BIOCHEMICAL SYSTEMATICS WITHIN PALAEOTROPIC (AVES: )

L. CHRISTIDIS Departmentof PopulationBiology, Research School of BiologicalSciences, Australian National University, P.O. Box475, CanberraCity 2601,Australia, and Divisionof Wildlifeand RangelandsResearch, C.S.I.R.O., Australia

ABSTR^CT.--Differentiationat 38 presumptiveloci was examined among 30 speciesof pa- laeotropicfinches (Estrildidae) by protein electrophoresis.Three speciesof Ploceidae,two of Fringillidae, and one of Emberizidaewere included for comparisonand the establishment of out-groups.Phenetic and cladisticanalyses were employed,and both producedconcordant patterns of relationshipsamong the species.I concludethat all four families are closely related, with Estrildidaeand Ploceidaegrouped on one major sublineageand Fringillidae and Emberizidae on the other. Within the Estrildidae, three distinct radiations are identified, correspondingto the waxbill (Estrildae),mannikin (Lonchurae),and grassfinch (Poephilae) groupingsin current classifications.Contrary to prevailing views, Aidemosyneis shown to be a memberof Poephilae,not Lonchurae,and to be allied with Neochmiaand Aegintha.Similarly, the relationshipsof Aeginthaas currently presumedwith the Estrildaeare not consistentwith the electrophoreticdata. Overall, the data suggesta monophyleticorigin for the Australasian Poephilae.Within this assemblage,however, Emblemaand Poephilaare clearly polyphyletic by current classifications.A major point of ambiguity in the data centerson the intercon- nectionsbetween the three major estrildid assemblages;at present, this can be treated only as an unresolvedtrichotomy. Received 24 October1985, accepted 15 August1986.

BEFORE1980, most ornithological studiesus- species(Barrowclough 1980) of limit its ing multilocus protein electrophoresiswere usein determininglevels of geneflow through confined to intraspecificvariation (Baker 1974, hybrid zones or between populations. This Corbin et al. 1974, Manwell and Baker 1975), weaknessbecomes its strength in determining and they revealed levels of variation compa- relationshipsamong species at genericand fam- rable to those of other vertebrates (Avise 1977). ily levels.It is only becauseof the low isozymic These studies also indicated a marked lack of divergenceencountered among birds that elec- differentiation between closely related species trophoresiscan be used productively up to in- (Smithand Zimmerman1976, Barrowclough and terfamilial comparisons(Avise et al. 1980c, Bar- Corbin 1978), a result that has now been found rowclough et al. 1981). Accordingly, I have to be a consistent feature in the Aves. In a series applied it here in an attempt to resolvespecies- of papers on comparativeprotein electropho- group relationships in the estrildine finches, resis within families, Avise et al. Estrildidae. (1980a-c, 1982) reported that genetic distances Delacour (1943) divided the estrildine finches in birds were much lower than those observed into three tribes on the basisof courtship,mouth in other vertebrates.Similar findings alsowere markingsof nestlings,and life habits.They are reported in nonpasserine families (Barrow- the waxbills (Estrildae), which are restricted cloughet al. 1981,Gutierrez et al. 1983,Adams largely to Africa; the grassfinches (Poephilae), et al. 1984). Whether the observedlow isozymic an Australasian-centeredgroup; and the man- genetic distancesare a consequenceof a slow nikins (Lonchurae),which are pan-palaeotropic rate of protein evolution (Avise et al. 1980c)or from Africa through southern Asia to Austra- an artifact of overestimating the age of avian lasia. While this grouping has been accepted taxa (Baker and Hanson 1966) remains unre- generally, the exact compositionof each tribe solved (Avise and Aquadro 1982). Such genetic has been in constantdispute. In particular, the propertiesplace obvious constraints on the use relationshipsof the genera Aegintha,Erythrura, of protein electrophoresisin evolutionary stud- Chloebia,Arnadina, and Aidernosynehave never ies. Thus, the low genetic distancesbetween been resolvedsatisfactorily. Mayr (1968) could populations(Barrett and Vyse 1982) and sub- not group the genera into three discrete tribes

380 The Auk 104: 380-392. July 1987 July 1987] BiochemicalSystematics of EstrildidFinches 381

TABLE1. Enzymesexamined, buffers used, and tissuedistribution of eachenzyme.

