sister taxa: central California and the paraphylyin Gray-breastedJays. In this case, southernBaja California (ACCAL and ACHY3); however, the two-branched form of the phy- northern and central Great Basin populations logeny makesany possiblerooting lead to para- (ACNE1 and ACNE4); northern and southern phyly in that species.From theseresults, I sug- Rocky Mountain populations (ACWO1 and gest that Gray-breased Jays represent a ACWO4); southern Mexican populations of paraphyletictaxon that gave rise to both Scrub ScrubJays (ACREM and ACSM2); and central- Jaysand Unicolored Jays. western Mexican (AUGRA and AUWO1) and Cladisticanalyses.--In cladistic analyses of re- eastern Mexican (AUPO2 and AUCO2 and then lationships of the 35 individual populations, AUPO3) populations of Gray-breastedJays. PAUP found at least 2,500 trees of 108 steps.A Many of these groupingscorrespond to those majority-ruleconcensus of thesetrees is shown describedbased on morphologyand geography in Figure 5. Of the 12 nodes in the concensus- (Pitelka 1951). Within the woodhouseiigroup, tree, eight were retained in all jackknife pseu- northern (ACNE1) and southern (ACNE4) neva- doreplicateanalyses. The randomizationtest of dae populationsare sister taxa, and northern Archie (1989) indicated the existenceof signif- (ACWO1) and southern (ACWO4) woodhouseii icant phylogeneticinformation in the data (P are also sistertaxa. This finding confirmsgeo- -< 0.03). graphicpatterns described by Pitelka(1951), and January1992] JayPhylogeny and Evolutionary Rates 141

AUCOL ACSP1

AUGRA ACCY1 AUPO2 ANCONi :NE4 ACREM • 75-99% WO1 -• 50-74% ACWO4 ACWO3 • 100%oftreesI • ACWO3ACOOC length 108 steps i1• mAccY2 ACOBkACCY2 • c.I.=0.352 •. • ACGR1 • ACGR3

AClMM ...... AUAZ1 • ^COBk i • ACSP3 AUULT ACTEX AUCO2 ACCY1 ANCONi AUPO3

ACNE1 ACSP3 ACSM2• F' ...... co.

ANOAX

AUWO1 • ANOAX AUPO3 Fi•. 5. Cladisti•a•alysis of 35 Aphdoro• jay pop- AUULT• ulations,based o• bi•a? codingof 3g alleles.A •a- jority-•le •o•se•s•s of 2,B00 10•-step trees show• ACSM2 by li•e ?pe. lodes •arked •ith asteriskfo•d to be a•stable to a jackknife manipulation. See •able 1 for ACNE6 abbreviations.Consistency i•dex (c.i.) i•dicated. ACGR3ACCAL • Fig. 4. Three replicateFitch-Margoliash analyses populationsof Gray-breastedJays are united with randomly chosenrepresentative populations of with the ScrubJay clade, making Gray-breasted eight subspeciesgroups of Aphelocomajays. Barsin- Jaysas a taxon paraphyletic(Fig. 5). In addition, dicate speciesmembership: white, Scrub Jay; gray, in all trees examined, a northwestern Gray- Gray-breastedJay; and black,Unicolored Jay. See Ta- breastedJay population sample (AUAZ1) groups ble 1 for abbreviations. with two Unicolored Jay populations(but see Discussion).Again, the exactposition of the root contradictspatterns described by Phillips (1986), doesnot changethe paraphylyin Gray-breasted who suggested that southern Great Basin Jays.Therefore, the currently recognizedspe- (southernnevadae) and southernRocky Moun- ciesGray-breasted Jay appears to be a paraphy- tain (southernwoodhouseii) populations are more letic taxon, with the other two recognizedspe- closelyrelated to eachother than to populations cies derived from different populationswithin to the north. it. In the original consensustree and in 33 of 35 Inspectionof patternsof characterevolution of the jackknife pseudoreplicates,the eastern on this tree indicates that the Scrub Jay clade 142 A. TOWNSENDPETERSON [Auk,Vol. 109

