sign in sign up
<<
Home , Aphelocoma, Corvidae, Cyanocitta, Cyanocorax, Perisoreus

The Condor 99:490-502 0 The Cooper Ornithological Society 1997

INTERGENERIC RELATIONSHIPS OF THE JAYS INFERRED FROM CYTOCHROME b GENE SEQUENCES

ALEJANDROESPINOSA DE LOS MONTEROS AND JOEL CRACRAFT Department of Ornithology, American Museum of Natural History, Central Park West @ 79th Street, New York, NY 10024, e-mail: [email protected]

Abstract. The six genera of corvids endemic to the (i.e., , Calo- citta, , , , and Gymnorhinus) form the assemblagethat most ornithologistscall the New World jays. The intergenericrelationships among these six gen- era are explored using complete sequencesfrom the cytochrome b gene (1,143 bp) along with 29 morphologicalcharacters. A consistentphylogenetic hypothesis was obtained when the data sets were analyzed independently or in a total evidence approach.The phylogeny presentedin this paper does not corroboratethe existence of two evolutionary lineages as previously proposedby Hardy (1961, 1969). The most basal of the New World jays is Cyanolyca, which also is supportedby independent evidence on cranial anatomy. The remainder of the genera are embeddedwithin two major clades. The most derived is formed by Cyanocitta, Aphelocoma, and Gymnorhinus, being Cyanocitta the most basal genuswith- in this clade. Calocitta and Cyanocorax constitutethe otherclade. A very closerelationship betweenthese two generais supportedby bootstrapvalues and branchsupport analysis. Theseresults corroborate the hypothesisthat the Piiion (Gymnorhinus cyanocephalus) is a true New World jay, and a putative relationship with nutcrackersis not supported.The phylogeny obtainedis usedto infer a biogeographicscenario, as well to explain the evolution of a very derived jaw articulationpresent only in the New World jays. The biogeographic scenario agrees with a relatively recent arrival of a corvine ancestor via Beringia, and a very rapid dispersaland radiation into the Americas. Key words: New World jays, , cytochrome b gene, total evidence analysis, phylogenetic analysis, biogeography.

INTRODUCTION ids, particularly to the genus Nucifraga, based Corvids (, jays and allies) are one of the on the fact that Gymnorhinus walks and runs most successful groups of . The instead of hops as do all the New World jays. family has a world-wide distribution and exhib- Hardy argued that the walking behavior must be its a high speciesdiversity. The corvids encom- correlated with a rather different neuromuscular pass a total of 25 genera and 113 ,which control system. Other workers have acceptedthe are found in numerous kinds of (Good- idea that the Pifion Jay is related to the other win 1976, Madge and Bum 1994). Six genera jays, but they have usually postulated that Gym- norhinus of corvids (i.e., Aphelocoma, Calocitta, Cyano- is basal and should be considered as the sister-group to the New World jays (Ashley citta, Cyanocorax, Cyanolyca, and Gymnorhin- us, sensu AOU 1983) are endemic to the Amer- 1941, Pitelka 1951, Goodwin 1976). icas and have successfully radiated within trop- Not only has the position of the Pifion Jay ical, subtropical, and mild temperate environ- been challenged, but the entire phylogenetic ar- ments. Some morphological characters, like a rangement of the New World jays has been con- buttress complex (Zusi 1987), shared by these stantly questioned. Earlier workers have at- taxa seem to support their , and for tempted to resolve corvid intergeneric relation- most ornithologists these six genera comprise ships based on morphological characters (e.g., the so-called New World jays. In contrast, others Ridgway 1904, Amadon 1944, Hardy 1961, have argued that Gymnorhinus (the Pifion Jay) 1969). Despite this attention, the phylogenetic should be excluded from this assemblage. For conclusions of these studies show little congru- instance, Hardy (1961) postulatedthat the Pition ence. Moreover, the methods used to establish Jay was most closely related to Old World corv- monophyly of the New World jays in those anal- yses were weak, because they were based on poorly defined similarities and without evidence ’ ’ Received10 May 1996. Accepted12 December for polarity of the characters. 1996. Using external morphological and behavioral

[4901 INTERGENERIC RELATIONSHIPS OF THE NEW WORLD JAYS 491 characters, Amadon (1944) postulated that all within the group showed highly unstable topol- New World jays formed a monophyletic group, ogies, her results supported the monophyly of and he placed them as the sister-group to Gym- the New World jays on the basis of four derived norhinus cyanocephalus(Pifion Jay). He consid- characters: (a) a very high rostra1 crest of the ered Gymnorhinus to be a specialized jay that lateral cotyla of the mandible, (b) a rostral artic- resembles nutcrackers (Nucifraga spp.) because ulation of the lower jaw, (c) the presence of a of similarities in feeding habits. Amadon’s clas- medial basal ridge on the quadrate, and (d) an sification divided the New World jays into four extensive fusion of the prepalatine bar to the genera, instead of the nine previously proposed ventral bar of the jaw. Hope (1989) presented by Ridgway (1904). The genera recognized by three different phylogenetic hypotheses for re- Amadon were: Cyanocitta (which included lationships within the New World jays. In all of Aphelocoma, Cyanolyca and Cissilopha), Psilo- them, however, the genera Aphelocoma and rhinus, Cyanocorax, and Calocitta (Fig. la). Cyanocorux were postulated to be paraphyletic. Hardy (1961) suggested that two main evo- Another constant result was the emergence of lutionary lineages of New World jays could be Cyanolyca as the most basal taxon. A strict con- identified using both behavioral and morpholog- sensustree of the three solutions highlights the ical evidence. He distinguished an “Inornate unresolved relationships among the New World line” (tribe Aphelocomini) containing only the members of the genus Aphelocomu, and an “Or- jays (Fig. Id). nate line” (tribe Cyanocorini) which encom- The development of molecular protocols has passed five of the remaining seven genera (Cul- allowed the integration of biochemical data into ocitta, Cissilopha, Cyanocitta, Cyanocorax, and phylogenetic analysis. Sibley and Ahlquist Psilorhinus). The genus Cyanolyca was omitted (1985) employed DNA-DNA hybridization dis- from either line becausethe speciesin this genus tances to explore the phylogenetic affinities were considered to be so poorly known behav- among corvids. Their analysis included three iorally that their affinities could not be assessed genera of New World jays (Cyanocitta, Aphel- (Fig. lb). Hardy (1969) later reviewed his pre- ocoma, and Cyanocorux). Their results show the vious classification and maintained the hypoth- New World jays as a paraphyletic group inside esis of two evolutionary lineages inside the New the tribe Corvini (sensu Sibley and Ahlquist), World jays, but this time only recognized three which also includes crows, , nutcrackers genera. The ornate lineage consisted of Cyano- and coughs. It should be noted, however, that cittu, as well as the largely Neotropical Cyano- the resolving power of DNA hybridization data, corax within which Calocitta, Cissilopha, and as well as the methodology used to analyze Psilorhinus were ranked as subgenera. The in- them, have been questioned (Cracraft 1987, ornate lineage contained the third genus, Aphel- Houde 1987, Harshman 1994). ocoma, which also included the speciesformerly Several other molecular studies have ad- placed in Cyanolyca. A major problem with both dressedthe phylogenetic relationships within the of Hardy’s classification schemesis that the in- corvine assemblage (Sibley and Ahlquist 1990, ornate clade was defined mainly on the absence Helm-Bychowski and Cracraft 1993). However, of ornate which seems to be a plesio- none of these studies has confronted relation- morphic condition among corvids. ships within the New World jays in detail. The Goodwin (1976) concurred with previous au- goal of the present study is to clarify the generic thors’ opinions that there was an early isolation interrelationships of these taxa by generating a of the New World jays within the corvid radia- tion. Although his classification followed Ama- new data set, namely sequencesfrom the mito- don’s hypothesis and placed the New World jays chondrial cytochrome b gene. as the sister-group of Gymnorhinus, relation- To avoid further confusion, we will use the ships within the New World jays were complete- noun “jays” when referring to the seven genera ly unresolved (Fig. lc). of corvids used in this study, and the noun Using a cladistic approach, Hope (1989) an- “New World jays” for the six genera endemic alyzed the phylogenetic interrelationships within to the Americas (Aphelocoma, Calocitta, Cya- the family Corvidae employing a set of 80 os- nocitta, Cyanocorax, Cyanolyca, and Gymno- teological characters. Although the analyses rhinus). 492 ALEJANDRO ESPINOSA DE LOS MONTEROS AND JOEL CRACRAFT

