Biochemical Genetics of Sunfish IV. Relationships of Centrarchid Genera Author(S): John C

Biochemical Genetics of Sunfish IV. Relationships of Centrarchid Genera Author(s): John C. Avise, Donald O. Straney, Michael H. Smith Source: Copeia, Vol. 1977, No. 2 (May 25, 1977), pp. 250-258 Published by: American Society of Ichthyologists and Herpetologists Stable URL: http://www.jstor.org/stable/1443906 Accessed: 25/11/2008 14:40

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=asih.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission.

JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We work with the scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that promotes the discovery and use of these resources. For more information about JSTOR, please contact [email protected].

American Society of Ichthyologists and Herpetologists is collaborating with JSTOR to digitize, preserve and extend access to Copeia.

http://www.jstor.org Biochemical Genetics of Sunfish IV. Relationships of Centrarchid Genera JOHNC. AVISE,DONALD O. STRANEYAND MICHAELH. SMITH We have examined electrophoretic variation in proteins encoded by 11-14 loci in species representing all nine genera of the Centrarchidae. A dendrogram based on allozyme information is compared to postulated relationships of sunfish genera based on general and specific morph- ologies, and on hybridizing propensity. The allozyme information correlates most strongly with that derived from a very detailed study of the acoustico-lateralis system by Branson and Moore (1962). Similarities between the two sets of data are observed 1) in the clustering together of species of Lepomis; 2) in the clustering of Lepomis with Micropterus; 3) in the placement of Acantharchus with Archoplites; and 4) in the very distant relationship of Elassoma to the other centrarchid genera. The levels of genetic similarity between centrarchid genera are compared to previously published levels of similarity between congeneric species of Lepomis, subspecies of Lepomis macrochirus, and geographic populations within the subspecies of L. m. macrochirus and L. m. purpurescens. Mean levels of genetic similarity (S) are as follows: between genera, S = 0.29; between congeneric species, S = 0.53; between subspecies of Lepomis macrochirus, S = 0.85; between geographic populations, S = 0.97.

THE Centrarchidae is a dominant component This arrangement is from Childers (1967). of the North American ichthyofauna. About Elassoma is not included; most authors agree 30 living species belonging to nine genera are that it is only distantly related to Centrarchidae, currently recognized. The earliest known fossils and perhaps should be placed in a distinct referable to Centrarchidae have been found in family (Branson and Moore, 1962). Eocene deposits (Romer, 1966), and some fossils In this study, we continue our survey of assigned to living centrarchid genera may be as genetic variability and differentiation among old as the Miocene (Miller, 1965). Centrarchids members of Centrarchidae. Previous work has are indigenous to North America and probably described geographic variation in Lepomis evolved from serranid (sea bass) ancestors early macrochirus (Avise and Smith, 1974a), and dif- in the Cenozoic. ferentiation among species of Lepomis, the most Attempts to determine phylogenetic relation- species-rich of centrarchid genera (Avise and ships among genera of sunfish have been based Smith, 1974b). Here we examine electrophoret- on morphological counts and measurements ically levels of genetic differentiation among (Schlaikjer, 1937; Bailey, 1938), the acoustico- species representing all nine sunfish genera. lateralis system (Branson and Moore, 1962), Results will be compared 1) to results of previous success of hybridization (Hester, 1970; West, morphological and hybridization studies, and 1970) and chromosome cytology (Roberts, 1964). 2) to levels of genic differentiation between As discussed later, results of these studies are congeneric species, and between conspecific far from uniform. a Nonetheless, general pic- populations. ture of the possible relationships among centrarchid genera has emerged, and is reflected MATERIALS AND METHODS in the current classification: Fish were frozen at time of and trans- Centrarchidae capture ported to the laboratory where routine pro- Centrarchinae Lepominae Enneacanthini cedures of horizontal starch gel electrophoresis Ambloplitini were to variation in Archoplites Enneacanthus employed assay specific are those Ambloplites Lepomini proteins. Electrophoretic techniques al. modified for fish Acantharchus Lepomis of Selander et (1971) as by Centrarchini Micropterini Avise and Smith (1974a,b). The common allele Pomoxis Micropterus at each locus in the bluegill (Lepomis macro- Centrarchus chirus) was arbitrarily assigned a value of 100. 250 AVISE ET AL.-SUNFISH BIOCHEMICAL GENETICS 251

TABLE 1. SAMPLESOF SPECIESBELONGING TO NINE GENERAOF NORTHAMERICAN CENTRARCHIDAE.