(minus ACTEX, ACREM, and ACSM2) is sup- OUTGR ported by two synapomorphies,and that sev- eral other internal nodesare supportedby syn- !- Apomorphy 2 1 apomorphic alleles. Because the GDA locus shows an odd pattern of variation in Gray- []-Parallelism 10 I 21 •_ COCAL breastedJay populations(Peterson 1990), I be- •- Revereal cameconcerned that it might be responsiblefor length 58 steps the patternsof paraphylyin Gray-breastedJays. c.I. = 0.569 3 This locus might be suspectbecause the two ! cowoo alleles observed differed in mobility by only ! * • COMEX about 2 mm. (Several rechecks of allele homol- 1 ogiesat maximummobility failed to distinguish 2 more than two alleles at this locus.) ! reran the i I COSUM analysesomitting the information at the GDA 21 locus. Although the strict consensusof the re- suiting2,500 most-parsimonious trees is lesswell resolvedthan the overall analysis,AUAZ1 was still more closelyrelated to two Unicolored Jay populations than to other Gray-breastedJay II CUL WOLPOT populations (but see Discussion).A node unit- ing easternGray-breasted Jay populations with ScrubJays was found in 69%of the trees.There- fore, support exists for the hypothesis of para- I phyly of Gray-breastedJays at loci other than UN E GDA. I *LULULT In analysesof relationshipsof the subspecies Fig. 6. Cladisticanalysis of eight subspeciesgroups groups,PAUP found one 58-steptree (Fig. 6), of Aphelocomajays and outgroupshowing character- nine 59-steptrees, and many longer trees. The state changesand support for nodes.Nodes marked shortest tree is almost identical to those pro- with asteriskfound to be unstableto a jackknifema- ducedby the Fitch-Margoliashanalyses (Fig. 4), nipulation.See Table 1 for abbreviations;consistency index (c.i.) indicated. but differs in that eastern and western Gray- breastedJays group together;it providesbetter resolutionamong populations of ScrubJays. The variably placed as basal offshootsof the trees 59-step trees differ in placement of a few (Fig. 7B). For reasonsoutlined below, this ar- branches, but the basic structure is similar. Sev- rangement seems unlikely, so a placement of eral alleles provide support for most nodes,in- thesethree populationsthat agreesmore closely cluding 10 alleles that represent synapomor- with geography(Fig. 7A) is also consideredin phies for all Scrub Jay populations except the the analysesof rate uniformity. sumichrastigroup, and one to five synapomor- Ratesof molecularevolution. --Because the phy- phies that support other nodes (Fig. 6). A jack- logeneticstudies described above are equivocal knife manipulation of the data set preserved asto the placementof three populations(ACINS, the general structureof the tree, with two un- ANGUE, and ACCO2), ! usedboth treesof Fig- stable nodes. The randomization test of Archie ure 7 in the rate tests. ! conducted a relative- (1989) indicated the presenceof significantphy- rate test based on the following three-taxon logeneticinformation in the data set (P < 0.01). statements(from Figs.7A and 7B):((ACINS, cal- Omitting the two GDA allelesfrom analysisdid ifornica group), woodhouseiigroup); ((ANGUE, not change the tree structure,except in that two east-slope unicolor), southern Gray-breasted nodes are unresolved. Jays);((ACCO2, rest of Scrub Jays),eastern ul- Although the tree of the eight subspecies tramarina);((ACINS, rest of ingroup), CYSTE); groups in the three speciesis well resolved, ((ANGUE, rest of ingroup), CYSTE); and placementof the remaining three distinct lin- ((ACCO2, rest of ingroup), CYSTE). In the six eagesis equivocal.The ScrubJays of SantaCruz tests,apparent departures from rate equalityex- Island (ACINS) and Florida (ACCOE), and the ist; ACINS, ANGUE, and ACCO2 in every com- Unicolored Jaysof Guerrero (ANGUE) are in- parison had greater genetic distancesand pa- January1992] JayPhylogeny and Evolutionary Rates 143

g g g o o o o g g g øøø

C.I. = 1.0 C.I. = 0.5 PLUMAGETYPE ß UnicoloredJay [] Gray-br.Jay [] ScrubJay Fig. 7. Two possiblephylogenies combining the resultsof Figures2-6, showingmost-parsimonious re- constructionsof patternsof plumageevolution. Abbreviations follow Table 1; consistency index (c.i.) indicated. tristic distances to the third taxon than did their an attempt to understand what factors, meth- sisterlineages (sign test; all P < 0.001).Presum- odologicalor biological,led to thosepatterns. ably, higher ratesof molecularevolution have Paraphylyof theGray-breasted Jay.