(a) Amadon, 1944 (b) Hardy, 1961

Cyanocorax Cissilopha

Calocitta Psilorhinus

L Cyanocorax

(c)Goodwin, 1976 (d) Hope, 1989

Other Corvids Other Corvids

Perisoreus

Corvus

Cyanolyca

Gymnorhinus

Aphelocoma

Cyanocitta

Psilorhinus

Cyanocorax

Calocitta

FIGURE 1. Some phylogenetic hypotheses previously proposed for the intergeneric relationships of New World jays.

METHODS craft (1993). The (Cyanocittu cristutu) TAXA EXAMINED complemented the sequencesof the New World jays presented here. Four species of birds-of- Complete cytochrome b gene sequences(1,143 paradise (Munucodiu keruudrenii, Epimuchus bp) were determined for six speciesof jays (Ta- fustuosus, Ptiloris paradiseus, and Cicinnurus ble l), each representing a different genus within mugnificus)and two bowerbirds (Ailuroedusme- the lineage. Four sequencesbelong to what con- lunotus and Ptilonorhynchus violuceus) were ventionally has been recognized as the core New used as representatives of closely related lin- World jays: the Scrub Jay (Aphelocoma coeru- eages within the corvine assemblage. Finally, lescens), Jay (Culocittuforrnosa), Plush- the Hermit Thrush (Cuthurus guttutus), a non- capped Jay (Cyunocorax chrysops), and White- collared Jay (Cyanolycu viridicyana). The fifth corvine passerine, was employed as an outgroup sequence corresponds to the Piiion Jay (Gym- to root the tree. norhinus cyunocephalus) which for some orni- DNA EXTRACTION AND SEQUENCING thologists should be considered as a close rela- tive to New World jays, but not as part of this Frozen tissue (muscle, liver and heart) of se- assemblage. Finally, the last sequence belongs lected taxa was obtained from the Museum of to the Gray Jay ( cunadensis). Natural Science, Louisiana State University, and The rest of the sequencesused for this anal- the Museo de Zoologia Facultad de Ciencias, ysis were taken from Helm-Bychowski and Cra- Universidad National Autonoma de . To- INTERGENERIC RELATIONSHIPS OF THE NEW WORLD JAYS 493

TABLE 1. Jay specimens sequenced.

Locality Collection Reference number Aphelocomacoerulescens USA; LSUMZ’ B-15539 Calocittaformosa MEXICO; Oaxaca MZUNAMb OMVP-168 Cyanocoraxchrysops BOLIVIA; SantaCruz LSUMZ” B-18785 Cyanolycaviridicyana PERU; Huanuco LSUMZ” B-3501 Gymnorhinuscyanocephala USA; California LSUMZ” B-14328 Perisoreuscanadensis USA: LSUMZ” B-10198

*Collection of Frozen Tissues, Louisiana State University Museum of Natural Saence. h Museo de Zoologia Facultad de Ciencias. Umversidad National Autonama de Mexico tal genomic DNA was extracted using Chelex were considered as unordered, and their polar- 5% following the protocol suggestedby Singer- ization was determined by the use of outgroup Sam et al. (1989). The cytochrome b gene comparison. The Gray Jay was employed as the (1,143 bp) was amplified by the use of conven- outgroup for this part of the analysis. A char- tional PCR techniques (Kocher et al. 1989). A acter-statematrix for all these data is presented total of 14 oligonucleotide primers were used for in Table 2. A detailed list of the characters as amplification and sequencing (primer sequences well as a description of the character-statescan documented in Helm-Bychowski and Cracraft be consulted in Appendix 1. 1993). Single-stranded template for sequencing was produced by asymmetric PCR using unbal- DATA ANALYSIS anced primer ratios (Gyllensten and Erlich Independent cladistic analyses of each data set 1988). DNA sequencing was performed using (i.e., molecules and morphology), as well as a T7 DNA polymerase (Sequenase version 2.0, data set that combined both morphological and United States Biochemical) following standard molecular characters, were performed. Phylo- dideoxy chain-termination protocols (Sanger et genetic analyses were undertaken with the com- al. 1977). Finally, sequencing reactions were puter program Phylogenetic Analysis Using Par- subjected to denaturing gel electrophoresisand simony (PAUP version 3.1.1; Swofford 1993). autoradiography. Sequenceswere easily aligned To insure the discovery of the most parsimoni- by eye with the aid of the ESEE editor program, ous solution, exhaustive searcheswere executed. version 1.07 (Cabot and Beckenbach 1989). Nucleotide transformations were considered unordered during all analyses. MORPHOLOGICAL DATA Many methodologies have been suggestedto Hope (1989) undertook an analysis of corvid evaluate cladistic signal in sequencedata (Swof- phylogeny using osteological characters. To ford and Olsen 1990, Cracraft and Helm-By- complement the present study, we employed a chowski 1991, Hillis 1991). We adopted the fol- total of 28 charactersfrom her analysis that were lowing approach: (a) only most parsimonious phylogenetically informative for the jays. The trees were used, (b) in those casesin which the remainder character(Character 1) was suggested solution included more than a single most par- by Peterson. In our analysis all character-states simonious tree, the signal was identified using

TABLE 2. Character-statematrix for 29 morphologicalcharacters for the sevengenera of jay.