Run- Run- No. ning ning of buf- time b Enzyme (E.C. no.) Abbreviation loci Tissue fera (h) Isocitrate dehydrogenase(1.1.1.42) IDH 2 Muscle E 2 Glutamate-oxaloacetate transaminase (2.6.1.1) GOT 2 Muscle E 2.5 Glucose-phosphateisomerase (5.3.1.9) GPI 1 Muscle B 3 Mannose-phosphateisomerase (5.3.1.8) MPI 1 Muscle C 1.5 Glycerophosphatedehydrogenase (1.1.1.8) GPDH 1 Muscle A 3 6 Phosphogluconatedehydrogenase (1.1.1.44) PGDH 1 Muscle C 2.5 Malate dehydrogenase(1.1.1.37) MDH 2 Muscle A 1.5 Glyceraldehyde-3-phosphatedehydrogenase (1.2.1.12) GAPDH 1 Muscle D 3 Malic enzyme (1.1.1.40) ME 2 Muscle A 1.5 Peptidase C (3.4.11) PEP L-A 1 Muscle A 2 Peptidase B (3.4.11) PEP L-G-G 1 Muscle B 1.5 Esterasec (3.1.1.1) EST 2 Muscle A 1 Aldolase (4.1.2.13) ALD 1 Muscle D 3 Triose-phosphateisomerase (5.3.1.1) TPI 1 Liver D 3 Guanine deaminase (3.5.4.3) GDA 1 Liver C 1 Glutamatedehydrogenase (1.4.1.3) GDH 1 Muscle D 3 General protein (--) GP 5 Muscle A 3 Fumerase (4.2.1.2) FUM 1 Liver E 2 Aconitase (4.2.1.3) ACON 2 Liver, muscle F 2 Lactatedehydrogenase (1.1.1.27) LDH 2 Muscle, heart D 3 Phosphoglucomutase(2.7.5.1) PGM 2 Liver E 3 Hexokinase (2.7.1.1) HK 2 Muscle C 2 Superoxide dismutase (1.15.1.1) SOD 1 Muscle A 2 NADP specificdehydrogenase a NADP nDH 1 Liver C 2 NAD specific dehydrogenased NAD nDH 1 Muscle, liver D 3 a A 50 mM TEM, B - 15 mM TEB, C 50 mM TEM + NADP, D = 50 mM TEM + NAD, E = 0.1 M Tris-citrate, F = 0.01 M citrate-phosphate. See Baverstock et al. (1980) for buffer recipes. bAt 5 mA per 12 cm gel. ßUsed method A in Harris and Hopkinson (1976) with 4-methyl-umbelliferyl-acetate. d "Nothing" dehydrogenases,i.e. bandswere observedwithout the addition of any substrateto the stainingmixture. and arbitrarily accepted Delacour's (1943) re- cladisticanalyses. The resulting data are used vision with minor modifications. This reflects to reassessrelationships within the Estrildidae the fact that the Estrildidae are unusual among and the affinities of this family to other seed- birds in that plumage patterns are a relatively eating birds. poor clue to relationships owing to extensive convergencesand parallelisms (Harrison 1963, Mayr 1968). Moreover, the value given to the MATERIALS AND METHODS various morphological and behavioral charac- Specimensused were obtained from the wild and ters (Goodwin 1982) used is often subjective, from aviaries. Wild-caught specimens of several both in interpretation and in application. For speciescame from the Fitzroy River region of north- example, there is a gradation in the patterns of western Australia, and, whenever possible, prefer- nestling mouth markings between the grass- ence was given to material from such sources.For finches and mannikins (Delacour 1943) such that other speciesonly aviary-bredbirds were readily the separation between these two tribes is ar- available. In such cases,each sample was obtained from diverse sources in Victoria and New South Wales bitrary. Thus, although Mayr's (1968) arrange- ment was intended to be temporary, it still re- to minimize the chanceof obtaining related individ- uals.A list of the specimensexamined and their col- mains the only practicalclassification. lection localitiesis presentedin the Appendix. I analyzed protein products of 38 presump- Liver, muscle, and heart tissue were excised and tive loci in 30 speciesof estrildid finches by storedin liquid nitrogen.Tissues were homogenized both pheneticand cladisticmethods. Six species using the recipe of Baverstocket al. (1980). Individ- of Ploceidae,Fringillidae, and Emberizidaewere uals were screened for 25 enzyme systemsrepre- included for comparisonand as out-groupsin senting 38 presumptive loci (Table 1). Enzymes were 382 L. CHRISTIDIS [Auk,Vol. 104

stained according to the method of Harris and Hop- respectively(summarized in Avise and Aquad- kinson (1976) except GOT, where the procedure of ro 1982). Shaw and Prasad(1970) was used.All systemswere Distancevalues for speciesrepresenting the run in a celluloseacetate matrix on a paper support three related families Ploceidae,Fringillidae, (Cellogel, Chemetron, Italy). and Emberizidaeare given in Table 4. The in- The measuresof Nei (1978) and Rogers(1972) were used to estimate genetic distances between taxa. A terfamilial genetic distancesare low, ranging from 0.765 to 0.455, in contrast to the recorded pheneticanalysis by the UPGMA method(Sneath and Sokal 1973) was performedsummarizing the matrix distance acrossavian families of 1.00 (Avise and of Nei's(1978)/•, while the matrixof Rogers'(1972) Aquadro 1982). /• wassubjected to a distance-Wagnerprocedure (Far- ris 1972). In the latter computation,three speciesof CLUSTER ANALYSIS Ploceidaewere usedas an out-group.Both theseanal- yseswere performed on the BIOSYS package(Swof- Phenetic analysis.--The UPGMA-generated ford and Selander 1981). In addition, a qualitative phenogram (Fig. 1) weakly separatesthe Estril- analysisusing the method of FIennig (1966) was per- didae from the Ploceidaeand groups the es- formed.The rationaleand procedureof this analysis has been outlined in Baverstocket al. (1979) and Pat- trildids in three clusterscorroborating the tribes ton and Avise (1983).After an initial dichotomywith- of Delacour (1943) and Mayr (1968). They are in the Estrildidae was determined through the plo- the Poephilae (cluster 1), Estrildae (cluster 2), ceid out-group,each of the sisterestrildine lineages and Lonchurae (cluster 3) in the nomenclature were then treated as out-groupsto each other from of Mayr (1968). dichotomy to dichotomy. Apart from Aeginthatemporalis and Aidemo- synemodesta, the speciesin cluster 1--all Aus-