Mean # in- hetero- Sample Genus Species Common name Collection sitel dividuals zygosities 1 Lepomis (10 spp.) sunfish See Avise & Smith (1974b) 1099 5.9 2 Pomoxis nigromaculatus black crappie Pond near Ellenton, S.C. 18 1.0 3 Enneacanthus obesus banded sunfish Creek near Ellenton, S.C. 27 0.0 4 Micropterus salmoides largemouth bass Pond near Ellenton, S.C. 18 8.3 5 Elassoma okefenokee pygmy sunfish Baker Co., Fla. 20 1.4 6 Elassoma evergladei pygmy sunfish Baker Co., Fla. 12 0.0 7 Centrarchus macropterus flier Swamp near Ellenton, S.C. 4 0.0 8 Acantharchus pomotis mud sunfish River Styx, Fla. 5 2.9 9 Ambloplites rupestris rockbass Central Tennessee 12 13.0 10 Archoplitesinterruptus Sacramentoperch Bishop Lake, Calif. 39 0.4 1More detailed information available on request.

Other alleles were designated according to from members of all other genera, we have proportional mobility when run along side the simply included in the dendrogram the single 100 allele and beside other alleles of known category "Lepomis"; the point at which this relative mobility. category joins the remainder of the dendro- The proteins examined in this study were gram is determined by the mean level of genetic chosen solely on the basis of availability of affinity of all Lepomis species to the other suitable staining procedures and clarity of band- genera. ing. Additional proteins banded well in some A question arises whether the species sampled species but not in others. We have included in adequately represent each of the sunfish genera. this paper only those loci which could be re- The most species-rich genus Lepomis is well liably scored in most or all species of Centrarchi- represented by 10 of its 11 species. Acantharchus, dae. The set of loci examined was chosen with- Centrarchus and Archoplites are monotypic out bias with respect to level of differentiation genera, and Pomoxis and Ambloplites each con- between species. tain only two living species. We have sampled Heterozygosities (mean percentages of indi- one of three species of Enneacanthus, and two viduals heterozygous per locus) were determined of the three species of Elassoma. We have sam- from allele frequencies using Hardy-Weinberg pled only one of six living Micropterus species, probabilities. Genetic similarities (S) between although a large-scale study of other species is species were calculated according to Rogers' planned. Since the possibility exists that Mi- (1972) formula. Rogers' similarity coefficient cropterus salmoides does not adequately repre- may assume values from zero to one, with one sent all Micropterus species for purposes of indicating genetic identity. Genetic distance is intergeneric comparisons, it is best to consider D = 1-S. From the genetic similarity matrix, the relationships of Micropterus presented in cluster analysis was performed using the un- this paper to strictly apply only to Micropterus weighted pair-group method with arithmetic salmoides. Similar reservations apply to the means (Sokal and Sneath, 1963). other genera not exhaustively sampled. The species examined for the first time in this We also admit that the number of loci which study are listed in Table 1. We have sampled could be reliably scored in most or all species one representative species from each of seven is regrettably small. To our knowledge, the 19 genera, two species of Elassoma and ten species species studied here represent the largest num- of Lepomis. The genetic relationships among ber of species of any taxon surveyed electro- Lepomis have previously been described (Avise phoretically (at a number of loci) to date. As and Smith, 1974b). In order to avoid unneces- more species are added to a survey, the diffi- sary repetition, mean values for these Lepomis culty of reliably scoring a large number of loci species are used here wherever possible and increases substantially. Because we have scored where they do not distort the intended meaning. relatively few loci, we prefer not to advocate For example, since all Lepomis species cluster changes in the current taxonomy of centrarchid together in a biochemical dendrogram, distinct genera based on the new biochemical data. TABLE 2. COMMON ALLELES (FREQUENCY ~> 0.05) AT 14 LOCI IN SPECIES OF CENTRARCHIDAE.1