Characters”

TZKlll I 2 3 4 5 6 7 8 9 IO 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 Cyanolyca 02211000000012112100110101001 Cyanocorax 0 1 0 0 1 1 1 1 1 1 0 0 1 1 2 1 1 3 0 0 0 0 1 1 1 0 0 1 0 Calocitta 01201111100010111200000110110 Cyanocitta 11210201111100111311112121010 Aphelocoma 1 1 0 1 0 2 0 1 1 1 1 1 0 0 1 1 1 3 2 1 1 1 2 1 2 1 0 1 0 Gymnorhinus0 0 0 2 0 2 0 1 2 1 1 1 0 0 2 1 0 3 2 2 1 2 3 1 2 0 1 1 0 Perisoreus 03120000210022002010220012022

*Characters and character-state descriptions are presented in Appendix 1 494 ALEJANDRO ESPINOSA DE LOS MONTEROS AND JOEL CRACRAFT strict consensus trees, (c) branch lengths of total of 332 phylogenetic informative characters, most-parsimonious trees were optimized by the from which 150 sites representtransversion sub- DELTRAN option in PAUP and (d) relative stitutions. Within the jays, the cytochrome b strengthof signal was examined by the bootstrap gene had a total of 177 phylogenetic informative test (1,000 replications), as well as by branch characters,and 43 when just transversionswere support (Bremer 1988, 1994). considered. The mean number of base substitu- tions was 153 (13.4%), a relatively high value RESULTS when compared to the paradisaeinines which had an average divergence of 111 (9.7%). The SEQUENCES AND SEQUENCE COMPARISONS transversion analysis showed an average diver- Complete sequencesof the cytochrome b gene gence of 47 substitutions (4.08%). The maxi- for Aphelocoma, Calocitta, Cyanocorax, Cyano- mum number of transversion differences among lyca, Gymnorhinus, and Perisoreus were ob- the corvids was found between Perisoreus and tained for this study. These sequenceshave been Calocitta at 76, and the minimum number was deposited in GenBank to allow an easier access 25 between Gymnorhinus and Cyanocitta. Table to these data. The accession number for the se- 3 summarizes the relative nucleotide divergence quences are: U77331 (Perisoreus canadensis), among the jays. U77332 (Gymnorhinus cyanocephalus), U77333 (Cyanolyca viridicyana), U77334 (Cyanocorax MOLECULAR PHYLOGENETICS chrysops), U77335 (Aphelocoma coerulescens), Examination of the eight basic types of nucleo- and U77336 (Calocitta formosa). Sequencesfor tide substitutions among jays and the outgroup the remainder taxa and their accessionnumbers taxa (i.e., Catharus, Ailuroedus and Ptilono- can be found in Helm-Bychowski and Cracraft rhynchus) reveals a clear saturation in third po- (1993). sition transitions (Fig. 2). Therefore, to avoid the No deletions or insertions in the 381 codons effect of excessive homoplasy in third positions, in the cytochrome b gene were detected among transversions at this position were up-weighted these birds. The termination codon for the six ten times over transitions (i.e., transitions weight sequenceswas TAA as commonly found in the = 0.1) when estimating phylogenetic trees. All chicken mitochondrial protein-coding genes other kinds of substitutions were considered (Desjardins and Morais, 1990). Recently, several with equal weight, assigning to them the same studies have reported nuclear copies of mito- weight as third position transversions (i.e., chondrial genes especially cytochrome b (Quinn weight = 1). We determined this weighting 1992, Smith et al. 1992, Kornegay et al. 1993). scheme via successive approximations decreas- These nuclear copies are characterizedby frame- ing the weight of third position transitions until shifts and nonsense codons, neither of which the phylogenetic analysis yielded a consistent were apparent in the sequencesthat we obtained. tree. Furthermore, based on the fact that the complete The result produced two equally most parsi- cytochrome b gene was isolated initially as a monious trees of 428 stepsin length with a con- single fragment, we assumethat these sequences sistency index of 0.607 and a retention index of are of mitochondrial origin only. Therefore, they 0.466. The only disagreement between the two do not represent pseudogenous or translocated topologies was the position of Cyanocitta as the fragments that might be located elsewhere in the sister taxon to the clade Aphelocoma-Gymno- jays’ genome. rhinus or to the clade Calocitta-Cyanocorax. The nucleotide ratio among the different taxa The signal contained in these trees was con- is typical for the avian cytochrome b gene. The densed by a strict consensustree and the result average relative frequencies of each kind of nu- is shown in Figure 3. cleotide for the DNA sequences was: 3 1.87% Using the Hermit Thrush to root the network, cytosine, 25.84% thymine, 29.22% adenine, and three monophyletic groups were resolved: the 13.08% guanine. Third positions showed an av- jays (Corvidae), birds-of-paradise (Paradisaei- erage divergence of 35.6%, followed by first po- dae), and bowerbirds (Ptilonorhynchidae). In all sitions with 8.28%, and finally second positions the analyses that we performed, the sister-group with 2.84%. to Corvidae was the birds-of-paradise. This was The complete data set (14 taxa), contained a a constant result even when the taxa within the INTERGENERIC RELATIONSHIPS OF THE NEW WORLD JAYS 495

TABLE 3. Pairwisecomparison of relativenucleotide divergence among the jays. TOT = Total; TS = Tran- sitions;TV = Transversions;lst, 2nd and 3rd = first, secondand third position;Its, 2ts and 3ts = first, second and third position transitions; ltv, 2tv and 3tv = first, second and third position transversions;K2 = Kimura’s two-parameter model.