RESULTS tralasian-always have been consideredto be closely related (Morris 1958, Mayr 1968). Both ALLELIC FREQUENCIESAND HETEROZYGOSITIES A. temporalisand A. modestaare included here AND GENETIC DISTANCE DATA in the Poephilae, aligned with Neochmiaruff- cauda(Fig. 1). The mostdecisive split is between Allelic frequenciesfor the 30 variable loci are Poephilaguttata, P. bichenovii,and the rest of the presentedin an appendixthat is availablefrom tribe. Although the genetic distancebetween the author. The following loci were monomor- P.guttata and P. bichenovii isrelatively high (/• = phic in all species:GOT-2, HK-2, ME-1, ME-2, 0.21), they stand together apart from other Poe- MDH-1, MDH-2, NADnDH, GP-4, GP-5, TPI, philae. The remainingspecies of Poephilacluster and GDH. The range of heterozygosity mea- closelywith the -Aegintha-Aidemosyne sures (Appendix) is large (0.00-0.06, mean = assemblage,while Emblemaguttata is grouped 0.03), and the average proportion of polymor- with Neochmiaphaeton and not its congener,E. phic loci is 12% (see Appendix). These values picta. are comparableto the levels of geneticvariation The limited number of speciesand genera reported for birds in general (Selander 1976, examinedprecludes comprehensive analysis of Avise and Aquadro 1982). the relationshipswithin cluster 2, the African Matrices of genetic distance values for the Estrildae. Two points, however, merit com- Estrildidaeare presentedin Table 2. Exceptfor ment. First, melbaand P. phoenicoptera Poephilabichenovii, in which two subspecieswere cluster together and are close to the rest of the examined,all comparisonsare betweenspecies. Estrildae,despite their extensivekaryotypic dif- The Nei (1978) distance between P. b. bichenovii ferences(Christidis 1983). If Nei's (1978) /• is and P. b. annulosaof 0.01 (Table 2) is similar to related to time since divergence, then the that observedbetween subspecies of otherbirds branchingpatterns (Fig. 1) suggestthat the two (Corbin 1977).In comparison,genetic distances speciesof Pytiliahave diverged rather recently between congenericspecies range from 0.10 to both from each other and from the rest of the 0.25, while those between genera are higher, Estrildae. This in turn indicates that the marked rangingfrom 0.30 to 0.77 (Table3). Thesevalues chromosomaldifferences within Pytilia (Chris- are generallyhigher than thosereported in sev- tidis 1983) also have arisen recently. Secondly, eral other passerinefamilies where mean intra- the speciesof Estrildaeare phenetically more andintergeneric values of/• are0.10 and 0.26, distant from one another than are those within July1987] BiochemicalSystematics of Estrildid Finches 383

the Poephilae. Such a result is not unexpected spectively. This also accountsfor the discrep- becauseit is generally agreed that the Estril- anciesin the position of Lagonostictasenegala of didae arose in Africa (Goodwin 1982), so Afri- the Estrildae(Figs. 1 and 2). can estrildine taxa could be expected to be older An alternative to the distance-Wagner pro- and more distantly radiated than their Asian cedurefor analyzingelectrophoretic data is the and Australian counterparts. use of percent fixed differences (Baverstocket The third cluster, Lonchurae, is subdivided al. 1982, Adams et al. 1984). For most species into five groups(Fig. 1). The first group,which this method producesa Wagner network sim- includesthe genera Erythruraand Chloebia,and ilar to that presented in Fig. 2, and so it is not the second,which is the genus ,are repeatedhere. Somediscrepancies occurred. On distinct from the rest of the Lonchurae. With percentfixed differences,N. ruficaudais 8 units the exceptionof Paddaoryzivora, the remaining from N. phaeton,Aegintha temporalis, and Aide- speciesbelong to Lonchuraitself and separate mosynemodesta, while A. temporalisand A. mo- into three groups. The first of these is mono- destaare 6 units from one another. The problem typic with L. pectoralis.The secondis also mono- lies in the position of N. phaeton,which is typic with L. bicolor,though it may reflect the 19 units from A. modesta, 14 units from A. tem- fact that neither L. cucullatanor L. fringilloides-- poralis,but only 8 unitsfrom N. ruficauda,a result both supposedclose allies of L. bicolor--wasin- that produces negative branch lengths. Such cluded in the analysis. oryzivora and the inconsistenciesare quite common in electro- remaining five speciesof Lonchuracomprise the phoretic (Avise et al. 1980c)and immunological final group.Here, extremelyshort branch points (Farris1981) data and are probablythe result of separatethe specieseven though someof them, back-mutationsand convergences.Here, fixed particularly P. oryzivora,are morphologically allelic differences indicate that it is more ap- distinct. propriate to group N. phaetonwithin the Neoch- Wagneranalysis.--Unlike phenetic analyses, mia-Aidemosyne-Aeginthacluster than to align it the Wagner tree doesnot assumea constantrate with Emblemaguttata (Figs. 1 and 2). This align- of protein evolution (Fig. 2). The dendrogram ment is corroboratedby morphological,behav- generated in Fig. 2 uses three speciesof Plo- ioral (Goodwin 1982), and chromosomal (Chris- ceidaeas an out-group.It is clearfrom the branch tidis 1986) data. lengthsthat the rate of protein evolution varied Cladisticanalysis.--A Hennigian cladogram along the different lineages.Despite this, there drawn from qualitative analysis of electro- is considerable concordance between the morphs (Fig. 3) was based on three speciesof UPGMA and Wagner networks. Ploceidaeas an out-group. This method of anal- The branching patterns for the Poephilae in ysistakes account of convergenceand back-mu- both networks(Figs. 1 and 2) are essentiallythe tation in the constructionof phylogenetictrees same,the only differencebeing a minor switch (Baverstocket al. 1979, Patton and Avise 1983). in position between Aegintha temporalisand The specific charactersdefining each of the Neochmiaruficauda. Moreover, the five clusters branchesin the cladogram,the most robust of within the Lonchurae are the same. In the which are thosedefined by three or more char- branching sequencesof the Lonchurae, how- acter states, are listed in Table 5. ever, there are differences. These are due to In agreement with UPGMA and distance- unequalamounts of protein changealong a lin- Wagner analyses,the cladogram resolves the eage,as shown in the differencesin the lengths Estrildidae into the three same base clades. The of branches leading to Amadinaand relationshipsof the Lonchurae to the two other pectoralisin both networks, As a result the clades, Estrildae and Poephilae, however, are UPGMA analysisplaces L. pectoraliscloser to the ambivalent. Three derived character states of genusLonchura than to the genusAmadina, while the electromorphsdefining theseclades--LDH- the Wagner network clusters Amadina closer. l(c), PGDH(g), and PGM-2(g)--link the Lon- The arrangementin the Wagnertree reflectsthe churaewith the Estrildae;an alternativegroup-- greater number of ancestral character states GPDH(c),SOD(j), and NADP nDH(g)--links the (symplesiomorphies) and derived character Lonchuraeto the Poephilaeinstead, an arrange- states(synapomorphies) that L. pectoralisand ment supportedby the distance-Wagnerden- Amadina share with the rest of Lonchura, re- drogram.The UPGMA phenogramgroups the 384 L. CHRISTIDIS [Auk,Vol. 104 July1987] BiochemicalSystematics of Estrildid Finches 385 386 L. CHRISTIDIS [Auk, Vol. 104