Locus

Species Ldh-l Ldh.2 Idh-1 6Pgd-l Es-1 Es-3 Mdh-l Got-1 Got-2 Ipo-1 Pgn-1 Pgi-1 Pgi-2 Pep-2 Lepomis 100(0.71) 137(0.87) humilis 100 100 A 100 96 93 100 78(0.29) 100 122 110 96 103(0.13) 123 Lepomis 122(0.58) 90(0.70) microlophus 100 100 A 87 96 81 102 100 109 96(0.42) 110 81(0.30) 100 123 Lepomis 86(0.61) 110(0.91) 95(0.80) 62(0.33) 83(0.94) auritus 100 100 C 100 96 88(0.38) 102 100 58 100 115(0.08) 89(0.19) 100(0.67) 88(0.06) Lepomis 100(0.70) gulosus 100 100 A 100 100 63 100 100 58 67(0.30) 110 96 198 100 Lepomis 104(0.63) 110(0.95) 100(0.82) n 96(0.60) 100(0.90) 0o megalotis 100 100 A 87 100(0.37) 100(0.32) 102 114(0.07) 58 74 100(0.05) 95 137(0.18) 115 97(0.07) Lepomis 88(0.50) 110(0.70) 98(0.86) cyanellus 100 100 A 86(0.50) 100 80 102 100 58 122 100(0.30) 95(0.14) 130 83 Lepomis 88(0.52) 110(0.78) 95(0.83) punctatus 100 100 A 100 96 86(0.48) 102 100 58 122 93(0.22) 90(0.17) 100 88 Lepomis 65(0.78) 95(0.65) - gibbosus 100 100 A 100 96 87 102 82(0.22) 58 100 110 90(0.35) 100 115 Lepomis 94(0.46) 58(0.79) marginatus 100 100 A 87 1100 83(0.39) 102 100 24(0.22) 74 110 95 100 95 88(0.15) Lepomis 100(0.50) 100(0.95) 100(0.64) macrochirus 100 100 A 100 100 96(0.40) 100 100 58(0.05) 100 110(0.26) 100 100 100 92(0.10) 90(0.10) Pomoxis 96(0.94) nigromaculatus 150 100 D 100 - - 102 100 70 - 125 92(0.06) 137 85 Enneacanthus obesus 200 100 E 100 - 102 100 100 115 102 140 124 Micropterus 102(0.61) 114(0.69) salmoides 100 100 E 100 - 150(0.39) 100(0.31) 58 - 110 100 140 90 AVISE ET AL.-SUNFISH BIOCHEMICAL GENETICS 253

04a Nonetheless, as described below, the results from Ca the allozyme survey do agree well with classical es systematic information, in broad In 0 o in 0 0 perspective. a') -4 ', alJ addition, they provide quantitative estimates of c oc, the of loci which 04100 00 o00 proportion genetic typically 00 -( 04 differ in allelic between centrarchid o t- 0000 composition

= 0- 00 0 0o O ifC O genera. kn nc __) r _ _. I RESULTS 0O0 Proteins encoded by a total of 14 0 C- 0 0 0C genetic C 00 0 0o loci were scored in all species of Lepomis. '1.4.-^ Eleven loci were scored in each of the other cn t- 00oo Common alleles at these loci are od species. listed in l') fn Lff ) I4f i)0 Table 2. The proteins examined represent _ _ _ a wide array of functional types, including de- hydrogenases, oxidases, esterases, transaminases, mutases, isomerases, and peptidases. The proteins scored and the abbreviations of the loci 0 I I I I I o - g which encode them are as follows: lactate de-

OCl hydrogenases (Ldh-1 and Ldh-2); isocitrate de- 0 o N hydrogenase (Idh-1); 6-phosphogluconate dehy- O 0 o o c: O 4 drogenase esterases and 1o 1o 30 (6-Pgd-l); (Es-l Es-3); malate dehydrogenase (Mdh-1); glutamate oxa- > late transaminases (Got-l and Got-2); indo- a) o3 oxidase if) if) t- phenol (Ipo-1); phosphoglucomutase 0 It 0 if S: C C0 - C O O (Pgm-1); phosphoglucose isomerases (Pgi-1 and 0 0 Pgi-2); peptidase (Pep-2). Q bii Be Most of these enzymes exhibit banding pat- '5! terns on gels which are similar to those ob- Cs 0 00 0 0 0- -0 served in other fishes where the genetic basis 54 has been documented through breeding studies "-4 (May, 1975; Allendorf, 1975). In addition, the I I I I I I codominant inheritance of alleles at eight loci (Es-3, Mdh-1, Got-2, Ipo-1, Pgm-1, Pgi-1, 6Pgd-l a ) and Pep-2) has been confirmed by the appear- c.0 ance of appropriate banding patterns in F1 I I I I I Q..0l-ca r hybrids between macrochirus and L. o f Lepomis microlophus (Avise and Smith, 1974b). At other 6 C3 polymorphic loci where sample sizes were ade- o 0 quate to permit tests, observed fre- 04 00 00 0 0 genotype 00 00 00 0000o 0 were close -. quencies to those predicted on the basis on Hardy-Weinberg equilibrium, further increasing confidence in our correct interpretation -4a a of zymogram patterns. One isocitrate did ex- 0 0 0 0 0 o a. enzyme, dehydrogenase, hibit unusual banding patterns in several species. Individuals of most species uniformly exhibited a 0CM CM0 CT0ff) oO C 0 0o single band, presumably representing C O - - 0 - - the product of alleles at a single monomorphic locus. However, five species (Lepomis auritus, aQ)C Pomoxis nigromaculatus, Ennaecanthus obesus, 'a< Micropterus salmoides and Ambloplites rupes- JJlQ i tris) showed various 2-banded phenotypes, both bands intense All members s^s-jo^s.ss^i-s.fc ? equally (Fig. 1). within a had identical ~^E |.|l|lilt?li^ P 1 J given species patterns, with no apparent variants. Isocitrate dehydro- 254 COPEIA, 1977, NO. 2

