Evolutionary Genetics of Birds. V. Genetic Distances Within Mimidae (Mimic Thrushes) and Vireonidae (Vireos)

Home , Crissal thrasher, Mimid, Mimus, Toxostoma

Biochemical Genetics, Vol. 20, Nos. 1/2, 1982

John C. Avise, 1 Charles F. Aquadro, 1 and John C. Patton I

Received 22 Apr. 1981--Final 17 June 1981

Genetic distances (D's) between five species within each of the families Mimidae and Vireonidae were estimated from frequencies of protein electromorphs at 23 loci. For three mimid species in the genus Toxostoma, D equals 0.084 (range, 0.069-0.104); and among three mimid genera, D equals 0.223 (0.167-0.278). These distances typify values previously reported in other birds at comparable levels of taxonomic recognition. In sharp contrast, the mean genetic distance among five congeneric species of Vireonidae is far higher, D = 0.360 (0.027-0.578). One possible explanation for these results is that Vireo species are considerably older, on the average, than are species of Toxostoma or than are members of severai other avian genera assayed to date. Conventional thought about the origin and relative age of the Vireoni- dae appears compatible with this explanation. Although genetic distances in the Vireonidae are large by "avian standards," they remain modest or even small in comparison with distances between many nonavian vertebrate congeners. Results for the Mimidae and the Vireonidae are directly contrasted with genetic distances in well-known genera of Amphibia and Reptilia.

KEY WORDS: genetic divergence; protein electrophoresis; phylogeny; Ayes.

This research was supported by NSF Grant DEB 7814195 and by a grant from the American Philosophical Society. ' Department of Molecular and Population Genetics, University of Georgia, Athens, Georgia, 30602. 95 0006-2928/82/0200-0095503.00/© 1982 Plenum Publishing Corporation 96 Avise, Aquadro, and Patton

INTRODUCTION An emerging generalization from electrophoretic comparisons of structural gene products is that genetic distances within genera and families of birds are usually smaller than genetic distances within most genera and families of nonavian vertebrates. This "conservative" pattern of avian protein divergence has been reported in the Icteridae (Smith and Zimmerman, 1976), Muscica- pidae (Avise et al., 1980a), Emberizidae (Avise et al., 1980b), Parulidae (Barrowclough and Corbin, 1978; Avise et al., 1980c), and Anatidae (Patton and Avise, 1982). The conservative pattern of protein evolution in birds might conceivably be a reflection of the recent origin of many avian species surveyed thus far. In this study we examine genetic distances at 23 loci in five species of vireos (Vireonidae). Vireos represent a relatively ancient group of Passeriformes that one might expect to exhibit large protein differences. Therefore, we hypothesize that the Vireonidae are an excellent group to test the generalization of conservative protein evolution in birds. Because specimens of "mockingbirds and thrashers" (Mimidae) were available, an opportunity was also provided to examine genetic distances between several species and genera in that family. Genetic distances in birds have been explicitly contrasted with those in selected fish and mamma!s (Avise et al., 1980b,c). Here we will compare the genetic distances of Vireonidae and Mimidae against previously reported distances in representatives of two other classes of vertebrates--the reptiles and amphibians.

MATERIALS AND METHODS The Mimidae and Vireonidae are compact New World Passeriforme families, comprised of about 31 species (13 genera) and 43 species (4 genera), respectively. [For general taxonomic discussion, we will arbitrarily follow the recent classification of Morony et al. (1975).] The sample of species studied (Table I) was governed solely by the availability of fresh specimens, most of which were TV "tower-kill migrants" kindly donated by staff of the Tall Timbers Research Station, Leon County, Florida. Specimens of Mimus polyglottos, Toxostoma curvirostre, and T. dorsale were collected in Brews- ter County, Texas, and two other specimens of Mimus polyglottos were taken in Athens, Georgia. Standard techniques of one-dimensional starch-gel electrophoresis were used to assay variation in proteins encoded by 23 genetic loci. With slight modifications, the techniques employed, the loci assayed, and their designa- tions are similar to those described in detail by Avise et al. (1980a). Electromorphs were assigned numerical values determined by their mobilities relative to the mobilities of the common electromorph at each locus in Dumetella carolinensis (Mimidae) or Vireo olivaceus (Vireonidae). Only Genetic Distances in Birds 97