TABLE3. Mean geneticdistances at different taxonomiclevels within the Estrildidae.

Comparisona,b /• (Nei 1978) SD SE Congenericin Poephilae (10) 0.124 0.073 0.007 Congeneric in Lonchurae (17) 0.048 0.049 0.003 Congenericin Estrildae(2) 0.208 0.084 0.042 Intergeneric in Poephilae (45) 0.281 0.097 0.002 Intergeneric in Lonchurae (49) 0.395 0.085 0.002 Intergeneric in Estrildae(19) 0.352 0.078 0.004 Interfamilial; Estrildidae vs. Ploceidae (93) 0.630 0.046 0.002

Number of pair-wise comparisonsis in parentheses. Speciescompositions of the genera are basedon the revision by Christidis(1987).

Estrildae with the Poephilae on the GOT-i(c) DISCUSSION synapomorphy. There is no singledominant factor that favors The electrophoretic data were analyzed by any of these three alternative branching pat- three different methods, after the approach of terns, although a closer connection between Patton and Avise (1983). I produced a best-fit Poephilae and Lonchurae is supported by the phylogenyby comparingthe resultsof all three Wagnerdendrogram and one of the alternative methodsfor agreementor discrepancy.In most Hennigian cladograms.Without other indepen- instancesthere was good agreementamong the dent evidence corroborating any of the alter- three analyses,and consequentlyphylogenetic native alignments, the relationship among the conclusions can be drawn with some confi- tribes remains an unresolvedtrichotomy. Al- dence.The comparisonalso pinpointed areasof though the degree of resolution of the Hen- conflict,identifying discrepanciessuch as in the nigian cladogram(Fig. 3) is less than in the relationshipsbetween Neochmiaruficauda and UPGMA (Fig. 1) and Wagner (Fig. 2) networks, N. phaeton.I testedthe phylogeniesobtained in the composition and relationships of species the present study by an analysis of estrildid within each of the estrildid tribes is generally karyotypes(Christidis 1986) and found broad consistentacross all three analyses.Thus, in the agreementwith the electrophoreticdata in the Hennigian cladogramboth Poephilaguttata and composition of Poephilae, Lonchurae, and Es- P. bichenovii form a distinct clade within the trildae, and the alignmentof speciesinternally. Poephilae;Emblema picta is separatedfrom all Although the electrophoreticdata subdivide other grassfinches on one electromorph;and the Estrildidae into three groupscorresponding the two Neochmiaspecies form a distinct clade, largely to the "tribes" of Delacour (1943) and consistent with data from fixed allelic differ- Mayr (1968), they do not establish definitive ences. Clades within the Lonchurae are all de- relationshipsamong the tribes. Other evidence fined by two or more derived characterstates shedslittle light on the question.From appen- (Fig. 3), and the clustersof speciesare similar dicularmusculature, Bentz (1979) suggested that to those obtained by the Wagner and UPGMA the Estrildaeand Lonchuraeform a clusterapart methods. There is parallel correspondencein from the Poephilae within the Estrildidae; but the Estrildae. this is not supported consistentlyby the elec-

TABLE4. Geneticdistance measures in the familiesexamined. Above diagonal: Rogers (1972) genetic distance; below diagonal:Nei (1978) unbiaseddistance.

Species Family 1 2 3 4 5 6 7 1. Poephilaguttata Estrildidae -- 0.433 0.450 0.419 0.533 0.504 0.538 2. Passer domesticus Ploceidae 0.547 -- 0.081 0.373 0.406 0.363 0.466 3. P. montanus Ploceidae 0.579 0.050 -- 0.403 0.422 0.361 0.473 4. Foudiamadagascariensis Ploceidae 0.530 0.456 0.504 -- 0.445 0.478 0.502 5. Cardueliscarduelis Fringillidae 0.751 0.514 0.530 0.589 -- 0.254 0.375 6. C. chloris Fringillidae 0.697 0.442 0.423 0.641 0.282 -- 0.375 7. Tiaris canora Emberizidae 0.765 0.621 0.632 0.693 0.460 0.455 -- July1987] BiochemicalSystematics ofEstrildid Finches 387