41,~~~~~~~ An~Allelefrequencies were used to calculate mea- 4*(i @ sures of relatedness (genetic similarity and ge- IDH netic distance) between each pair of the 19 sunfish examined (Table 3). Genic similarities between members of different genera cover a broad spectrum, ranging from a low of 0.09 in I^B_~~_ mmmthe comparison between Pomoxis nigromacula- mm^_ _ = ^tus and Elassoma evergladei to a high of 0.53 in the comparison between Enneacanthus obesus and Mean level of t U Ambloplites rupestris. ge- A B CD E~ netic similarity (Rogers' coefficient) between PHENOTYPE all 36 possible pairwise intergeneric comparisons equals 0.29. Genic similarities between genera Fig. 1. Zymogrampatterns of isocitrate dehydro- of centrarchids have not previously been re- genase in 19 species of North American Centrarchi- ported. However, data have dae. Most exhibited either A or electrophoretic species phenotype been collected for a few species, primarily by B. Phenotype C was observed in Lepomis auritus; phenotype D in Pomoxis nigromaculatus; and Whitt and associates. Probably the most com- phenotype E in Enneacanthus obesus, Micropterus plete set of data includes information on 14 salmoidesand Ambloplitesrupestris. genetic loci in Micropterus salmoides and Lepomis cyanellus (Whitt et al., 1973). Using their data, we have calculated genetic similarity genase is a dimeric molecule in most animals, using Rogers' coefficient; it equals 0.43. Based including fishes. If two loci (perhaps arising on our own data, genetic similarity between a are involved in through gene duplication) de- these two species equals 0.42. the two banded in termining zymogram these An expanded version of the matrix of ge- five the of the two loci do not species, products netic similarities in Table 3 was employed to form a heterodimer band of inter- expected form a dendrogram by cluster analysis (Fig. 2). mediate The lack of active en- mobility. hybrid The estimates of the genic relationships among molecules between of different zyme products centrarchid genera may be compared to previ- loci has however been in fishes of Idh reported ously inferred relationships based on other cri- the Since genus Xiphophorus (Siciliano, 1973). teria. In order to make these comparisons as we cannot determine the exact unambiguously objective as possible, several procedures have number of Idh loci in sunfishes with the present been used. First, we have arbitrarily assigned and since the do indicate at data, zymograms similarity values to the various levels of the least some differences between in genetic Idh "formal classification" of Centrarchidae as pre- different for of species, purposes interspecies sented in the introduction. The assigned sim- we have comparisons conservatively scored each ilarity values were chosen to cover roughly the distinct Idh phenotype (Fig. 1) as the product same range of values as had been obtained from of a different allele at a locus single genetic the genic evidence-that is, from about 0.10 to The level of at 11 loci genic heterozygosity in 0.55. The simplest and most reasonable method several of sunfish is consider- species (Table 1) of assignment is to equally divide the range of lower than the mean of about ably 6% reported values to be assumed by the number of recog- for all vertebrates examined and (Selander nized levels of the taxonomic hierarchy. Assign- No was observed Kaufman, 1973). variability ments were made as follows: 1) species belong- in Enneacanthus Elassoma or obesus, evergladei ing to the same tribe, 0.55; 2) species belonging Centrarchus these re- macropterus. However, to different tribes in the same subfamily, 0.40; sults should be with extreme cau- interpreted belonging to different subfamilies, tion. First, only~~~~~~~3)tion.belongngly small samplesFirst, wdfferewere obtainedobnt subfamilies, 0.25; 4) species of Elassoma versusversus other cen- of some species such as Centrarchus. More im-02 4s e portantly, the number of loci examined is not trarchds The va are listed below large, and some classes of enzymes which are the diagonal in Table 3. known to be particularly variable could not The most thorough published study of cen- be reliably scored in all genera of Centrarchi- trarchid phylogeny is based upon a very detailed dae, and hence are not included in the hetero- analysis of components of the acoustico-lateralis zygosity estimates. Among Lepomis species, for system (Branson and Moore, 1962). From their which 14 loci including two esterases were hypothetical dendrograph, we have constructed scored, mean heterozygosity was 5.9%. a dendrogram from which similarity estimates AVISE ET AL.-SUNFISH BIOCHEMICAL GENETICS 255