--In all anal- characterizedthese three lineagesrelative to yses,the Gray-breastedJay is a paraphyletictax- the rest of the group. on. UnicoloredJays invariably group with the Analysesinvolving resamplingof popula- southernGray-breasted Jay lineage, and Scrub tionsrevealed a high degreeof heterogeneity Jayswith the easternGray-breasted Jay lineage. of ratesamong lineages (Fig. 8). Statisticalcom- The positionof the westernGray-breasted Jay parisonsindicate that evolutionary rates are sig- lineage is less clear. On the whole, it appears nificantly different in every pair of sisterlin- that the Gray-breastedJay is a relatively old eages(Mann-Whitney U-test; all P < 0.05).The lineagethat hasbeen in the mountainsof north- validity of thesestatistical tests depends on the ern Mexico for a long period of time, and that effectivenessof the resamplingmanipulation in Scrub Jaysand Unicolored Jayswere derived estimatingtrue levels of branch-lengthvari- independentlyfrom different partsof the Gray- ability. The manipulationemployed here eval- breastedJay complex. uateserror introducedby choiceof samplelo- Although unexpected(Pitelka 1951),this idea cality, but not samplingerror in estimationof receivessome support from other evidence.First, genetic distances.Nevertheless, it is clear that Gray-breastedJays are intermediate in habitat molecularevolutionary rates have been uneven use between the other two species.Where the in the historyof the Aphelocomajays. three species'ranges overlap in central Vera- cruz, ScrubJays live on dry interior slopesand DISCUSSION UnicoloredJays on humid coastalslopes. Gray- breasted Jays inhabit the - forests at The consistencywith which geographically higherelevations intermediate between the two, relatedpopulations clustered together indicates and so direct invasion of either habitat would that thesedata containphylogenetic informa- have been possible.Second, Gray-breasted Jays tion. For example,in mostanalyses, the 10wood- in the southern and western portions of the houseii-grouppopulations cluster together, as do range show delayed maturation of beak color, the nine californica-grouppopulations. Below I a characterpresent in rudimentaryform in Uni- discussthe unexpectedfeatures of the trees in colored Jays and otherwise rare in (Pe- 144 ^. TowNsœNt)PœTœRSON [Auk,Vol. 109 terson199 lb). EasternGray-breasted Jays do not delay beak color maturation, which is similar to the ontogenetic patterns of Scrub Jays(Pe- terson,in press).Similarly, western and south- ern Gray-breastedJays do not give the female- specific"rattle" vocalization, which is found in easternGray-breasted Jay and all ScrubJay pop- ulations (Brown and Horvath 1989). Therefore, the hypothesisof independent derivations of Scrub and Unicolored jays from Gray-breasted Jaysis supportedby other characters.Taxonom- ic implicationsof thesefindings will be treated elsewhere (Peterson, in prep.). Originof thecalifornica group of ScrubJays.- In a numberof Fitch-Margoliashtrees (e.g. Fig. 3), the californicagroup of ScrubJays is derived from within the woodhouseiigroup of ScrubJays. Somesupport for this hypothesisis alsofound in cladisticresults although these nodes are not preserved in the consensusanalyses (Fig. 5). This pattern is to be expectedbased on a prob- able Central American/Mexican origin of the specieswith subsequentinvasion into the Great Basin and then into California (Peterson 1990). Rapid evolutionversus old splits.--Three dis- junct populations--the ScrubJays of SantaCruz Island (ACINS), the Scrub Jays of Florida (ACCO2), and the Unicolored Jaysof Guerrero Fig. 8. Trees showing resultsof resamplinganal- (ANGUE)--consistently are placedas basallin- ysesof ratesof molecular evolution in different lin- eagesin the phylogeneticanalyses (Fig. 7; in a eages,based on two treesof Figure 7. See Table 1 for few alternative trees, the Florida population abbreviations. ACCO2 groupedwith southernMexican Scrub Jay populationsbased on one allele unique to the two groups).This resultis surprising,given ternative B. First, although