(Estrildae) invasion from Africa. This hypoth- esis was based on the assumption that these generawere more closelyrelated to African Es- trilda than to other Australian elements. A sec- ond invasion by primitive mannikins (Lon- churae) was said to have given rise to and Aidemosyne.In contradicting this hypoth- esis,my resultsand the myologicaldata of Bentz (1979) suggestinstead a single, monophyletic origin for the Australian grassfinches(Poephi- lae) that includes Emblema,Neochmia, and Ae- gintha.The behavioraland morphologicalsim- ilarities between membersof these genera and African Lagonosticta(Mitchell 1962) and (Delacour1943, Mitchell 1962)are evidently the result of convergencesand parallelisms, not common ancestry. Fig. 1. UPGMA phenogram based on Nei's Neochmia,Aegintha, and Aidemosyneoften are (1978)/•. placed in three separatetribes (Wolters 1957, 1981; Mitchell 1962). Goodwin (1982) argued for a closerelationship among them on the basis trophoreticevidence (compare Figs. 1-3). Opin- of nestling mouth markingsand plumage pat- ion has fluctuated, moreover, as to whether the terns, but the evidence was inconclusive be- Australian grassfinch fauna has been derived cause Aidemosyneshared several behavioral from one or several successiveinvasions (Mor- characterswith the Lonchurae.The electropho- ris 1958; Harrison 1963, 1967). Immelmann retic data nonethelesssupport a closerelation- (1962) and Harrison (1967) maintained that the ship among thesethree genera.The phylogeny Australian genera Emblema,Neochmia, and Ae- of the Poephilae (Figs. 1-3) also highlights the ginthawere all direct descendantsof a waxbill shortcomingsin plumage patterns as determi-

AFA LATLMA x/• AER "?',PE

EGUATE • UBE

N ASU

Fig.2. Distance-Wagnernetwork based on Rogers' (1972) b. Thetree is rooted by the"ut-group method" (Farris 1972)using the Ploceidae(PDO, PMO, FMA). Speciesabbreviations are defined in the Appendix. 388 L. CHRISTIDIS [Auk, Vol. 104

.PGU TABLE5. Electromorphsthat define dades (Fig. 3).

Clade Apomorphic characters PBI 1 ACON-2(b), PGM-2(d), Ndh(h) a PAC 2 PGDH(k) PCI 3 PGDH(h) 4 IDH-2(b), PGM~i(g) ß .• PPE 5 GP-2(e) -- AMO 6 GPDH(a), ACON-i(e), a PGM-2(c), LDH-i(f), a LDH-2(b) NRU 7 IDH-i(e), GOT-i(e), SOD(j) NPH 8 PGDH(g), MPI(d), HK-I(b), LDH-2(a) 9 GDA(a) • ATE 10 PGM-2(g) • EGU 11 GOT-i(d), Aid(d), GP-2(c), CP-3(a), LDH- l(e) • EPI 12 IDH-i(a), GOT-i(c), ACON-i(a), LA-2(h) 1• GOT-i(i) LPE 14 IDH-i(f), ACON-i(i), a PGM-i(c), PGM-2(f), CGO Est-l(d) 15 LDH-i(c), PGDH(g), PGM-2(g) ETR 16 GAPDH(a), ACON-I(h), SOD(c) lO LCA 17 Ndh(i), Est-l(c) LFL 18 GDA"(e) 19 LA-2(f) 8 ß LAT 20 GP-2(a) LMA • Charactersin which the plesiomorphicstate could not be deter- mined. LPU POR Goodwin (1982) contended that Amadina was a LBI specialized offshoot of the Estrildae becauseit 15 AFA shared several morphological and behavioral AER traits (songand nestling mouth markings)with ASU Pytilia.Such similarities, however, are evidently convergentor parallelisms.The genera Chloebia AAM and Erythrura,moreover, are sister members of ,71- EAS the Lonchurae and not related to Poephila(De- PME lacour 1943). The closeness of Chloebia and Erythrura(Fig. 1) is consistentwith the views PPH 16t, of Mitchell (1958) and Schodde and McKean UBE (1976) that the former is an arid-adapted rep- 1811•LSE resentativeof rain forest-inhabiting Erythrura. In this assemblage,only Lonchuraas presently constitutedis paraphyletic.In particular, L. pec- toralisand L. bicolorare separatefrom each other Fig. 3. Cladogrambased on electromorphs.Num- and from the rest of the genus, a distinction bers at branch refer to synapomorphies(Table 5). consistentwith their behavior (Guttinger 1976). Apart from A mandava,the speciesof Estrildae nants of relationship. A case in point is the examined here have always been grouped to- red upper-tail covertsof Emblemapicta, E. guttata, gether. Harrison (1961) believed that and Aeginthatemporalis, which have been used was a specializedderivative of the Lonchurae, to unite thesespecies in a single genus(McKean while Wolters (1957, 1981) saw similarities in 1975).This grouping,however, is clearlypoly- behavior and plumage patternsbetween Aman- phyletic (Figs. 1-3). dava and poephiline Aegintha.Neither view is The electrophoreticresults also demonstrate supported by the protein data, which demon- that the Lonchurae are a monophyletic assem- strate conclusively that Amandavais a member blageand include Amadina. Gutfinger (1976) and of the Estrildae. July1987] BiochemicalSystematics ofEstrildid Finches 389