TOT TS TV I St 2nd 3rd Its 1tv 2ts 2tv 3ts 3tv K2 Taxa % % % % % % % % % % % % %

Gymnorhinus Perisoreus 14.4 10.1 4.3 5.5 1.3 36.5 4.5 1.0 1.0 0.3 24.9 11.5 16.3 Cyanolyca Perisoreus 13.8 9.2 4.6 5.8 2.1 33.6 4.2 1.6 1.3 0.8 22.0 11.5 15.5 Cyanolyca Gymnorhinus 13.2 9.7 3.5 5.8 1.8 32.0 4.7 1.0 1.3 0.5 23.1 8.9 14.8 Cyanolyca Cyanocorax 14.3 9.8 4.5 7.3 3.4 32.0 6.0 1.3 2.1 1.3 21.3 10.8 16.1 Cyanolyca Cyanocitta 12.9 9.7 3.2 8.4 1.6 28.9 7.1 1.3 1.0 0.5 21.0 7.9 14.5 Cyanolyca Calocitta 14.9 9.8 5.1 8.7 4.5 31.5 6.6 2.1 3.1 1.3 19.7 11.8 16.8 Cyanolyca Aphelocoma 13.6 10.0 3.7 8.1 2.6 30.2 6.3 1.8 1.6 1.0 22.0 8.1 15.4 Cyanocorax Perisoreus 15.8 9.4 6.5 7.1 2.9 37.5 5.2 1.8 1.8 1.0 21.0 16.5 18.0 Cyanocorax Gymnorhinus 12.3 8.6 3.8 5.8 1.6 29.7 4.5 1.3 0.8 0.8 20.5 9.2 13.7 Cyanoc0ra.x Cyanocitta 12.6 8.9 3.7 6.6 1.8 29.4 5.0 1.6 1.0 0.8 20.7 8.7 14.0 Cyanocorax Calocitta 10.8 7.7 3.1 5.5 2.6 24.1 3.7 1.8 1.6 1.0 17.8 6.3 11.8 Cyanocorax Aphelocoma 13.6 9.6 3.9 8.4 2.9 29.4 6.8 1.6 1.6 1.3 20.5 8.9 15.2 Cyanocitta Perisoreus 14.3 9.5 4.7 5.8 1.6 35.4 5.0 0.8 1.3 0.3 22.3 13.1 16.1 Cyanocitta Gymnorhinus 10.7 8.5 2.2 5.0 0.3 26.8 4.2 0.8 0.3 0.0 21.0 5.8 11.8 Cyanocitta Aphelocoma 12.3 9.6 2.7 7.6 1.6 27.8 6.0 1.6 1.0 0.5 21.8 6.0 13.8 Calocitta Perisoreus 15.7 9.1 6.6 6.8 3.9 36.2 4.2 2.6 2.9 1.0 20.2 16.0 17.7 Calocitta Gymnorhinus 13.7 9.7 4.0 5.8 2.6 32.8 3.7 2.1 1.8 0.8 23.6 9.2 15.5 Calocitta Cyanocitta 13.0 8.9 4.1 7.3 2.9 28.9 5.0 2.4 2.1 0.8 19.7 9.2 14.5 Calocitta Aphelocoma 14.3 10.1 4.2 7.9 3.4 31.8 5.5 2.4 2.1 1.3 22.8 8.9 16.2 Aphelocoma Perisoreus 14.6 9.8 4.8 6.6 2.1 35.2 4.7 1.8 1.8 0.3 22.8 12.3 16.5 Aphelocoma Gymnorhinus 11.6 9.2 2.4 6.6 1.3 27.0 5.2 1.3 0.8 0.5 21.5 5.5 12.9 Averagedivergence 13.4 9.4 4.1 6.8 2.3 31.3 5.2 1.6 1.5 0.8 21.4 9.8 15.1 ingroup were changed and alternative taxa were Higher-level relationships in this phylogenetic used as an outgroup. Although the consensus hypothesis are consistent with the findings of tree depicts the phylogenetic relationships within other studies (Sibley and Ahlquist 1985, 1990, the New World jays as incompletely resolved, Helm-Bychowski and Cracraft 1993). However, the Gray Jay (Perisoreus) always appearsas the a branch support analysis showed that three of sister taxon to the New World jays. the five nodes in the jay clade collapse only one step away from the most parsimonious tree. In 26 - addition, the low bootstrap values registered 25 - . 0 0 within the jays can be considered as another in- dicator of clade instability. One explanation for spurious topologies is the presence of relatively small internodal distances compared to the lengths of terminal branches. Smith (1994) sug- gested that this imbalance between internodal distancesand terminal branches can increase the probability of rooting the network in the wrong position.

164 , , , , , , , , An important factor that affects tree topolo- 10 11 12 13 14 15 16 17 IS 19 20 gies is the use of outgroupsthat are too distantly Relative Divergence (%) related to the ingroup (Maddison et al. 1984, FIGURE! 2. Saturation curve third-position transi- Wheeler 1990, Smith 1994). To investigate this, tions plotted against relative sequence divergence. Catharus and the bowerbirds were eliminated Squares represent pairwise comparisons among the from the analysis and the birds-of-paradise were genera of jays; circles representpairwise comparisons chosen as the outgroup to root the jay phyloge- between the jays and the outgroup.Growing transitions differences leads to a plateau indicating a saturationin ny. Removing these taxa apparently eliminates third positions (line was adjustedby eye). the observed saturation in third position transi- 496 ALEJANDRO ESPINOSA DE LOS MONTEROS AND JOEL CRACRAFT

Cyanocitta A constant result in all these phylogenies was the high support for the clade Calocitta-Cyano- Aphelocoma corux. In the last analysis, the bootstrap value GyINIOIhiIIUS for that clade was 99%, and to break apart this

Calocitta clade it was necessary to incur an additional 13 steps from the most parsimonious solution. Cyanocorax CLADISTIC ANALYSIS OF MORPHOLOGICAL Cyanolyca CHARACTERS Perisoreus A cladistic analysis of the morphological data (Table 2; Appendix I), which included only the New World jay genera and Perisoreus, resulted in two equally parsimonious trees of 62 steps each. Using Perisoreus as the root, the relation- ships depicted in these trees are consistent with FIGURE 3. Strict consensustree of two equally par- the results obtained using the molecular data simonious trees (length: 428 steps; CI: 0.607; RI: 0.466). Third position transitionswere down-weighted alone. One of the trees shows exactly the same ten times with respectto transversions.Percentages are topology as the one obtained for the molecular bootstrapvalues obtained for 1,000 replications;num- data set. The alternative tree differs in the po- bers represent the branch support for the specified sition of Gymnorhinus which is placed as the node. Outgroup brancheshave been collapsed,for de- sister taxon to the clade Aphelocoma-Cyanocit- tails on species used refer to Taxa Examined under Methods. ta. Thus, the strict consensusof these two trees presentsan unresolved node for these three taxa. In this analysis, two nodes received significant tions; therefore, for this and the subsequentanal- values from the bootstrap test and branch sup- yses all characters were considered as equally port. The clade formed by Aphelocoma-cyano- weighted. Once this was done, a single most par- citta-Gymnorhinus was represented in 82% of simonious tree of 935 stepsin length was found the bootstrap replicates, and had a branch sup- (Fig. 4). This topology is consistent with the re- port value of three. The clade Calocitta-Cyano- sult obtained in the previous analysis, and the polytomy inside the jay clade is resolved. In this tree, Cyanocitla is the sister taxon to the clade Aphelocoma-Gymnorhinus, the sister-group to them is the clade Calocitta-Cyanocorax, and the most basal genus of the New World jays is Cy- anolyca. Finally, the sister-group to the New World jays is the Gray Jay (Perisoreus canaden- sis). A bootstraptest was computed for this phy- logeny and only the clade formed by Calocitta- - 13 Cyanolyca Cyanocorux had a substantially high value (98%). Although the rest of the nodes show low Perisoreus bootstrap values, an increase in the branch sup- port for each node suggeststhat the phylogeny Birds-of-Paradise has gained stability. FIGURE 4. Single most parsimonious tree for the Finally, if the interrelationships of the jays are jays using birds-of-paradise as the outgroup (length: examined using Perisoreus as the root of the 935 steps; CI: 0.588; RI: 0.388). Using this outgroup, no evident saturationis observed in the data, therefore tree, a single tree with a length of 582 steps, in this analysisall characterswere consideredwith the consistency index of 0.729, and retention index same weight. Percentagesare bootstrap values; num- of 0.325 is found. This result yields an identical bers represent the branch support for the specified pattern of relationships to that of the previous node. When the birds-of-paradise were deleted and Perisoreuswas used to root the phylogeny the same analysis. Although no changes are registeredfor topology was obtained using molecular, morphologi- the relationships among the genera, no improve- cal, and total evidence analysis. Letters are the refer- ment in bootstrapor branch support is observed. ence name for each node as used in Table 4. INTERGENERIC RELATIONSHIPS OF THE NEW WORLD JAYS 491