Table I. Estimatesof Mean Heterozygosity(H) per Sample, Based on 23 Assayed Loci

Species English name Sample size H _+ SE

Mimidae (1) Dumetella carolinensis Gray catbird 24 0.018 _+ 0.007 (2) Mimus polyglottos Northern mockingbird 8 0.010 +_ 0.007 (3) Toxostoma rufum Brown thrasher 7 0.048 _+ 0.024 (4) Toxostoma curvirostre Curve-billedthrasher 4 0.000 (5) Toxostoma dorsale Crissal thrasher 1 0.000 Totals 44 0.015 Vireonidae (1) Vireo griseus White-eyedvireo 5 0.036 _+ 0.028 (2) Vireo philadelphicus Philadelphiavireo 1 0.000 (3) Vireoflavifrons Yellow-throatedvireo 4 0.043 _+ 0.026 (4) Vireo solitarius Solitary vireo 16 0.054 _+ 0.021 (5) Vireo olivaceus Red-eyed vireo 58 0.048 _+ 0.015 Totals 84 0.036 intrafamily comparisons were attempted, so shared electromorph designa- tions do not imply electrophoretic identity between the two families. Heterozygosities (H's) were determined by averaging the proportions of heterozygous individuals observed per locus. Genetic similarities (I's) and distances (D's) between species were calculated using Nei's (1972) formulas, and matrices of D were used to generate phenetic dendrograms by UPGMA [unweighted pair-group method of analysis with arithmetic means (Sneath and Sokal, 1973)]. The "fits" of distances implied by the dendrograms to distances in the matrices from which the dendrograms were derived were evaluated by the F statistic of Prager and Wilson (1978). Smaller values of F indicate greater congruence, but any F < 0.10 implies good fit. Several of the species assayed were represented by few individuals (Table I). The use of small sample sizes (when larger samples were unavaila- ble) for the purpose of calculating mean genetic distance between species can be justified on both empirical (Gorman and Renzi, 1979) and theoretical (Nei and Roychoudhury, 1974) grounds. Although larger samples would undoubtedly result in slightly different estimates of D, the magnitude of this sampling bias is trivial when compared against the magnitude of differences in D's among birds, amphibians, and reptiles reported here.

RESULTS

Genetic Variation For the Mimidae, the per-species counted H was 0.015, and for the Vireoni- dae, H was 0.036 (Table I). Three species (Toxostorna curvirostre, T. 98 Avise, Aquadro, and Patton

ss sssss s ss ss ~OOOOO~OOOO OOOOOO

O ~ O o % O

r~ I I

O OOO~ ..= I

? OO c- I I

ol Oo O I I

c~ OO I I

I o

m I I

I I I Genetic Distances in Birds 99 dorsale, and V. philadelphicus) lacked detectable variation at the loci assayed, but these species were represented by few individuals (<4). When mean heterozygosities per species are weighted by sample size, values become H = 0.019 and H = 0.048 in the Mimidae and Vireonidae, respectively. The estimates of H for Vireonidae are very close to mean values previously reported in other Passeriformes (Avise et al., 1980c), while estimates of H for the Mimidae average about one-third to one-half as large. Frequencies of alleles in the two families (Table II) form the empirical base for D calculation and for the following brief discussion of genetic relationships.

Relationships Within Mimidae The three assayed species of Toxostoma comprise a singl__e branch of the dendrogram and cluster at a low level of genetic distance (D = 0.09; Fig. 1). The samples of Mimus polyglottos from Texas and Georgia shared common

MIMIDAE (mimic thrushes) Toxostoma dorsale Toxostoma curvirostre Toxostoma rufum Mimus polyglottos Dumetella carolinensis _•1• Hylocichla mustelina J 1.14 i 1.12 i 1,0i i 0 !8 10!6 I 01.4 I 01.2 I 0

Plethodon (woodland salamanders) ~ fourchensis ouachitae caddoensis ~ glutinosus yonahlossee 1,14 i 1.12 i 1.10 i 01.8 i 01,8 i 01,4 i 012 i 0i

Anolis (anole lizards)~ poncensis cristatellus evermanni --~"- - 4 cuvieri I. cybotes 1. |4 I t,12 I 1,10 I 018 I 01,6 I 014 I 01,2 I 0I