The electrophoreticresults can also be used to provide relative time scalesof divergence. According to the calibration of the avian "mo- • • [ Poephi/ac•eti=gutIotach•oris lecular clock" by Gutierrez et al. (1983), most estrildid genera diverged about 8 million years ago and the three underlying tribes about 14 million yearsago. Basedon an incomplete and Fig. 4. UPGMA of the relationships between the depauperatefossil record, however, the average Estrildidae(Poephila guttata), Ploceidae (Passer domes- age of most passerine genera is 3.75 million ticus,P. montanus,and Foudiamadagascariensis), Frin- years (Romer 1966, Prager and Wilson 1980). gillidae (Caraduelischloris and C. carduelis),and Em- Given the large discrepancy between the pro- berizidae (Tiaris canora). tein and fossil time scales--a factor of 2--the accuracyof the latter must be questionedseri- [Nei's(1978)/3 = 0.46;see Table 4] and consis- ously. Here a combination of protein electro- tent with the classificationof Mayr and Amadon phoresiswith other corroborative data setssuch (1951), who treated them as subfamilies within as DNA-DNA hybridization, microcomplement the Fringillidae. Second, the separation be- fixation, and mitochondrial DNA sequence tween Estrildidae-Ploecidaeand Fringillidae- analysishas the potential to provide much more Emberizidae (Table 4) is not as great as would accurateestimates of times of divergenceamong be expected if the two groups were unrelated passerine taxa. (Bock 1960). On the other hand, affinity be- Relationshipsamong the Estrildidae,Ploceidae, tween thesetwo assemblagesis corroboratedby Fringillidae,and Emberizidae.--Inattempting to DNA-DNA hybridization data (Sibley and Ahl- determine an outgroup for the Estrildidae,three quist 1985).Thus, the similar palatine processes related families were examined (Fig. 4). The in these four families may be homologousand phenogram supportsthe current view that the evidence for monophyly, and are not conver- Ploceidaeand Estrildidae are more closely re- gent as argued by Bock (1960, 1963). The av- lated to one another than they are to any other erage Nei (1978) distance among the four fam- group (Bock 1960, Sibley 1970, Bock and Mo- ilies is only 0.59, which indicatesthat they are rony 1978). This is substantiatedfurther by the closephylogenetically. Before their phylogeny low overall genetic distance of 0.63 (Table 3) can be placed in context among other passer- between the ploceidsand estrildids compared ines, however, further work is required on such with the averageof 1.0 acrossavian families in families as the Nectariinidae, Alaudidae, and general (Avise and Aquadro 1982:fig. 2). How- Motacillidae.According to evidencefrom DNA- ever, the placing of Passerin the Ploceidaewas DNA hybridization (Sibley and Ahlquist 1985), questionedby Pocock(1966) and Sibley (1970), these three families form a natural assemblage who argued in favor of fringillid affinities. The with the Ploceidae,Estrildidae, Fringillidae, and electrophoreticdata do net support sucha con- Emberizidae. The results of my study clearly clusion(Fig. 4); insteadthey show that Passeris indicatethat multilocusprotein electrophoresis allied to other ploceidsand the Estrildidae,cor- could make a significant contribution in ex- roboratingthe resultsof DNA-DNA hybridiza- amining the affinities within this assemblage. tion (Sibley and Ahlquist 1985).

The relationships between Old World and ACKNOWLEDGMENTS New World seed-eatingoscines have been con- troversial, centering on whether similarities in I thank Prof. B. John, Dr. B. J. Richardson, Dr. R. the bony palate are due to monophyly (Tordoff Schodde,and Dr. D. D. Shaw for reading the manu- 1954, Mayr 1955) or reflect convergenceresult- script critically. Financial support for this study was ing from similar feeding strategies(Bock 1960, provided through a CommonwealthPostgraduate Re- search Award, an Australian Museum Postgraduate Raikow 1978).Although my electrophoreticdata Award, and the Australian National University. are limited, they are relevant to thesequestions. First, there is a clear separation between the LITERATURE CITED ploceid-estrildid and the fringillid-emberizid assemblages.The similarity between the Frin- ADAMS, M., P. R. BAVERSTOCK,D. A. SAUNDERS,R. gillidae and Emberizidae is surprisingly high SCHODDE,& H. Y. SMITH. 1984. Biochemicalsys- 390 L. CHRISTIDIS [Auk, Vol. 104