TABLE 4. Descriptive values for the analyses using general points with other higher-level hypothe- the Gray Jay as outgroup,and statisticsfor each node ses proposed for passerine interrelationships. as indicated in Figure 4. Our phylogenetic analysis (hypothesis) was able

Trees to identify as monophyletic the three main

Molecular Morphology Combined groups of birds represented (i.e., Corvidae, Par- adisaeidaeand Ptilonorhynchidae). The relation- Statistics ships depicted here among jays, birds-of-para- Length” 582 (1) 62 (2) 644 (1) C.Ib 0.129 0.823 0.738 dise and bowerbirds are consistent with previous R.I.’ 0.325 0.676 0.369 reviews of corvid phylogeny (Sibley and Ahl- Length next treea 583 (1) 63 (2) 652 (2) quist 1985, 1990, Edwards et al. 1991, Helm- BootstrapNode Bychowski and Cracraft 1993). With respect to 43% 44% 84% the New World jays, this phylogeny agrees with : 34% 82% 87% the proposal that these birds are a monophyletic 57% 60% 90% group. There are disagreements,however, when : 99% 90% 100% our hypothesisis contrastedto other phylogenies Branch Support Node at finer levels. 1 0 8 For instance, one of the hypothesesabout the : 2 3 8 history of the New World jays (hereafter re- I 1. 10 : 13 3 18 ferred to as Hardy’s hypothesis) proposesa sep- aration into two distinct lineages, the so called “The number in parentheses indicates the number of trees with that particular length. “ornate” and “inomate” lines (Hardy 1961, h Consistency index. / Retention index. 1969). The phylogenetic hypothesisproposed by us (Fig. 4), does not supportHardy ’s hypothesis. Incongruences arise with the position of Cyano- corux had a bootstrapvalue of 90% and a branch lyca. Based on external morphology and behav- support of three. The remaining node collapsed ior, Hardy’s hypothesis sustainsthe idea that the one step from the shortesttree. species in the genus Cyanolyca are the most COMBINED ANALYSIS primitive of the New World jays, which is con- The morphological and molecular data setswere sistent with our phylogeny. However, Hardy’s combined in a total evidence approach. The hypothesisalso postulatesthat Cyanolyca should search yielded a single most parsimonious tree be placed inside the inomate lineage along with of 644 steps, a consistency index of 0.738 and the genusAphelocoma. Hardy’s hypothesisarose a retention index of 0.369. The total evidence mainly because some speciesof Cyanolyca pro- tree corroborated the results obtained with the duce a relatively simple call very similar to other two data sets. In this analysis all the nodes those of Aphelocoma. Indeed, Cyanolyca was have bootstrapvalues over 80% and branch sup- discarded as a generic name, and its species port of at least eight. Bull et al. (1993) suggested were included within Aphelocoma. Our phylog- that combining different data sets in which re- eny suggeststhat these two genera are not each sults are not significantly different from one an- other’s closest relatives: Cyanolyca appears as other will result in a strongly corroboratedphy- the most basal clade within the New World jays, logenetic hypothesis. When the data sets were whereasAphelocoma is the sister taxon to Gym- analyzed independently, alternative trees were norhinus comprising one of the most derived only one step away from the most parsimonious clades. One of the major problems in Hardy’s tree, but in the combined analysis the nearest systematic study is that the so-called inomate trees were eight steps longer than the most par- lineage was established using plesiomorphic simonious solution. A comparison among the characters(e.g., no crests,uniform bluish above, three analyses and a detailed list of statistics is gray below, simple calls), therefore this lineage presented in Table 4. can not be accepted as a cladistic entity. DISCUSSION Hardy’s hypothesis recognized the remainder COMPARISONS WITH PREVIOUS of the genera as the ornate line, and lumped PHYLOGENIES them as a subgenus within Cyanocorax (Hardy The molecular phylogeny obtained for the com- 1969). Proposed evolutionary trends showed by plete set of taxa (Fig. 3) coincides in several the ornate line included: (a) shortening of crest, 498 ALEJANDRO ESPINOSA DE LOS MONTEROS AND JOEL CRACRAET

(b) increase in melanin leading to a darker and plex that is a shared character of all genera of simpler pattern of plumage, and (c) a reduction jays endemic to the New World, including Gym- of song repertoire. Based on these tendenciesthe norhinus, but which is absent from all Old subgenera Calocitta and Cyanocitta were con- World genera. This charactersupports the mono- sidered to be close to each other and to be the phyly of New World jays to the exclusion of sister-group to Cyanocorax. This conclusion Nucifraga and Perisoreus. There is additional contradicted a previous observation made by Pi- evidence supporting the hypothesis that Gym- telka et al. (1956) when describing a hybrid be- norhinus is a specialized New World jay that is tween Calocitta forrnosa and Cyanocorax morio. superficially similar to Nuczfraga. Both species The hybrid led Pitelka et al. to suggestthat these live in the same kind of coniferous forest under two speciesof jays were not very divergent and similar environmental conditions. The Pifion Jay probably should not be considered two separate and Clark’s base their diet on genera. seeds,and both use their for stabbing into Our phylogenetic hypothesis,in contrast, does unripe cones, hammering on bark, and probing not recognize the ornate line as a monophyletic the ground (Marzluff and Balda 1992). Similar- group, and in all analyses Cyanocorax was the ity in external morphology and behavior pre- sister taxon of Calocitta. Therefore, we do not sumably have converged in the two. Only Nu- accept the hypothesis that two evolutionary as- czfraga species possess a specialized structure semblagesexist within the New World jays such (i.e., sublingual pouch) for carrying as proposed by Hardy (1961, 1969). seeds, whereas the Pinion Jay uses a distensible Finally, Peterson (in Edwards and Naeem esophagus to resolve this problem. Thus, the 1993) proposed a phylogenetic hypothesis for presence of different morphological structures the New World jays based on electrophoretic for carrying seedssuggest that these two genera data. The phylogeny shows an unresolved clade independently acquired their specialized depen- formed by Aphelocoma, Cyanocitta, and Gym- dence on pine seeds (Ligon 1974). In addition, norhinus, which is the sister-group to an unre- the Pifion Jay sharesother features with the New solved clade consisting of Calocitta, Cyanocor- World jays that are lacking in Nucifraga: (1) the ax, Psilorhinus, and Cissilopha; the sister-taxon Pirion Jay is completely blue as are many New to these clades was the genus Cyanolyca. Al- World jays, (2) chicks of the Pifion Jay are na- though the tree is not resolved, Peterson’s results ked at birth, whereas down is present in nut- are consistent with the hypothesis presented crackers, crows and , and (3) the presence here. of the buttress complex in Gymnorhinus.