TABLE 3. LEVELS OF SIMILARITYBETWEEN CENTRARCHIDGENERA.1

1 2 3 4 5 6 7 8 9 10

1. Lepomis (mean of 10 species) 0.32 0.30 0.45 0.21 0.22 0.18 0.30 0.31 0.33 (0.68) (0.70) (0.55) (0.79) (0.78) (0.82) (0.70) (0.69) (0.67) 2. Pornoxis nigronaculatus 0.25 0.37 0.36 0.11 0.09 0.18 0.19 0.25 0.37 (0.25) (0.63) (0.64) (0.89) (0.91) (0.82) (0.81) (0.75) (0.63) 3. Enneacanthus obesus 0.40 0.25 0.36 0.19 0.18 0.45 0.37 0.53 0.46 (0.35) (0.25) (0.64) (0.81) (0.82) (0.55) (0.63) (0.47) (0.54) 4. Micropterus salmoides 0.40 0.25 0.40 0.12 0.11 0.21 0.30 0.40 0.34 (0.40) (0.25) (0.35) (0.88) (0.89) (0.79) (0.70) (0.60) (0.66) 5. Elassoma okefenokee 0.10 0.10 0.10 0.10 0.83 0.26 0.29 0.20 0.28 (0.10) (0.10) (0.10) (0.10) (0.17) (0.74) (0.71) (0.80) (0.72) 6. Elassoma evergladei 0.10 0.10 0.10 0.10 0.70 0.18 0.28 0.19 0.27 (0.10) (0.10) (0.10) (0.10) (0.70) (0.82) (0.72) (0.81) (0.73) 7. Centrarchus macropterus 0.25 0.55 0.25 0.25 0.10 0.10 0.28 0.41 0.27 (0.25) (0.30) (0.25) (0.25) (0.10) (0.10) (0.72) (0.59) (0.73) 8. Acantharchus pomotis 0.25 0.40 0.25 0.25 0.10 0.10 0.40 0.27 0.37 (0.20) (0.20) (0.20) (0.20) (0.10) (0.10) (0.20) (0.73) (0.63) 9. Ambloplites rupestris 0.25 0.40 0.25 0.25 0.10 0.10 0.40 0.55 0.28 (0.30) (0.25) (0.30) (0.30) (0.10) (0.10) (0.25) (0.20) (0.72) 10. Archoplites interruptus 0.25 0.40 0.25 0.25 0.10 0.10 0.40 0.55 0.55 (0.15) (0.15) (0.15) (0.15) (0.10) (0.10) (0.15) (0.15) (0.15)