GENETIC DISTANCE Fig. I. Protein dendrograms of the Mimidae, and of representative salamanders (Plethodon) and lizards (Anolis), plotted on a common scale of Nei's (1972) genetic distance. Data for the subset of closely related species within Plethodon are from Duncan and Highton (1979), and Anolis data are from Wyles and Gorman (1980). Fit to the distance matrix for Mimidae is F = 0.054. I00 Arise, Aquadro, and Patton electromorphs at all loci and hence were lumped for further analyses. In the dendrogram Mimus polyglottos joins Toxostoma at D = 0.19, and Dume- tella carolinensis follows at D = 0.25 to complete the UPGMA array. Many authors have considered the Mimidae to be closely allied to the thrushes, Turdidae [Muscicapidae in the classification of Morony et al. (1975)]. Beecher (1953), for example, classified mimic thrushes (Mimidae) as a subfamily of Turdidae, and several other recent classifications agree (Sibley, 1970). For this reason, we compared all of our samples of the Mimidae to the wood thrush (Hylocichla mustelina) (N = 7). At 23 loci, the mean genetic distance of Hylocichla to the assayed mimids (D = 0.68; Fig. 1) is nearly 2.5 times as large as the greatest assayed distance between any two mimids (D = 0.28 between Dumetella carolinensis and Toxostoma rufum). Earlier, we compared Dumetella carolinensis against seven species of the Turdidae at 25 to 27 loci (Avise et al., 1980a). The mean genetic distance, D = 0.76, was very close to the distance of Hylocichla to the Mimidae. These distances are slightly larger than the D = 0.48 observed between the chat (Icteria) and other genera of the Parulidae (Avise et al., 1980c) or the D = 0.59 observed between selected representatives of two related seed-eating families (Morony et al., 1975), Emberizidae and Fringillidae (Avise et al., 1980b). Thus, the limited data currently available are consistent with the suggestion that genetic distances between highly divergent, confamilial genera, or between very closely related avian families, often lie in the approximate range of D = 0.40-0.80. The use of Hyloeichla as an "outgroup" taxon for the Mimidae revealed no apparent synapomorphs (shared-derived alleles) within the Mimidae. For this reason we were unable to employ qualitative Hennigian analyses to our data (Patton et al., 1981).

Relationships Within Vireonidae Vireo solitarius and V. flavifrons are genetically very similar (D = 0.027) and form a cluster later joined by V. philadelphicus, V. griseus, and V. olivaeeus, in that order (Fig. 2). V. solitarius and V. flavifrons are generally thought to be closely related species within the solitarius group of the subgenus Vireo (Hamilton, 1962), and V. griseus has been placed in a basally distinct species group (griseus) within that subgenus. V. olivaceus and V. philadelphicus are sometimes placed in the subgenus Vireosylva but are thought to be only basally related species within that subgenus (Hamilton, 1962). Thus, it is somewhat surprising that V. philadelphicus appears closer genetically to members of the subgenus Vireo than to V. olivaceus. However, the most striking aspect of our data is the large genetic distance between V. olivaceus and other assayed vireos. V. olivaceus appears completely distinct Genetic Distances in Birds 101

solitarJus V)reo (vireos) ~ .•IP b fiavifrons ~ .~.~ ohiladelphicus / ~ gri~eus - ~ ohvaceus i J 1 ~ l J l q i i i , 20 1.8 1.8 L'4 1.12 ....1.0 08 0'.6 .t 0.4 0.2 0

clnerea Hyla (tree frogs) gratJosa

squirella ~- ~ femoralis ~ chrysosce~is

avivoca

2 1.B 1.6 1.4 1.2 1.0 (3.8 . 0,4 0.2 0

aorora Rana (true frogS) nloscosa boyli(

~ cascadae ~ pretiosa ~1, catesbeiana = tarahumarae

1.8 1.6 1.4 1. 1. 0~8 0.6 04 O. GENETIC DISTANCE Fig. 2. Protein dendrograms of the genus Vireo, and of representatives of two genera of frogs (Hyla and Rana), plotted on a common scale of Nei's (1972) genetic distance. Data for the genus Rana are from Case (1978; see also Case et al., 1975), and the genus Hyla data are from Etges (1979). Fit to the distance matrix for Vireonidae is F ~ 0.080. in allelic composition from other vireos at a minimum of 9 of 23 loci (Table II). We have earlier compared Vireo olivaceus with Catharus ustulatus (Muscicapidae) and 28 species of warblers (Parulidae) (Avise et al., 1980c). Mean genetic distances were D = 1.45 and D = 1.24, respectively, much larger than any D's within Vireonidae or between Mimidae and Muscicapi- dae.