ON THE POSITION OF THE PIfiON JAY ON THE EVOLUTION OF THE BUTTRESS A key question investigated by those interested COMPLEX in the evolutionary history of New World corv- The development and complexity of the buttress ids has been the phylogenetic position of the Pi- complex varies among the different genera of non Jay (Gymnorhinus cyanocephalus).The re- New World jays. Specimens of Cyanolyca ex- sults presented here agree with the hypothesis amined in the American Museum of Natural that Gymnorhinusbelongs to the New World jay History collection showed a very poorly devel- assemblage,and they also supportthe placement oped buttress complex. Within Cyanocorax the of this genus as the sister-taxon to Aphelocoma, buttress is always present, but in some species well inside the New World jay assemblage. the buttress is lower, less vertical, and more Thus, another question can be raised: Why are widely separatedfrom the rostra1condyle of the there so many similarities in morphology and quadrate. In nestlings of Aphelocoma, the man- behavior between the Pifion Jay and Clark’s Nut- dibular buttressis fully developed at the time of cracker (Nucifraga columbiana)? Although this hatching, and the same seems to be true for Cy- analysis did not include any representativeof the anocitta and Gymnorhinus (Zusi 1987). genus Nucifraga, alternative morphological ev- The lack of a robust phylogeny for the New idence supports the hypothesis that nutcrackers World jays has made an interpretation of the are not members of the New World jay lineage. evolutionary history of this structure difficult. Zusi (1987) described a structural configuration Zusi was unable to decide if the intermediate of the jaw articulation called the buttress com- stages found in Cyanocorax represented the INTERGENERIC RELATIONSHIPS OF THE NEW WORLD JAYS 499

FIGURE 5. Biogeography of jays and scenario for the evolution of the buttresscomplex in the New World jays. Shaded areas are current distributionsof each genus. Letters representthe most parsimoniousexplanation in the change of the buttress complex: (a) ancestral state, (b) origin of the complex in the New World jays’ ancestor,(c) increment in complexity, and (d) maximum complexity reached. starting point of a morphocline or if they were BIOGEOGRAPHIC SCENARIO secondarily reduced. Furthermore, it was not Although absolute dates of diversification can- possible to infer whether the buttress complex not be obtained without reliable paleontological developed only once or multiple times within data, some studies have suggestedthat relative New World jays. Zusi’s interpretation of the evo- times can be estimated based on differences in lution of the group based on this charactermade transversion substitutionsin mitochondrial DNA him conclude that CyanoZycashould be the sis- (Brown et al. 1982, DeSalle et al. 1987, Miya- ter-group to the remaining genera of New World moto and Boyle 1989, Irwin et al. 1991). This jays. Building on the ideas presented by Zusi, clock principle bases its assumptions on a uni- the descriptions of the jaw articulation for the form rate of molecular evolution within a clade. jays can be interpreted within the context of the Peterson (1992) showed that some Aphdocoma phylogeny proposed in this analysis (Fig. 5). jay populations present evolutionary rates three The result is a gradual increase in complexity to four times greater than those of their sister toward the most derived lineages. It is possible, taxa, making the assumption of a molecular moreover, to identify three general patterns in clock for this group unreliable. Sibley and the development of the jaw articulation and each Ahlquist (1985) proposedthat the origin of cor- apparently evolved once. The rudimentary but- vines can be traced to Australia. Between 20 to tress complex of Cyanolyca probably resembles 30 million years ago the ancestor of crows and the condition that was present in the ancestorof jays was able to disperse to and the group the New World jays. An increase in complexity radiated in Asia and Europe. More recently an occurred in the ancestorof the rest of the genera invasion through Beringia resulted in the colo- and a similar form remains in Cyanocorux and nization of the New World by the corvids. Culocittu. Finally, the highest complexity was Figure 5 shows the distribution of each genus reached in the ancestor of Aphelocoma, Cyano- of jays along with their relative transversion di- citta, and Gymnorhinus. vergence. The result is compatible with the sce- 500 ALEJANDRO ESPINOSA DE LOS MONTEROS AND JOEL CRACRAET nario presented by Sibley and Ahlquist for the the Lewis B. and Dorothy Cullman ResearchFacility recent invasion of the New World. The basal po- at the AMNH and has received generoussupport from the Lewis B. and Dorothy Cullman Program for Mo- sition of a genus with a clear Palaearctic affinity lecular Studies, a joint initiative of the (Perisoreus) suggeststhat the origin of the New New York Botanical Garden and the AMNH. AEM World jays was the result from a single dispersal was supportedby a fellowship from the Mexican gov- event into the Americas. This event could have ernment (CONACYT # 80276). taken place around the early when the LITERATURE CITED climatic conditions were dominated by warm temperaturesand deciduousforest covered Alas- AMADON, D. 1944. The genera of Corvidae and their ka (Potts and Behrensmeyer 1992). A jay-like relationships.Am. Mus. Novit. 1251:1-51. AMERICAN ORNITHOLOGISTS' UNION. 1983. Check-list from late Miocene sediment of Colorado of North American birds. 6th ed. American Or- (Brodkorb 1972) suggests that the ancestor of nithologists’ Union, Washington,DC. jays must have dispersed throughout the Amer- ASHLEY,J. E 1941. A study of the structure of the icas rather rapidly. The available paleogeograph- humerusin the Corvidae. Condor 43: 184-195. BREMER,K. 1988. The limits of amino acid sequence ic data suggest that the older genera (i.e., Cy- data in angiosperm phylogenetic reconstruction. anolyca, Cyanocorax, and Calocitta) might have Evolution 42:795-803. entered and radiated in probably BREMER,K. 1994. Branch supportand tree stability. by the early when the Isthmus of Pan- 10:295-304. ama emerged. More recently, a secondary radi- BRODKORB,P 1972. Neogene fossil jays from the Great Plains. Condor 74:347-349. ation within took place and gave BROWN,W. M., E. M. PRAGER,A. WANG, ANDA. C. rise to Cyanocitta, Gymnorhinus, and Aphelo- WILSON. 1982. Mitochondrial DNA sequencesof coma. primates: tempo and mode of evolution. J. Mol. Along these analyses, we have followed the Evol. 18:225-239. BULL, J. J., J. l? HUELSENBECK,C. W. CUNNINGHAM,D. assumption that all the genera as defined are L. SWOFFORD,AND F! J. WADDELL. 1993. Parti- monophyletic. We acknowledge the need for de- tioning and combining data in phylogenetic anal- tailed analyses at the specieslevel, especially for ysis. Syst. Biol. 42:384-397. the genus Cyanocorax in which former genera CABOT,E., AND A. T. BECKENBACH.1989. Simulta- have been lumped (i.e., Cissilopha, Psilorhinus neous-editing of multiple nucleic acid sequences with ESEE. Comput. Appl. Biosci. 5:233-234. and Xanthoura). Another genus that certainly CRACRAF~,J. 1987. DNA hybridization and avian deserves attention is Cyanolyca, because the phylogenetics.Evol. Biol. 21:47-96. specific statusof some of their populations have CRACRAFT,J., AND K. HELM-BYCHOWSKI.1991. Par- been debated for a long time. Therefore, until simony and phylogeneticinference using DNA se- quences:some methodologicalstrategies, p. 184- some of these problems are settled the overall 220. In M. M. Miyamoto and J. Cracraft [eds.], intergeneric relationships among the New World Phylogenetic analysis of DNA sequences.Oxford jays presented here should be considered tenta- Univ. Press, New York. tive. DESALLE,R., T FREEDMAN,E. M. PRAGER,AND A. C. WILSON. 1987. Tempo and mode of sequence ACKNOWLEDGMENTS evolution in mitochondrial DNA of Hawaiian Drosophila. J. Mol. Evol. 26:157-164. We especially thank Sylvia Hope for allowing us to DESJARDINS,l?, AND R. MORAIS. 1990. Sequence and use her morphological data for this analysis. George gene organization of the chicken mitochondrial Barrowclough, Robert Rockwell, A. Townsend Peter- genome: a novel gene order in higher vertebrates. son, and three anonymousreviewers provided helpful J. Mol. Biol. 212:599-634. suggestionsto improve the final manuscript. We are EDWARDS,S. V., AND S. NAEEM. 1993. The phvloge- very grateful to the following curators for supplying netic componentof cooperativebreeding in perch- us with tissue samples from their collections: Mark ine birds. Am. Nat. 141:754-787. Hafner and Frederick Sheldon (Museum of Natural EDWARDS,S. V., l? ARCTANDER,AND A. C. WILSON. Science, Louisiana State University) and Adolf0 Na- 1991. Mitochondrial resolution of a deep branch varro-Sigiienza (Muse0 de Zoologia Facultad de Cien- in the genealogical tree for perching birds. Proc. cias, Universidad National Autonoma de Mexico). R. Sot. Lond. 243:99-107. Comments on earlier versions of this manuscriptwere GOODWIN,D. 1976. Crows of the world. Cornell Univ. provided by Gina Gould and Valerie Schawaroch.Par- Press, Ithaca, NY. tial support for laboratory work was provided by GYLLENSTEN,U. B., AND H. A. ERLICH. 1988. Gener- grants from the Frank M. Chapman Memorial Fund ation of single-strandedDNA by the polymerase and Sanford Fund of the Department of Ornithology, chain reaction and its application to direct se- American Museum of Natural History (AMNH). The quencing of the HLA-DQA locus. Proc. Natl. researchreported in this paper is a contribution from Acad. Sci. 85:7652-7656. INTERGENERIC RELATIONSHIPS OF THE NEW WORLD JAYS 501