placement of these that eachshares obvious similarities of plumage three populationsas basal splits (Fig. 7B) re- coloration, shape and size, vocalizations,and duceshypothesized levels of heterogeneityof behavior with other membersof its species.For rates of molecular evolution, it leads to prob- example,ACINS sharesseveral unique plumage lems in understandingthe evolution of plum- characterswith populationsof California and, agecoloration. For the tree in Figure 7B, plum- on that basisas well ason groundsof geograph- age colorationmust have undergonea number ic proximity, seemsto have been derived from of reversalsand episodesof rapid evolution adjacentmainland populations (Pitelka 1951). (plumagecoloration mapped onto tree; consis- Two alternative hypothesesare apparent: (A) tency index = 0.5). Under the view in Figure that these lineagesare exceptionallydivergent 7A, plumagecoloration shows only two changes: populationsderived from within ScrubJays and one at the derivation of ScrubJays; and another Unicolored Jays,but for nonhistorical reasons at the derivationof UnicoloredJays (consisten- appear as basalsplits on the trees (Fig. 7A); or cy index = 1.0). Hence, hypothesis B is less (B)that theselineages are exceptionallyold, and consistentwith patterns of variation in plum- the charactersthat each shareswith its conspe- age-color charactersthan hypothesis A. Fur- cific populationsare actually retained similar- thermore, these populationsshare not only as- ities (Fig. 7B).Alternative B is better supported pectsof plumagecoloration, but a wide range by the allozyme data than alternative A (tree- of characters(e.g. body size and shape,ecolog- length of 62 stepsin B; 73 stepsin A). icalrequirements, vocalizations, behaviors) with Still, two biologicalpoints argue againstal- conspecifics(Pitelka 1951), so whole suites of January1992] JayPhylogeny andEvolutionary Rates 145 charactersnot included in my analyseswould of alleles. Thus, the tempo and mode of evo- have had to have changedunpredictably. lution in a groupcan affect the outcomeof phy- Second,even if these populationsrepresent logeneticanalyses; in this case,it leadsto the basal splits (alternative B), they still have ex- basal placement of ACINS, ANGUE, and ceptional branch lengths. Unequal branch ACCO2. lengths can result from incorrect choice of out- This problem may affect other parts of the group (i.e. an outgrouptaxon actuallybelongs phylogeneticreconstruction of the ApheIocoma to the ingroup).Outgroups in thisstudy include jaysas well. For example,it couldbe responsible representativesof the entire family Corvidae, for the basalplacement of ACTEXwithin Scrub soderivation of outgrouptaxa from the ingroup Jaysand the groupingof ANCON, ANGUE,and is unlikely. Nevertheless,if to avoid unequal AUAZ1, all of which show reduced genetic branch lengths, the trees are rerooted at the variation (Fig. 5). Grouping of Florida and Santa midpoint of the longest branch (in this case, Cruz Island populations(ACCO2 and ACINS) that leading to ACINS or ANGUE), not only do in someanalyses (a mostunlikely outcomegeo- at leasttwo other ingroup taxaremain with ex- graphically)is almostcertainly a result of this ceptionalbranch lengths,but the outgroupap- phenomenon.Unexpected features of the phy- pearsas an exceptionallydivergent branchde- logeniesdiscussed above, such as the derivation rived from within ApheIocoma. of the caIifornicagroup from the woodhouseii Removalof the three long branchesfrom the groupand paraphylyof Gray-breastedJays, most tree (after Swofford and Olsen 1990) produces likely do not result from this problem because results that agree with geography and mor- they do not involve groupsthat differ markedly phology. The populationson the three long in levels of geneticvariability. branchesall represent small, isolated popula- The besthypothesis for the positionof ACINS, tionsisolated from the remainderof the species' ACCO2, and ANGUE in the Aphelocomajay phy- ranges. Hence, ! considered features