tematics of the Australian cockatoos (Psittaci- maxilla in the passeres.Bull. Mus. Comp. Zool. formes: Cacatuinae). Australian J. Zool. 32: 363- 122: 361-488. 377. 1963. Evolution and phylogeny in morpho- AvIsE, J. C. 1977. Is evolution gradular or rectan- logically similar groups. Amer. Natur. 97: 265- gular? Evidencefrom living fishes.Proc. Natl. 285. Acad. Sci. USA 74: 5083-5087. --, & J. J. MORONY,JR. 1978. Relationships of --, & C. F. AQUADRO.1982. A comparativesum- the passerinefinches (Passeriformes: Passeridae). mary of geneticdistances in the vertebrates.Pat- Bonner Zool. Beitr. 29: 122-147. terns and correlations. Evol. Biol. 15: 151-185. CHRISTIDIS,L. 1983. Extensive chromosomal repat- --, & J. C. PATTON. 1982. Evolutionary terning in two congeneric species:Pytilia melba, geneticsof birds. V. Genetic distanceswithin L. and Pytilia phoenicopteraSwainson (Estrildidae: Mimidae (mimic thrushes) and Vireonidae (vir- Aves). Cytogenet. Cell Genet. 36: 641-648. eos). Blochem. Genet. 20: 95-100. 1986. Chromosomal evolution within the --, J. C. PATTONß& C. F. AQUADRO. 1980a. Evo- family Estrildidae (Aves) I. The Poephilae. Ge- lutionarygenetics of birds.I. Relationshipsamong netica 71: 81-97. North American thrushes and allies. Auk 97:135- --. 1987. Phylogenyand systematicsof estrildid 147. finchesand their relationshipsto other seed-eat- --, & --. 1980b. Evolutionary ge- ing .Emu 87: in press. neticsof birds. II. Conservativeprotein evolution CORBIN,K.W. 1977. Genetic diversity in arian pop- in North American sparrowsand relatives.Syst. ulations. Pp. 291-302 in Endangeredbirds (S. A. Zool. 29: 323-334. Temple, Ed.). Madison, Univ. Wisconsin. --, & --. 1980c. Evolutionary ge- --, C. G. SIBLEY,A. FERGUSON,A. C. WILSON, A. netics of birds. J. Hered. 71: 303-310. H. BRUSHß& J. E. AHLQUIST. 1974. Genetic poly- BAKERßC. M. A., & H. C. HANSON. 1966. Molecular morphismin New Guinea starlingsof the genus geneticsof avian proteins.VI. Evolutionaryim- Aplonis.Condor 76: 307-318. plicationsof blood proteinsof elevenspecies of DELACOUR,J. 1943. A revision of the subfamily Es- geese.Comp. Blochem.Physiol. 17: 997-1006. trildinae of the family Ploceidae.Zoologica 28: BAKER,M. C. 1974. Genetic structure of two popu- 69-86. lations of White-crowned Sparrowswith differ- FARRIS,J. S. 1972. Estimatingphylogenetic trees from ent songdialects. Condor 76: 351-356. distance matrices. Amer. Natur. 106: 645-668. BARRETTßV. g., & E. R. VYSE. 1982. Comparative --. 1981. Distancedata in phylogeneticanalysis. geneticsof three Trumpeter Swan populations. Pp. 3-23 in Advancesin cladistics.Proc. 1st meet- Auk 99: 103-108. ing. Willi Hennig Soc. (V. A. Funk and D. R. BARROWCLOUGH,G.F. 1980. Geneticand phenotypic BrooksßEds.). differentiation in a wood warbler (genus Den- GOODWIN, D. 1982. Estrildid finches of the world. droica)hybrid zone. Auk 97: 655-668. LondonßBritish Museum (Natural History) and ß& K. W. CORI•N. 1978. Genetic variation and Oxford Univ. Press. differentiation in the Parulidae. Auk 95: 691-702. GUTIERREZ,R. J., R. M. ZINK, & D. Y. YANG. 1983. --, & R. M. ZINK. 1981. Genetic differ- Geneticvariationß systematicß and biogeographic entiation in the Procellariiformes. Comp. Blo- relationships of some galliform birds. Auk 100: chem. Physiol. 69B: 629-632. 33-47. BAVERSTOCK,P. R., M. ARCHERßM. ADAMS, & B. J. GUTWINGER,H. R. 1976. Ethology and of RICHARDSON.1982. Genetic relationshipsamong the generaAmadina, Lepidopygia and Lonchura(Es- 32 speciesof Australiandasyurid marsupials. Pp. trildidae). Bonn. Zool. Beitr. 27: 218-244. 641-650 in Carnivorous marsupials(M. Archerß HARRIS,H., & D. g. HOPKINSON. 1976. Handbook of Ed.). SydneyßAustralia, Royal Zoological Society enzyme electrophoresisin human genetics.Am- of New South Wales. sterdam, North Holland Publ. Co. -, S. R. COLEßB. J. RICHARDSON,& C. H. S. WATTS. HARRISONßC. J.O. 1961. Affinities of the Red Ava- 1979. Electrophoresisand cladistics.Syst. Zool. davat, Amandava amandava (L.). Bull. Brit. Orni- 28: 214-219. thol. Club 82: 126-132. ßC. H. S. WATTSßM. ADAMSß& M. GELDER. 1980. 1963. The taxonomic position of the Crim- Chromosomal and electrophoretic studies of son and Red-browed Finch. Emu 63.' 48- AustralianMelomys (Rodentia: Muridae). Austra- 56. lian J. Zool. 28: 553-574. 1967. Apparent zoogeographicaldispersal BENTZ,G. C. 1979. The appendicular myology and patterns in two avian families. Bull. Brit. Orni- phylogeneticrelationships of the Ploceidaeand thol. Club 87: 63-72. Estrildidae (Aves: Passeriformes).Bull. Carnegie HENNIG,W. 1966. Phylogeneticsystematics. Chica- Mus. Nat. Hist. No. 15. go, Univ. Illinois Press. BOCK,W. J. 1960. The palatine processof the pre- IMMELMANN,K. 1962. Beitrage zu einer vergleich- July1987] BiochemicalSystematics ofEstrildid Finches 391