DISCUSSION Magnitudes of genetic distance within and between genera of the Mimidae are characteristic of values previously reported in other avian families. Congeneric species typically exhibit D's near 0.10, and confamilial genera typically exhibit D's in the range of about 0.20-0.40. Results are summarized in Fig. 3. 102 Avise, Aquadro, and Patton

WITHIN GENERA tCTERIDAE (1) I MUSCICAPIDAE (6) ~,~ EMBERIZIDAE (7) l PARULIDAE (74) I ANATIDAE (59) I • I MIMIDAE 13) ', • , VlREONIDAE (10) I • [ 070I I o' o o',o o' o o2o' o11o 000I BETWEEN GENERA ICTERIDAE (20) i U i MUSCICAPIDAE (15) i | u EMBERIZlDAE (59) o i o PARULIDAE (304) • I • I ANATIDAE (266) •

MIMIDAE (7) I • ',

0.70' 0.60' 0.50' 0.40' 0. '3 0 0.20' 0 I 10 0.00' GENETIC DISTANCE Fig. 3. Means (n) and ranges (~) of observed genetic distance in Aves, as estimated by conventional multilocus electrophoresis; above, distances among congeneric species; below, distances among confamilial genera. Numbers in paren- theses indicate numbers of pairwise species comparisons. Data were summarized from the references listed in the first paragraph of the Introduction.

In contrast, by "avian standards" the observed genetic distances in the genus Vireo are exceptionally large (Fig. 3), averaging about 4 to 10 times greater than mean values between many other avian congeners. The Vireoni- dae is comprised of only four genera, and nearly 60% of living species are in Vireo. From a strictly taxonomic perspective, it might be argued that the total array of observed genetic diversity in the genus Vireo should approximate that observed within other avian families, such as the Muscieapidae, Emberi- zidae, or Parulidae, which are divided into many genera. To a first approxi- mation, this appears to be the case (Fig. 3). One hypothesis to account for the observed genetic distances within the genus Vireo is that species in the genus might be much older, on the average, than those in previously assayed avian genera. A few Pleistocene and recent fossils of vireonids are known (Romer, 1966), but these are of little or no use in estimating divergence times. However, several investigators have suggested, primarily from morphological considerations, that living vireonids are descendents of the stem group which gave rise to all "New World nine-primaried oscines" [Beecher, 1953; Bock, 1960; Tordoff, 1954; but see Raikow (1978) for conflicting opinion]. This latter group includes the Emberizidae, Parulidae, Icteridae, Fringillidae, and several other families in the classification of Morony et al. (1975; see Sibley, 1970). Thus Sibley (1970) writes, "there is general agreement among avian systematists today Genetic Distances in Birds 103 that the vireos are probably allied to the New World nine-primaried oscines, but not as closely as the members of the latter group are to one another. The egg-white data agree with this position." While these conclusions do not address directly the question of times of divergence of species within the Vireonidae, they are compatible with the possibility that some Vireo specia- tions are relatively ancient. Thus the morphological uniformity of Vireo (Hamilton, 1962), summarized in present taxonomy, may reflect the long- term retention of generalized insect-feeding adaptations (Bock, 1960) and may not evidence a very recent common ancestry. Whatever the explanation for the large D's in Vireo, the genus now appears to be an ideal candidate for larger-scale electrophoretic surveys involving additional species, since major evolutionary lineages within the group could likely be identified. The small electrophoretic distances else- where in Aves have inhibited attempts to unambiguously resolve species lineages within many genera. Genetic distances based on similar electrophoretic conditions and comparable suites of protein loci have also been reported for several well- known genera of Amphibia and Reptilia. In Figs. 1 and 2 we summarize these results in a fashion that permits direct visual comparison with observed distances in Mimidae and Vireonidae. The conservative pattern of protein differentiation in birds is again apparent. For example, all five assayed species of the Mimidae, belonging to three genera, cluster at a slightly lower level of genetic distance than do the two most closely related species of Plethodon examined by Duncan and Highton (1979); and clustering distances in some Anolis lizards (Wyles and Gorman, 1980) are more than five times as great as any clustering levels observed in Mimidae. Further- more, although several Vireo species exhibit atypically large D's in comparison with other birds, these distances remain modest or even small in comparison with those of many nonavian genera such as Hyla and Rana (Case, 1978; Etges, 1979; Fig. 2). We are not yet certain of the explanation(s) for the small genetic distances observed in birds. Several lines of evidence argue that protein evolution is decelerated in birds relative to most nonavian vertebrates (Avise et al., 1980c; Patton and Avise, 1982). Since the conservative pattern of electrophoretically assayed protein divergence now appears general to many avian families, hypotheses about the probable deceleration can begin to focus on biological characteristics prevalent in or unique to birds.