of small logeny appearsto be that they are exceptionally populationsthat could possiblyaffect accurate derived lineages from within Scrub Jaysand reconstructionof phylogenies. Unicolored Jays.This conclusionis supported A possibleresolution of thesedifficulties may by severallogical problems with the alternative lie in the extremely reduced levels of within- hypothesis.The methodologicalreason for their populationgenetic variation in thesesame three basalplacement in the analysesis clearly iden- populations (Peterson 1990). Heterozygosities tifiable, and has nothing to do with time since for the three populationsaveraged 1.9, as com- divergencefrom the ancestralstock. pared with 3.9 for the remainder of the popu- My resultsserve to illustrate someof the dif- lationssampled (Mann-Whitney U-test,P < 0.05; ficultiesof using electrophoreticcharacters in Peterson1990). Other measuresof within-pop- phylogeneticanalyses, as have beenpointed out ulation variation (e.g. proportion of loci poly- by other workers(e.g. Buth 1984,Swofford and morphic, number of alleles) are similarly re- Berlocher 1987, Swofford and Olsen 1990). The duced in the three populations. Many of the highly polymorphicnature of thesecharacters characterson which the phylogeniesare based makescoding and analysisdifficult; effectsof are alleles at low or moderate frequencies (e.g. population history can reduce the probability PGD-e, which is found in all but one of caIifor- of correctly reconstructingthe history of the nica-group populations, never at a frequency group. Although ! believe that useful and ac- >0.136). If a population with low-frequency al- curate information was retrieved from the data leles that identify its sister-taxonrelationships set, the difficulty of exactplacement of several losesvariation, perhapsdue to population bot- important populationsdemands further study tlenecks, those same low-frequency alleles with other types of biochemicalcharacters. would most likely be lost. Consequently,many Ratesof molecularevoIution.--The analyses of of the charactersshared with the ingroup (syn- rates of molecular evolution indicated that at apomorphies) are secondarily absent. These leastthree ApheIocomajay populationshave un- changeswill move such lineagesto the baseof dergonemarked accelerationsin rate of differ- cladograms.Phenetic techniques that assume entiation relative to their sisterpopulations. Re- rate equality obviously will be sensitive to in- gardlessof whether they are placed as basal creasesin evolutionary rates in particular lin- splits or with apparently conspecificpopula- eagesand, potentially, also to secondarylosses tions (Figs. 7B or 7A), these populationshave 146 A. TOWNSENDPETERSON [Auk,Vol. 109 evolvedfaster than other populationsof Aphelo- to John Bates,Shannon Hackett, Scott Lanyon, and comajays. Inspection of branch lengths in the two anonymousreviewers for helpful commentson tree (Fig. 8) suggeststhat heterogeneityof evo- the manuscript. I thank Pamela Austin, John Hall, lutionaryrates may not be restrictedto justthose Victor Pacheco,Mary Anne Rogers,Douglas Stotz, ThomasSchulenberg, David Willard, Adolfo Navarro three lineages,but that rate heterogeneityon a S., and Patricia EscalanteP. for their interest in my finer scalemay be common in the genus. work. Shannon Hackett provided invaluable assis- Becauseall tree-buildingmethods are to some tance with testing phylogenetic information in the degree sensitive to rate differences(Swofford cladisticdata sets.I thank the Direcci6n de Flora y and Olsen 1990), trees estimated are not inde- Fauna Silvestre, Secretaria de Desarollo Urbano y pendent of the question of whether rates of Ecologla,of the Mexicangovernment, and the state evolution are uniform. When presentedwith a governmentsof Oregon,California, Nevada, , caseof extremerate inequality, algorithmsfor Utah, ,Colorado, and Texasfor providing phylogeny estimationerr on the side of reduc- permits for scientific collecting. The Field Museum ing