1 Above diagonal: allozyme-based genic similarities (distances) calculated from allele frequencies at 11-14 genetic loci, using Rogers' (1972) formulas. Below diagonal: similarities obtained from the "formal classification" of Cen- trarchidae as described in text; and in parenthesis, similarities obtained from the dendrogram in Fig. 3 based on Bran- son and Moore's (1962) study of the acoustico-lateralis system. may be obtained (Fig. 3). Again, the range of primitive centrarchid stock or else have com- similarities was chosen to roughly correspond to pletely different affinities. Branson and Moore those based on the genic and "formal classifica- (1962) argue that Elassoma should be raised to tion" estimates, and the range was divided family status. equally by the number of branch points in the In the biochemical dendrogram (Fig. 2), Elas- dendrogram. Estimates of similarity taken from soma appears to be the most distinct of the this dendrogram have been entered in Table 3, centrarchid genera, joining the other genera at and form another basis of comparison with the S = 0.22. It is conceivable that this low level of genic data. Of course we realize that the sim- genic similarity is the result of chance identities ilarity estimates summarized for the "formal in mobility of a few proteins, or even to conver- classification" and Branson and Moore data gence to a common allelic state at a few loci. Or are imperfect, and may be criticized to detail. perhaps the loci at which alleles appear to be However, we are interested in main effects, and feel that attempts to objectively quantify information from various sources are necessary for LEPOMIS comparative purposes. MICROPTERUS In this section, we will describe the results POMOXIS ENNEACANTHUS of the genic information in relation to previ- AMBLOPLITES CENTRARCHUS ously published accounts of the probable rela- ACANTHARCHUS of centrarchid [L ARCHOPLITES tionships genera. okefenokeeE ELASSOMA 1, I, 1, 1 , 1, 1, 1, I, I, 1, Systematic relationships of Elassoma.-Jordan 0 0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1.0 (1877) first described Elassoma and believed it GENETIC SIMILARITY to be a member of the Cichlidae. Berg (1946) considered Elassoma to belong to Centrarchidae, Fig. 2. Allozyme-based dendrogram of centrar- where it is Most workers feel chid genera based on unweighted pair-group cluster presently placed. analysis of genic similarity coefficients derived from that elassomids either diverged very early from 11-14 genetic loci. 256 COPEIA, 1977, NO. 2 shared with some other species (Ldh-1, Ldh-2, and Moore (1962). Nonetheless, Branson and Idh-1) are more slowly evolving. We cannot Moore note that the relationship at best is an distinguish between these possibilities with the obscure one; again the retention of supposedly present data. It would be interesting to ex- primitive characters rather than recent shared amine very distantly related fish (such as those ancestry may explain its taxonomic affinity to belonging to different orders) to establish Acantharchus. Acantharchus and Ambloplites whether shared electrophoretic mobilities due show S = 0.27. to chance and convergence are common phe- The tribe Centrarchini contains Poxomis and nomena. At any rate, by all criteria Elassoma Centrarchus. There is general agreement among is the most distinct of centrarchid genera. most workers that these genera are more closely Three living species are currently assigned to related to one another than to other centrar- Elassoma, and we have examined two of them. chid genera. This conclusion is not apparent Elassoma okefenokee and E. evergladei are geni- from the genic data; Pomoxis and Centrarchus cally very similar to one another, S = 0.82. share alleles at only two of eleven loci, Ldh-2 Nonetheless they are completely distinct in al- and Mdh-l, and overall biochemical similarity lelic composition at Pgi-1 and Pgi-2. These two is only S = 0.18. species were obtained from the same collecting site. Since two distinct are Relationships within the Lepominae.-Three gene pools repre- tribes are included in this sub- sented, the species status of these forms is con- monogeneric the Enneacanthini, and Mi- firmed, despite their overall biochemical affinity. family, Lepomini These two about as similar cropterini. The banded sunfish genus Ennea- species appear geni- canthus is to be the most cally as do the two most closely related species of generally thought and L. S distinct member of this subfamily, with Lepomis Lepomis (L. marginatus megalotis, the former and MIi- 0.79-Avise and Smith, 1974b). (including Chaenobryttus) cropterus more closely related. Relationships within the Centrarchinae.-This Lepomis is the most species-rich of centrarchid subfamily contains two tribes, Ambloplitini and genera, and contains eleven living species. The Centrarchini. In Ambloplitini are currently warmouth, formerly Chaenobryttus gulosus, has included Archoplites, Ambloplites and Acan- recently been placed in the genus Lepomis tharchus. Acantharchus and Archoplites cluster (Bailey et al., 1970) on the basis of overall simi- together in the biochemical dendrogram, not- larity and especially its ready ability to hybridize withstanding the fact that they are genically with other lepomine forms (Hubbs, 1955; West, quite different from one another, S = 0.37. 1970; West and Hester, 1966; Birdsong and However, Ambloplites appears genetically some- Yerger, 1967). Hybrids between L. gulosus and what closer to Enneacanthus and not par- either L. cyanellus or L. macrochirus are at least ticularly close to the other Ambloplitini. partially fertile (Childers, 1967; West, 1970). All Archoplites, the Sacramento perch, is the only ten Lepomis species examined cluster together living sunfish native to waters west of the in a biochemical dendrogram, distinct from all rockies. Since sunfish are characteristic of low- other genera (Avise and Smith, 1974b). land waters, Archoplites was probably separated Biochemically, Micropterus clusters closest to from other centrarchids not later than the for- Lepomis, even though these genera are com- mation of the Rocky Mountains, which began pletely distinct in allelic composition at more during the Miocene or early Pliocene (Schuchert than 50% of their loci (S = 0.45). Micropterus and Dunbar, 1941). Branson and Moore (1962) and Lepomis are thought to be fairly closely argue that the Sacramento perch was early iso- related on other grounds as well; in particular, lated from other sunfish and "has become Lepomis gulosus X Micropterus salmoides hy- evolutionarily stagnated." Its placement in brids are completely viable although sterile Ambloplitini may thus reflect its apparent re- (West, 1970). tention, with Ambloplites and Acantharchus, of DISCUSSION certain traits considered characteristic of prim- In order to the corre- itive centrarchid stock, and may not necessarily objectively quantify between reflect a recent shared ancestry. The biochemi- spondence postulated relationships of sunfish cal evidence does not contradict this hypothesis. genera based on different criteria, we Ambloplites is placed near Acantharchus and have calculated correlation coefficients between Micropterus by Schlaikjer (1937), near Acan- similarity estimates obtained from (A) genic in- tharchus, Centrarchus and Pomoxis by Bailey formation, (B) the acoustico-lateralis system, (1938), and nearest Acantharchus by Branson (C) the "formal classification," and (D) general AVISE ET AL.-SUNFISH BIOCHEMICAL GENETICS 257