HARDY,J. W. 1961. Studies in behavior and phylog- TORO. 1956. A hybrid jay from Chiapas Mexico. eny of certain New World jays (Garrulinae).Univ. Condor 58:98-106. Kansas Sci. Bull. 42:13-149. POTTS,R., AND A. K. BEHRENSMEYER.1972. Late Ce- HARDY,J. W. 1969. A taxonomic revision of the New nozoic terrestrial ,p. 419-541. In A. World jays. Condor 71:360-375. K. Behrensmeyer,J. D. Damuth, W. A. DiMichele, HARSHMAN,J. 1994. Reweaving the tapestry: what R. Potts, H. D. Sues, and S. L. Wing [eds.], Ter- can we learn from Sibley and Ahlquist (1990)? restrial ecosystemsthrough time: evolutionary pa- Auk 111:377-388. leoecology of terrestrialplants and . Univ. HELM-BYCHOWSKI,K., AND J. CRACRAF~.1993. Re- Chicago Press, Chicago. covering phylogenetic signal from DNA sequenc- QUINN, T W. 1992. The genetic legacy of Mother es: relationships within the Corvine assemblage Goose: phylogeographicpatterns of Lesser Snow (Class Aves) as inferred from complete sequences Goose Chen caerulescens caerulescens maternal of the mtDNA cytochrome b gene. Mol. Biol. lineages. Mol. Ecol. 1:105-l 17. Evol. 10:1196-1214. RIDGWAY~R. 1904. The birds of North and Middle HILLIS, D. M. 1991. Discriminating between phylo- America. Part III. U.S. Natl. Mus. Bull. 50. genetic signal and random noise in DNA sequenc- SANGER,E, S. NICKLON,AND A. R. COULSON.1977. es, p. 278-294. In M. M. Miyamoto and J. Cra- DNA sequencing with chain terminating inhibi- craft [eds.], Phylogenetic analysis of DNA se- tors. Proc. Natl. Acad. Sci. 74:5463-5467. quences.Oxford Univ. Press, New York. SIBLEY,G. C., AND J. E. AHLQUIST.1985. The phy- logeny and classification of the Australo-Papuan HOPE,S. 1989. Phylogeny of the avian family Cor- passerinebirds. Emu 85: I-14. vidae. Ph.D. diss., City Univ. New York, New SIBLEY,G. C., AND J. E. AHLQUIST.1990. Phylogeny York. and classification of birds: a study in molecular HOUDE,l? 1987. Critical evaluation of the DNA hy- evolution. Yale Univ. Press, New Haven, CT bridization studiesin avian systematics.Auk 104: SINGER-SAM.S.. R. L. TANGUAY.AND A. D. RIGGS. 17-32. 1989. Use’ of Chelex to improve PCR signals IRWIN, D. M., T. H. KOCHER,AND A. C. WILSON. 1991. from a small number of cells. Amplifications 3: Evolution of cytochrome b gene of . J. 11. Mol. Evol. 32:128-144. SMITH,A. B. 1994. Rooting moleculartrees: problems KOCHER,T D., W. K. THOMAS,A. MEYER, S. V. ED- and strategies.Biol. J. Linn. Sot. 51:279-292. WARDS,S. PAABO,E X. VILLABLANCA,AND A. C. SMITH,M. E, W. K. THOMAS,AND J. L. PA~ON. 1992. WILSON. 1989. Dynamics of mitochondrial DNA Mitochondrial DNA-like sequencein the nuclear evolution in animals: amplification and sequenc- genome of an akodontinerodent. Mol. Biol. Evol. ing with conservedprimers. Proc. Natl. Acad. Sci. 9:204-215. 86:6196-6200. SWOFFORD,D. L. 1993. PAUP version 3.1.1. Smith- KORNEGAY,J. R., T D. KOCHER,L. A. WILLIAMS,AND son. Inst., Washington,DC. A. C. WILSON. 1993. Pathways of lysozyme evo- SWOFFORD,D. L., AND-G. J. OLSEN. 1990. Phylogeny lution inferred from the sequencesof cytochrome reconstruction.D. 411-501. In D. Hillis and C. b in birds. J. Mol. Evol. 37:367-379. Moritz [eds.], -Molecular systematics. Sinauer, LIGON,J. D. 1974. Comments on the systematic re- Sunderland,MA. lationshipsof the Piilon Jay (Gymnorhinuscyan- WHEELER,W. C. 1990. Nucleic acid sequencephylog- ocephalus). Condor 76:468-470. eny and random outgroups.Cladistics 6:363-368. MADDISON,W. P, M. J. DONOGHUE,AND D. R. MAD- Zusr, R. L. 1987. A feeding adaptation of the jaw DISON. 1984. Outgroup analysis and parsimony. articulation in New World jays (Corvidae). Auk Syst. Zool. 33:83-103. 104:665-680. MADGE, S., AND H. BURN. 1994. Crows and jays: a APPENDIX 1 guide to the crows, jays and magpiesof the world. Houghton Mifflin, New York. Morphological charactersthat are phylogenetically in- formative for the genera presented in this analysis. MARZLUFF,J. M., AND R. F! BALDA. 1992. The Pinyon Character 1 was suggestedby Peterson(pers. comm.). Jay: behavioral ecology of a colonial and coop- The rest of the characterswere modified from Appen- erative corvid. Academic Press, San Diego. dix II of Hope (1989 and pers. comm.). MIYAMOTO,M. M., AND S. M. BOYLE. 1989. The po- tential importance of mitochondrial DNA se- CRANIAL CHARACTERS quence data to eutherian phylogeny, p. (1) Bony lateral bar rimming one side of the sclerotic 437-450. In B. Femholm, K. Bremer and H. Jom- ring. 0 = Absent; 1 = Present. vall [eds.], The hierarchy of life. Elsevier Science, (2) Fontanelles on the posterior walls of the orbit. 0 Amsterdam. = Absent; 1 = Bilateral fontanelles, smaller than PETERSON,A. T 1992. Phylogeny and rates of molec- the optic foramen; 2 = Larger, up to two times ular evolution in the Aphelocomajays (Corvidae). the size of the optic foramen; 3 = Over two times Auk 109:133-147. the size of the optic foramen. PITELKA,E A. 1951. Speciation and ecologic distri- (3) Interorbital fontanelle. 0 = Round shape, small bution in American jays of the genusAphelocoma. or similar to the diameter of the optic foramen; Univ. Calif. Publ. Zool. 50:195-464. 1 = Larger, up to three times the diameter of the PITELKA,E A., R. K. SELANDER,AND M. ALVAREZDEL optic foramen; 2 = Irregular shape and very 502 ALEJANDRO ESPINOSA DE LOS MONTEROS AND JOEL CRACRAFT