heterogeneityof branch lengths.A search generouslyprovided accessto its facilities for bio- chemicalanalysis. Financial support was provided by for rate differencesbased on trees potentially the National Science Foundation Dissertation Im- affectedby such differencesis a conservative provement Grant Program (BSR-8700850),National test;errors introducedwill be of not finding a GeographicSociety, Field Museum of Natural His- difference that exists, rather than of finding tory, Chapman Fund (American Museum of Natural spuriousheterogeneity. Studies of variation in History), SigmaXi, and the University of Chicago. other moleculesthat will permit critical inde- pendenceof phylogenyestimation and branch- length estimationare in progress. LITERATURE CITED More interesting is the degree to which evo- lutionary rates differ. If ACINS and ANGUE ARCHIE,J. W. 1989. A randomization test for phy- representbasal splits from the genus,they have logenetic information in systematicdata. Syst. Zool. 38:239-252. been evolving 20-25% more rapidly than the rest of the genus.If, as seemsmore likely, they ARCHIE,J. W., C. SIMON,AND A. MARTIN. 1989. Small sample size does decreasethe stability of den- were derived from within ScrubJays and Uni- drograms calculated from allozyme-frequency coloredJays, their evolutionaryrates have been data. Evolution 43:678-683. three to four times greater than those of their BLEDSOE,A. H. 1987. DNA evolutionaryrates in nine- sisterlineages. Correlations between rate of dif- primatied passefinebirds. Mol. Biol. Evol. 4:559- ferentiationand featuresof populationbiology 571. are explored in Peterson(1990, in prep.). BRITTEN,R.J. 1986. Rates of DNA sequenceevolu- Electrophoreticdata for populationsof Aphe- tion differ between taxonomic groups. Science locomajays clearly violate the assumptionof rate 231:1393-1398. equality. Although studiesof severalother avi- BROOKS,D. R., ANDD. A. MCLENNAN.1990. Phylog- an groupshave cometo the oppositeconclusion eny, ecology,and behavior.Univ. ChicagoPress, Chicago. (e.g. Bledsoe1987), my resultsand the resultsof BROWN,J.L. 1987. Helping and communalbreeding others (e.g. Britten 1986, Springer and Kirsch in birds. Princeton Univ. Press, Princeton, New 1989) indicatethat the assumptionof rate uni- Jersey. formity is not always reasonable.Rate-depen- BROWN,J. L., ANDE.G. HORVATH.1989. Geographic dent algorithms and interpretations should be variationof groupsize, ontogeny, rattle calls,and testedfirst (e.g. Page 1990)and used cautiously, body size in Aphelocomaultramarina. Auk 106:124- if at all, in studiespurporting to derive phy- 128. logenetic or biogeographicinformation from BUTH,D.G. 1984. The applicationof electrophoretic molecular data alone. data in systematicstudies. Annu. Rev. Ecol.Syst. 15:501-522. CURTIS, E. L., AND R. C. MILLER. 1938. The sclerotic ACKNOWLEDGMENTS ring in North American birds. Auk 55:225-243. FELSENSTEIN,J. 1979. Alternative methods of phy- Many people have lent invaluable assistanceto me logenetic inference and their interrelationship. during this project. I thank especiallyScott M. Lan- Syst. Zool. 28:49-62. yon, Frank A. Pitelka, and John W. Fitzpatrick for FELSENSTEIN,J. 1982. Numerical methods for infer- their generousinterest throughout the project.Thanks ring evolutionarytrees. Q. Rev. Biol. 57:379-404. January1992] JayPhylogeny and Evolutionary Rates 147

FELSENSTEIN,J. 1989. PHYLIP (phylogenyinference $ARICH,V. M. 1977. Rates, sample sizes, and the package),version 3.2 manual.Univ. Washington, neutrality hypothesisfor electrophoresisin evo- Seattle. lutionary studies. Nature 265:24-28. FITCH, W., AND E. MARGOLIASH. 1967. Construction SHAW,C. R., AND R. PRASAD.1970. Starch gel elec- of phylogenetictrees. Science155:279-284. trophoresis--A compilationof recipes.Blochem. FITZPATRICK,J. W., AND G. E. WOOLFENDEN.1986. De- Genet. 