TABLE4. CORRELATIONCOEFFICIENTS BETWEEN SIMI- LEPOMIS LARITYESTIMATES OF CENTRARCHIDGENERA.1 MICROPTERUS ENNEACANTHUS A B C D ___ AMBLOPLITES A - 0.45 0.22 0.23 CENTRARCHUS - - B 0.46 0.05 I-~ ' POMOXIS C - 0.05 ACANTHARCHUS D ARCHOPLITES 1 Similarity estimates based on (A) genic data from the ELASSOMA present study, (B) the acoustico-lateralis system (Branson and Moore, 1962), (C) the "formal classification" of Centrarchidae, and (D) an early morphological study O.I 0.2 0.3 0.4 0.5 (Schaikjer, 1937). Similarity values used to generate SIMILARITY several of the correlations are given in Table 3. Fig. 3. Dendrogram of centrarchid genera based on a detailed study of the acoustico-lateralis system morphology. These correlations are presented (from Fig. 11 of Branson and Moore, 1962). in Table 4. The correlations between various sets of information are in all cases positive, but not terus with Lepomis, 3) the placement of Acan- strong. The highest correlation is that between tharchus with Archoplites and 4) the distant Branson and Moore's (1962) study of the lateral relationship of Elassoma to the other sunfish line system and the "formal classification." This genera. The correlation between the formal is not too surprising since, in fact, the formal classification and the allozyme data is somewhat classification in most cases follows the sugges- lower, rAc = 0.22. Nonetheless, this correlation tions of Branson and Moore. What is surprising is still considerably stronger than that calculated is that this correlation is not higher. There between another set of morphological data are at least two possible reasons for this. 1) As (taken from Schlaikjer, 1937) and either the already mentioned, it appears that some genera acoustico-lateralis system (rBD = 0.05) or the are placed in the same tribe or subfamily formal classification (rCD = 0.05). primarily on the basis of sharing particular The mean level of genic similarity between traits, for example primitive traits, rather than the nine centrarchid genera is low, S = 0.29. In on the basis of a relatively recent shared another words, on the average, species belonging cestry. Yet the dendrogram (Fig. 3) used to to different genera appear similar in allelic estimate similarities between genera from Bran- composition at only about 30% of their loci. son and Moore's work is based on the assump- And this is perhaps a conservative guess, since tion that similarity is primarily a (negative) electrophoretic methods consistently show a bias function of time since divergence from a com- towards underestimating the differences be- mon ancestor. Of course, the lack of close tween species (Avise, 1974). Clearly a very large correspondence between phenetic similarity and number of genic differences have accumulated time since divergence is a commonly reported during the independent evolutionary histories phenomenon and forms the heart of the prob- of these fishes. Miller (1965) postulates that at lem of determining a truly phylogenetic system- least five living centrarchid genera had already atics. 2) It is likely that the similarity coeffi- made their appearance in the Miocene or at the cients presented in Table 3 distort somewhat Miocene-Pliocene boundary (Pomoxis, Microp- the results of various studies, due to the neces- terus, Lepomis, Chaenobryttus and Amblop- sarily artificial and simplistic manner in which lites). Thus the considerable genetic divergence they were calculated. However, this potential between a number of genera may have accumu- bias should not favor any particular correlations lated over a period of at least 20 million years. over others, but likely lowers them all. If this is How do genetic distances between centrarchid true, the relative levels of the correlation co- genera compare to distances representing more efficients are of more significance than their recent stages of evolutionary divergence in the absolute levels. Centrarchidae? Four levels of comparison are The allozyme based information correlates now available: 1) between at least twelve geo- most strongly with information from the graphic populations within each of two bluegill acoustico-lateralis system. Similarities are most subspecies, Lepomis macrochirus macrochirus apparent in 1) the clustering together of all and L. m. purpurescens (Avise and Smith, Lepomis species, 2) the clustering of Microp- 1974a); 2) between these two bluegill subspecies 258 COPEIA, 1977, NO. 2