large, over three times the diameter of the optic rostrally about half way to the attachmentprocess foramen. of the muscle pseudotemporalissuperficialis; 2 = (4) Nasal processesof the premaxilla at the craniofa- Longer, nearly all the way to the attachmentpro- cial hinge. 0 = Narrow, less than one third the cess. width of the hinge, incomplete fusion of the (18) Rostra1cotyla of the mandible. 0 = Absent; 1 = paired processes;1 = Wider, nearly one third the Present,with a very small facet on the dorsalrim; width of the hinge; 2 = More than one third the 2 = Facet larger forming an articulation with a width of the hinge, the paired processescom- rostra1condyle of the quadrate;3 = High cotyla, pletely fused. as high as the dorsal rim of the ramus. (5) Lateral view of the upper jaw. 0 = Dorsal and (19) Ventral suprameaticcrest. 0 = Short, not elon- ventral border straight; 1 = Dorsal and ventral gated caudally; 1 = Long, the bulla notched at border continuouslydecurved, the ventral border the suprameaticcrest; 2 = Longer, the crest ex- continuouslydecurved. tended caudally. (6) Tip of the maxillopalatine. 0 = Not inflated; 1 = (20) Dorsomedial lobe on the suprameaticprocess of Inflated; 2 = Inflated, with one spur projected the ventral suprameaticcrest. 0 = Absent; 1 = caudally. Very small, sometimesinconspicuous: 2 = Large (7) Median septum of nasal capsule. 0 = No Ossi- and conspicuous. fication; 1 = Ossificationpresent. (21) Ventral suprameaticcrest at the quadrate. 0 = (8) Pit on the ventral surface of the transpalatine.0 Located close to the squamosalarticulation; 1 = = Absent; 1 = Present. In ventrolateral position from the squamosalar- (9) Premaxilla in the palate. 0 = Light ossification; ticulation, extendedslightly as a shelf; 2 = Form- 1 = Ossificationdense but the symphysisnot ful- ing a wide ventrolateralshelf. ly fused on the ventral view; 2 = Ossification (22) Dorsal suprameaticcrest at the bulla. 0 = Absent; dense, symphysiscompletely fused on the ventral 1 = Small, merges with the ventral suprameatic view. crest before joining the bulla; 2 = Well devel- (10) Fusion of the prepalatinebar to the ventral bar of oped, dorsal to the ventral suprameaticcrest at the jaw. 0 = Partially fused; 1 = Entirely fused. the quadrateand at the bulla. (11) Ventral lobe of the prefrontal. 0 = Inflated but (23) Abutment againstthe otic processof the quadrate small, lying mostly along the lateral border of the just medial to the squamosalcondyle. 0 = Ab- anteorbital plate; 1 = Highly inflated laterally sent; 1 = Presentbut small; 2 = Well developed and contractingthe jugal arch extensively. abutment but not ossified; 3 = Well developed (12) Postorbitalprocess. 0 = Located near to the cau- and heavily ossified. da1 wall of the orbit; 1 = Located on a more (24) Ventral view of the suprameaticprocess at the caudal position, continues along the lateral wall quadrate.0 = Narrow; 1 = Wide. of the cranium. (25) Subzygomatic sulcus. 0 = Absent; 1 = Present, extended slightly caudally up to the quadrate- (13) Zygomatic process.0 = Very small, looks like a small rostroventral projection from the orbital squamosal articulation; 2 = Extending beyond wall, the dorsal border abruptly curved; 1 = the quadrate-squamosalarticulation. Slightly longer, the dorsal border forming a shal- POSTCRANIALCHARACTERS low curve at the junction with the dorsal border (261 Pectoral crest of the humerus. 0 = The distal of the orbit; 2 = Long, wedge-shapedin lateral margin forms a continuousridge with the shaft; profile. 1 = The distal margin slightly outset from the (14) Jugal brace on the quadrate. 0 = Absent; 1 = shaft; 2 = The distal margin widely outset from Present but small; 2 = Present, forming a big the shaft. abutment againstjugal bar. (27) Dorsal epicondyle of the humerus.0 = Small; 1 (15) Otic processof the quadrate.0 = Narrow with a = Longer, extending beyond the level of the crest on the lateral border; 1 = Wider, the lateral proximal rim of the dorsal condyle. border roundedwith a small shaft; 2 = Shaft very Sacral vertebrae.0 = Six; 1 = Seven; 2 = Eight. big. Synsacrum.0 = Rectangularshape, the width be- (16) Rostral crest of the lateral cotyla of the mandible. tween the dorsolateralilliac crest is the same as 0 = Absent; 1 = Present. the length from trochanterto caudal limit of the (17) Rostra1groove of the medial cotyla of the man- fused vertebrae; 1 = Slightly wider; 2 = Very dible. 0 = Very short; 1 = Longer, extending wide.