enden Biologieaustralischer Prachtfinken (Sper- ROGERSßJ. S. 1972. Measuresof genetic similarity mestidae).Zool. Jahrb. Syst.90: 1-196. and geneticdistance. Studies in Genetics7. Univ. MANWELL, C., & C. M. A. BAKER. 1975. Molecular Texas Publ. 7213: 145-153. geneticsof avian proteinsXIII. Protein polymor- ROMER,A. S. 1966. Vertebratepaleontology, 3rd ed. phism in three speciesof Australianpasserines. Chicago,Univ. ChicagoPress. Australian J. Biol. Sci. 28: 545-557. SCHODDE,R., & J. MCKEAN. 1976. The relationships MAYR, E. 1955. Comments on some recent studies of somemonotypic genera of Australianoscines. on phylogeny. Wilson Bull. 67: 33-34. Proc. 16th Intern. Ornithol. Congr.: 531-541. 1968. The sequenceof genera in the Estril- SELANDER, R. K. 1976. Genic variation in natural didae (Aves). Breviora 287: 1-14. populations.Pp. 21-45 in Molecularevolution (J. --, & D. AMADON. 1951. A classification of re- AyalaßEd.). Sunderland, Massachusetts,Sinauer cent birds. Amer. Mus. Novitates No. 1496. Assoc. MCKEAN,J. L. 1975. "Ploceidae." Pp. 22-23 in In- SHAW,C. R., & R. PRASAD.1970. Starch gel electro- terim list of Australian (R. Schodde, phoresis--acompilation of recipes.Blochem. Ge- Ed.). MelbourneßRoyal AustralasianOrnitholo- net. 4: 297-320. gists Union. SIBLEY,C.G. 1970. A comparativestudy of the egg- MITCHELLßJ.G. 1958. The taxonomicposition of the white proteins of passerinebirds. Bull. Peabody GouldJan Finch. Emu 58: 395-411. Mus. Nat. Hist. 32. 1962. The taxonomicposition of the Crim- ß & J. E. AHLQUIST.1985. The phylogeny and son Finch. Emu 62: 115-125. classificationof the Australo-Papuanpasserine MORR•S,D. 1958. The comparativeethology of grass- birds. Emu 85: 1-14. finches (Erythruae) and mannikins (Amadinae). SMITH, J. K., & E.G. ZIMMERMAN. 1976. Biochemical Proc. Zool. Soc. London 131: 389-439. geneticsand evolution of North American black- NEI, M. 1978. Estimationof averageheterozygosity birds (family: Icteridae).Comp. Blochem.Phys- and genetic distance from a small number of in- iol. 538: 319-324. dividuals. Genetics 89: 583-590. SNEATH, P. H. A., & R. R. SO•AL. 1973. Numerical PAttON, J. C., & J. C. gvIsE. 1983. An empirical eval- .taxonomy.San FranciscoßW. H. Freeman. uation of qualitative Hennigian analysesof pro- SWOFFORD,D. L., & R. K. SELANDER.1981. Acomputer tein electrophoreticdata. J. Mol. Evol. 19: 244- program for the analysis of allelic variation in 254. genetics.J. Hered. 72: 281-283. Po½ocK,T.N. 1966. Contributionsto the osteology TORDOFF,H.B. 1954. A systematicstudy of the avian of African birds. Proc. 2nd Pan African Ornithol. family Fringillidae basedon the structureof the Congr. 14: 83-94. skull. Misc. Publ. Mus. Zool. Univ. Michigan 81: PR•GER,E. M., & A. C. W•LSON. 1980. Phylogenetic 1-41. relationshipsand ratesof evolutionin birds. Proc. WOETERS,H.E. 1957. On the generaEstrilda(Swains) 17th Intern. Ornithol. Congr.: 1209-1214. and Lagonosticta(Cub). Bull. Brit. Ornithol. Club RA•Icow,R.J. 1978. The appendicularmyology and 77: 62-63. relationshipsof the New World nine-primaried 1981. Die systematischeStellung des dor- oscines(Aves: Passeriformes). Bull. CarnegieMus. nastrilds,Aegintha temporalis (Latham) (Aves, Es- Nat. Hist. 7: 1-43. trildidae). Bonn. Zool. Beitr. 32: 137-144. 392 L. CHRISTIDIS [Auk, Vol. 104

APPENDIX. Speciesexamined, samplesizes, localities, and genetic variability measures.

Percentage Locality• of loci Mean Species Abbreviation (samplesize) polymorphicb heterozygosity Estrildidae Poephilaguttata PGU A (15) 21.1 0.053 P. bichenoviibichenovii PBI A (4), C (6) 15.8 0.016 P. b. annulosa PAN B (3) 5.3 0.026 P. acuticauda PAC A (8), B (9) 26.3 0.062 P. cincta PCI A (4) 18.4 0.059 P. personata PPE A (2), B (2) 7.9 0.026 Aidemosynemodesta AMO A (5) 18.4 0.058 Aeginthatemporalis ATE A (5), C (2) 15.8 0.041 Emblemaguttata EGU A (4) 13.2 0.046 E. picta EPI A (3) 10.5 0.440 Neochmiaruficauda NRU A (9), B (16) 26.3 0.011 N. phaeton NPH B (5) 10.5 0.037 Chloebiagouldiae CGO A (6) 10.5 0.013 Erythruratrichroa ETR A (3) 2.6 0.009 Lonchurapectoralis LPE A (1), B (11) 13.2 0.015 L. fiaviprymna LFL A (3), B (1) 2.6 0.013 L. malaccaatricapilla LAT A (2) 2.6 0.013 L. maja LMA A (1) 0.0 0.000 L. castaneothorax LCA A (4), B (5) 15.8 0.012 L. bicolor LB! A (2) 2.6 0.013 Amadinafasciata AFA A (4) 2.6 0.000 A. erythrocephala AER A (4) 5.3 0.013 Amandavasubfiava ASU A (3) 5.3 0.018 A. amandava AAM A (1) 0.0 0.000 Estrildaastrild EAS A (3) 2.6 0.000 Uraeginthusbengalus UBE A (3) 7.4 0.044 Lagonostictasenegala LSE A (3) 5.3 0.009 Pytiliamelba PME A (5) 5.3 0.011 P. phoenicoptera PPH A (6) 2.6 0.000 Ploceidae Passerdomesticus PDO C (16) 13.2 0.023 P. montanus PMO D (3) 7.9 0.018 Foudiamadagascariensis FMA A (2) 0.0 0.000 Fringillidae Cardueliscarduelis CCA D (2) 2.6 0.009 C. chloris CCH D (2) 5.3 0.026 Emberizidae Tiariscanora TCA A (3) 5.3 0.018 = aviarystock; B = FitzroyRiver region, northwestern Australia; C = AustralianCapital Territory region; D = Melbourne,Australia. locuswas considered polymorphic if the frequencyof the mostcommon allele did not exceed0.99.