(Avise and Smith, 1974a); 3) between ten spe- HESTER,F. E. 1970. Phylogenetic relationships of cies within the and Smith, sunfishes as demonstrated by hybridization. Trans. genus Lepomis (Avise Amer. Fish. Soc. 1970:100-104. 1974b); and 4) between nine genera (present H C L. 1955. Hybridization between fish study). Mean levels of genetic similarity and species in nature. Syst. Zool. 4:1-20. distance (Rogers' coefficients) at each of these JORDAN,D. S. 1877. Contributions to North Amer- levels are as follows: ican ichthyology based primarily upon the collec- tions of the U.S. National Museum. Bull. U.S. S ~~g-D Natl. Mus. 1877:50-51. May, B. 1975. Electrophoretic variation in the 1) geographic populations 0.97 0.03 genus Oncorhynchus: the methodology, genetic 2) subspecies 0.85 0.15 basis, and practical applications to fisheries research and management. Unpubl. Masters Thesis, 3) species 0.53 0.47 Univ. of Wash. 4) genera 0.29 0.71 MILLER, R. R. 1965. Quaternary freshwater fishes of North America, p. 569-581. In: The Quater- These values cover a very large range, and nary of the United States. H. E. Wright and D. in broad perspective are closely correlated with G. Frey (eds.) Princeton Univ. Press, Princeton, in broad perspective are closely correlated with NewNe JerseyJersey. levels of evolutionary divergence. ROBERTS, F. L. 1964. A chromosome study of twenty species of Centrarchidae. J. Morph. 115: ACKNOWLEDGMENTS 401-418. ROGERS,J. S. 1972. Measures of genetic similarity This work was supported by AEC grant AT and genetic distance. Studies in Genetics VII. (38-1)-310. Univ. Texas Pub. 7213:145-153. ROMER,A. S. 1966. Vertebrate paleontology. Univ. Chicago Press, Chicago. LITERATURECITED SCHLAIKJER,E. M. 1937. New fishes from the ALLENDORF,F. W. 1975. Genetic in a continental Tertiary of Alaska. Bull. Amer. Mus. variability Natur. Hist. 741-23. species possessing extensive gene duplication: genetic interpretation of duplicate loci and ex- SCHUCHERT,C., AND C. DUNBAR. 1941 A text- amination of genetic variation in populations of book of geology. Part II-Historical geology. John rainbow trout. Unpubl. Ph.D. thesis, Univ. of Wiley and Sons, New York. Wash. SELANDER,R. K., ANDD. W. KAUFMAN. 1973. Genic AvISE, J C. 1974. Systematic value of electropho- variability and strategies of adaptation in animals. Proc. Nat. retic data. Syst. Zool. 23465-481. Acad. Sci. 70:1875-1877. ANDM. H. SMITH. 1974a. Biochemical ge- , M. H. SMITH,S. Y. YANG,W. E. JOHNSON netics of sunfish. I. Geographic variation and ANDJ. B. GENTRY. 1971. Biochemical polymor- subspecific intergradation in the bluegill, Lepomis phism and systematics in the genus Peromyscus. macrochirus. Evolution 23:42-56. I. Variation in the old-field mouse (Peromyscus m 1974b. Biochemical of sunfish. polionotus). Studies in Geneticsntc VI.LuvTea Univ. Texas ? Biochemical genetics of sunfish. Pub. 7103:49-90. II. Genic between similarity hybridizing species. ILIAN, M. J. 1973. Evidence for multiple un- Amer. Nat. 108:458-472.108:458-472SICILIANO, M. J. 1973. Evidence for multiple un- BAILEY R M. 1938.1 . A systematc revisionr n of thete linked genetic loci for isocitrate dehydrogenase in BAILEY.h systematic fish of the genus Xiphophorus. Copeia 1973:158- centrarchid fishes, with a discussion of their dis- 161 and tribution, variations, probable interrelation- SOKALR. R. ANDP. H A. SNEATH. 1963 Princi- Ph.D. ships. Unpubl. dissertation, Univ. of Mich- ples of numerical taxonomy. W. H. Freeman, San igan. Francisco. E. E. S. E. A. J. FITCH, HERALD, LACHNER, WEST, J. L. 1970. The gonads and reproduction of C. C. LINDSEY, C. R. ROBINSAND W. B. SCOTT. three intergeneric sunfish (family Centrarchidae) 1970. A list of common and scientific names of hybrids. Evolution 24:378-394. fishes from the United States and Canada. Amer.. HESTER. 1966. Intergeneric Fish.Fish. Soc. Serial Pub 6.6 , AND F. E. HESTER. 1966. Intergeneric S.erial Pub.*Casfcto ofhybridization of centrarchids. Trans. Amer. Fish. BERG, L. S. 1947. Classification of fishes both Soc. 95:280-288. Edwards recent and fossil. Brothers, Inc., Ann WHITT, G. S., W. F. CHILDERSAND P. L. CHO. 1973. Arbor, Michigan. Allelic expression at enzyme loci in an intertribal R. AND R. W. BIRDSONG, S., YERGER.1967. A nat- hybrid sunfish. Journ. of Heredity 64:54-61. ural population of hybrid sunfishes: Lepomis macrochirus X Chaenobryttus gulosus. Copeia 1967:62-71.DEPARTMENT OF ZOOLOGY,UVIVERSITY OF GEOR- BRANSON,B. A., AND G. A. MOORE. 1962. The GIA, ATHENS, GEORGIA30602, DEPARTMENTOF lateralis components of the acoustico-lateralis sys- ZOOLOGY,UNIVERSITY OF CALIFORNIA,BERKE- tem in the sunfish family Centrarchidae. Copeia LEY, CALIFORNIA 94720, SAVANNAH RIVER ~~~1962:1-108. ~ECOLOGY LABORATORY,LAORATO DRAWER E, AIKEN, CHILDERS,W. F. 1967. Hybridization of four spe- OLO AIK cies of sunfishes (Centrarchidae). Bull. Illinois SOUTH CAROLINA 29801. Accepted 5 May Natur. Hist. Surv. 29:159-214. 1976.