ACKNOWLEDGMENTS We remain indebted to the staff of Tall Timbers Research Station, and particularly to Robert Crawford, for collecting and preserving the WCTV tower-kill specimens used in our work. 104 Avise, Aquadro, and Patton

REFERENCES Avise, J. C., Patton, J. C., and Aquadro, C. F. (1980a). Evolutionary genetics of birds. I. Relationships among North American thrushes and allies. Auk 97:135. Avise, J. C., Patton, J. C., and Aquadro, C. F. (1980b). Evolutionary genetics of birds. II. Conservative protein evolution in North American sparrows and relatives. Syst. Zool. 29:323. Avise, J. C., Patton, J. C., and Aquadro, C. F. (1980c). Evolutionary genetics of birds. III. Comparative molecular evolution in New World warblers (Parulidae) and rodents (Criceti- nae). J. Hered. 71:303. Barrowclough, C. F., and Corbin, K. W. (1978). Genetic variation and differentiation in the Parulidae. Auk 95:691. Beecher, W. J. (1953). A phylogeny of the Oscines. Auk 70:270. Bock, W. J. (1960). The palatine process of the premaxilla in the Passeres. Mus. Comp. Zool. Bull. 122:361. Case, S. M. (1978). Biochemical systematics of members of the genus Rana native to western North America. Syst. Zool. 27:299. Case, S. M., Haneline, P. G., and Smith, M. F. (1975). Protein variation in several species of Hyla. Syst. Zool. 24:281. Duncan, R., and Highton, R. (1979). Genetic relationships of the eastern large Plethodon of the Ouachita mountains. Copeia 1979:95. Etges, W. J. (1979). Ecological Genetic Relationships in Selected Anurans of the Southeastern United States, M.S. thesis, University of Georgia, Athens. Gorman, G. C. and Renzi, J., Jr. (1979). Genetic distance and heterozygosity estimates in electrophoretic studies: Effects of sample size. Copeia 1979:242. Hamilton, T. H. (1962). Species relationships and adaptations for sympatry in the avian genus Vireo. Condor 64:40. Morony, J. J., Beck, W. J., and Farrand, J., Jr. (1975). Reference list of the birds of the world. Spec. Publ. Am. Mus. Nat. Hist., N.Y. Nei, M. (1972). Genetic distance between populations. Am. Nat. 106:283. Nei, M., and Roychoudhury, A. K. (1974). Sampling variances of heterozygosity and genetic distance. Genetics 76:379. Patton, J. C., and Arise, J. C. (1982). Evolutionary genetics of birds. IV. Rates of protein divergence in waterfowl (Anatidae) (submitted for publication). Patton, J. C., Baker, R. J., and Avise, J. C. (1981). Phenetic and cladistic analyses of biochemical evolution in peromyscine rodents. In Smith, M. H., and Joule, J. (eds.), Mammalian Population Genetics, University of Georgia Press, Athens, pp. 288-308. Prager, E. M., and Wilson, A. C. (1978). Construction of phylogenetic trees for protein and nucleic acids: Empirical evaluation of alternative matrix methods. J. Mol. Evol. 11:129. Raikow, R. J. (1978). Appendicular myology and relationships of the New World nine-primaried oscines. (Ayes: Passeriformes). Bull Carnegie Mus. Nat. Hist. 7. Romer, A. S. (1966). Vertebrate Paleontology, University of Chicago Press, Chicago. Sibley, C. G. (1970). A comparative study of the egg-white proteins of passerine birds. Peabody Mus. Nat. Hist. Bull. 32. Smith, J. K., and Zimmerman, E. G. (1976). Biochemical genetics and evolution of North American blackbirds, family Icteridae. Comp. Biochem. Physiol. 53B:319. Sneath, P. H. A., and Sokal, R. R. (1973). Numerical Taxonomy, W. H. Freeman, San Francisco. Tordoff, H. B. (1954). Relationships in the New World nine-primaried oscines. Auk 71:273. Wyles, J. S., and Gorman, G. C. (1980). The albumin immunological and Nei electrophoretic distance correlation: A calibration for the Saurian genus ,4nolis (Iquanidae). Copeia 1980:66.