4:297-330. mographicroutes to cooperativebreeding in some SHELDON,F.H. 1987. Ratesof single-copyDNA evo- jays. Pages137-160 in Evolution of lution in herons. Mol. Biol. Evol. 4:56-69. behavior (M. H. Nitecki and J. A. Kitcheil, $IBLEY,C. G., ANDJ. E. AHLQUIST.1981. The phylog- Eds.). Oxford Univ. Press, Oxford. eny and relationshipsof the ratite birds as in- HARDY,J.W. 1969. A taxonomic revision of the New dicatedby DNA-DNA hybridization.Pages 301- World jays. Condor 71:360-375. 335 in Evolution today: Proceedingsof the Sec- HARRIS,H., AND D. g. HOPKINSON. 1978. Handbook ond International Congressof Systematicand of enzyme electrophoresis in human genetics. Evolutionary Biology (G. G. E. Scudder and J. L. North-Holland Publishing Co., Amsterdam. Reveal, Eds.). Carnegie-Mellon Univ., Pitts- HOPE,$. 1989. Phylogeny of the avian family Corvi- burgh. dae. Ph.D. dissertation,City Univ. New York, $IBLEY,C. G., AND J. E. ALQUIST. 1983. Phylogeny New York. and classification of birds based on the data of JOHNSON,N. K., AND J. A. MARTEN. 1988. Evolution- DNA-DNA hybridization. Curr. Ornithol. 1:245- ary geneticsof flycatchers.II. Differentiation in 292. the Empidonaxdifficilis complex. Auk 105:177-191. $IBLEY,C. G., J. E. ALQUIST,AND F. H. SHELDON. 1987. LANYON,S.M. 1985. Detecting internal inconsisten- DNA hybridization and avian phylogenetics:Re- cies in distancedata. Syst.Zool. 34:397-403. ply to Cracraft.Evol. Biol. 21:97-125. NEVO, E., R. BEN=$HLOMO,g. BEILES,J. U. g. JARVIS, SPRINGER,M. S., AND J. g. W. KIRSCH. 1989. Rates of AND G. C. HICKMAN. 1987. Allozyme differen- single-copy DNA evolution in phalangeriform tiation and systematicsof the endemic subter- marsupials.Mol. Biol. Evol. 6:331-341. ranean mole rats of South . Blochem.Syst. SWOEEORD,D.L. 1989. PAUP: Phylogeneticanalysis Ecol. 15:489-502. using parsimony, version 3.0. Illinois Natural PAGE,R. D.M. 1990. Temporal congruenceand cla- History Survey, Champaign. disticanalysis of biogeographyand cospeciation. $WOFFORD,D. L., AND $. H. BERLOCHER. 1987. Infer- Syst. Zool. 39:205-226. ring evolutionarytrees from genefrequency data PETERSON,A.T. 1990. Evolutionary relationshipsof underthe principleof maximumparsimony. Syst. the Aphelocomajays. Ph.D. dissertation, Univ. Zool. 36:293-325. Chicago,Chicago. $WOFFORD,D. W., ANDG. J. OLSEN.1990. Phylogeny PETERSON,h. T. 1991a. New distributional infor- reconstruction.Pages 411-501 in Molecular sys- mation on the Aphelocomajays. Bull. Br. Ornithol. tematics (D. M. Hillis and C. Moritz, Eds.). Si- Club 111:28-33. nauer Associates, Sunderland, Massachusetts. PETERSON,A. T. 1991b. Sociality and ontogeny of WOOLFENDEN,G. E., AND J. W. FITZPATRICK.1984. The colorationin the blue-and-blackjays. Wilson Bull. Florida ScrubJay: Demography of a cooperative 103:59-67. breeding . Princeton Univ. Press,Princeton, PETERSON,h. T. 1992. Geographic variation in the New Jersey. ontogenyof beakcoloration of Gray-breastedJays WILSON, A. C., S. CARLSON,AND T. J. WHITE. 1977. (Aphelocomaultramarina). Condor 93:448-452. Biochemical evolution. Annu. Rev. Blochem. 46: PETERSON,A. T., ANDN. VARGASB. 1992. Ecological 573-639. diversity in ScrubJays (Aphelocoma coerulescens). WRIGHT, S. 1978. Evolutionand the geneticsof pop- In Biologicaldiversity of Mexico: Origins and ulations. Vol. 4, Variability within and among distribution(T. P. Ramamoorthy,R. Bye, A. Lot, populations.Univ. ChicagoPress, Chicago. and J. Fa, Eds.). Oxford Univ. Press, Oxford. ZINK,R.M. 1982. Patternsof genicand morphologic PHILLIPS, A.R. 1986. The known birds of North and variation among sparrows in the genera Zono- Middle America,vol. 1. A. R. Phillips, Denver. trichia, Melospiza,Junco, and Passerella.Auk 99: PITELKA,F.A. 1951. Speciationand ecologicdistri- 632-649. bution in Americanjays of the genusAphelocoma. ZUCKERKANDL,E., AND L. PAULING. 1962. Molecular Univ. Calif. Publ. Zool. 50:195-464. disease,evolution, and genicheterogeneity. Pages ROHLF,F. J. 1988. NTSYS-pc:Numerical 189-225 in Horizons in biochemistry(M. Kasha and multivariate analysissystem, version 1.40. and B. Pullman, Eds.).Academic Press,New York. Exeter Software, Setauket, New York.