Evolution of Intrinsic Post-Zygotic Reproductive Isolation in Fish

Ann. Zool. Fennici 40: 321–329 ISSN 0003-455X Helsinki 29 August 2003 © Finnish Zoological and Botanical Publishing Board 2003

Evolution of intrinsic post-zygotic reproductive isolation in ﬁ sh

Stephen T. Russell

Gatty Marine Laboratory, School of Biology, University of St. Andrews, St. Andrews, Scotland, KY16 8LB (e-mail: [email protected])

Received 5 Feb. 2003, revised version received 3 June 2003, accepted 4 June 2003

Russell, S. T. 2003: Evolution of intrinsic post-zygotic reproductive isolation in ﬁ sh. — Ann. Zool. Fennici 40: 321–329.

The comparative approach is important in investigating the evolution of reproductive isolation. Yet this method has not been extended to teleost fi sh, the most diverse vertebrate lineage. Many experimental interspecifi c crosses have been performed in aquacultural research, and some of the resultant data is analysed in relation to mitochondrial cytochrome b divergence to investigate the evolution of intrinsic postzygotic isolation. Such isolation was found to increase gradually, suggesting that hybrid unfi tness is usually due to the gradual accumulation of deleterious epistatic interactions among species. Hybrid sterility evolved more rapidly than inviability, although both forms of hybrid fi tness are probably important in isolating natural populations. Because of general diffi culties in characterising fi sh genetic sex-determination systems, and because of the plethora of systems potentially represented in any given cross, too few crosses are currently available to evaluate the generality of Haldaneʼs rule in fi sh.

Introduction emerge from these analyses. Hybrid sterility tends to evolve sooner than hybrid inviability Speciation is a typically slow process whose in taxa as divergent as Lepidoptera (Presgraves genetical basis requires substantial effort to elu- 2002), Drosophila (Coyne & Orr 1989, 1997) cidate. Furthermore, taxa differ in their modes and birds (Price & Bouvier 2002). This is prob- and rates of speciation (Mayr 1942). Thus, com- ably because sex- and reproduction-related parative approaches are an important means of genes evolve more rapidly than viability-related identifying general phenomena characterising ones (Singh & Kulathinal 2000). Also evident speciation (e.g., Coyne & Orr 1989, 1997), and is Haldaneʼs rule: “when, in the F1 offspring of potentially of discerning the relative importance two different animal races, one sex is absent, of different speciation mechanisms (e.g., Pres- rare, or sterile, that sex is the heterozygous [het- graves 2002). erogametic] sex (Haldane 1922).” Haldaneʼs rule Comparative analyses considering the evo- is very widespread and has been documented lution of intrinsic post-zygotic reproductive in both groups with heterogametic males (e.g., isolation (RI) relate the magnitude of RI (as Mammalia, Amphibia, Reptilia), and those with assayed by hybrid sterility and inviability) to heterogametic females (Diptera, Orthoptera, genetic divergence. Several notable patterns Heteroptera, Lepidoptera, birds) (Coyne & Orr 322 Russell • ANN. ZOOL. FENNICI Vol. 40

1989, 1997, Laurie 1997, Orr 1997, Presgraves trends are apparent in the development of intrin- 2002, Price & Bouvier 2002). Haldaneʼs rule is sic post-zygotic RI. Equivalent data for sexual probably a composite phenomenon refl ecting isolation are relatively lacking, so this isolation both faster male evolution (through sexual selec- mechanism is not considered here. tion) and the recessivity of x-linked genes gen- erating hybrid unfi tness (the Dominance theory) (Laurie 1997, Orr 1997). Material and methods Various mechanisms can induce intrinsic post-zygotic RI; for example, founder-effect spe- Data on hybrid unfi tness were obtained from ciation (Carson & Templeton 1984), differences interspecifi c crosses largely described in Argue in ploidy between related taxa (Stebbins 1950), and Dunham (1999) and Ryabov (1981) (see and microbe-induced isolation (Werren 1998). Appendix). Consideration was limited to F1 However, the Dobzhansky-Muller model has the viability and fertility as in equivalent compara- greatest general explanatory power. Dobzhansky tive studies of other organisms. Thus, although (1936) and Muller (1940) considered that hybrid many studies have documented gross hybrid sterility and inviability result as pleiotropic by- breakdown only in later hybrid generations products of independent evolution in allopatric (Edmands 2002), due to break-up of parental populations. In the Dobzhansky-Muller (D-M) gene combinations through recombination, model, alleles that enhance fi tness on the normal these were discarded. Analysis was limited to genetic background may lower fi tness when studies that had examined fertility and viability brought together in hybrids with alleles from of both sexes. Asymmetries in the strength of another taxa. Thus, RI results from the accu- post-zygotic RI between reciprocal crosses often mulation of deleterious epistatic interactions exist (e.g., Tiffi n et al. 2001). However, recipro- in hybrids. Empirical and theoretical studies cal crosses were usually not distinguished in implicate natural selection in the accumulation of the source literature, so they are generally not D-M incompatibilities (e.g., Coyne et al. 1997, distinguished here; in the occasional instances Turelli et al. 2001). Theoretical analyses indicate when reciprocal cross data was available, crosses that D-M incompatibilities are rare and that the were averaged across. Similarly, different races severity of hybrid unfi tness and the number of of a particular species can be isolated to differing loci underlying incompatibilities should increase degrees with the same taxon (Argue & Dunham as the square of time separating two taxa (Orr 1999); although any such instances identifi ed in 1995, Orr & Turelli 2001). So, the D-M model the literature happened to be discarded from this predicts an approximately regular rate in the analysis because of phylogenetic non-independ- acquisition of RI, in contrast to other post- ence, this caveat remains potentially important. zygotic RI mechanisms that typically induce The magnitude of RI was coded from 0 (both speciation instantaneously (e.g., founder-effect sexes viable and fertile) to 4 (both sexes invi- speciation, ploidy differences, etc.). Compara- able). A code of “2” means that both sexes are tive studies may therefore be used to evaluate viable but sterile, such that effectively complete the relative importance of different isolation intrinsic RI exists (see Fig. 1 for isolation index mechanisms. classifi cation). Fish are the most diverse vertebrate lineage, Mitochondrial cytochrome b sequences were yet RI evolution in fi sh has not been considered obtained from GenBank (accession numbers are within a comparative context. Many experimen- given in the Appendix) and aligned using Clus- tal crosses have involved teleost fi sh taxa for talX (Thompson et al. 1994). Pairwise distance aquacultural research (for example, Schwartz estimates were estimated using the Kimura-2- (1981) lists over 1800 studies), and these con- parameter implemented in Mega v. 2.0 (Kumar stitute a large and important, but unexploited et al. 2001). The cytochrome-b gene, unlike resource. Here, I examine a small subset of these allozymes, is unlikely to be affected by selection crosses in relation to mitochondrial cytochrome on traits conferring RI (Fitzpatrick 2002), and is b divergence, and investigate whether general therefore an appropriate basis for genetic dis- ANN. ZOOL. FENNICI Vol. 40 • Evolution of post-zygotic isolation in fi sh 323

family corrected tance estimation. However, mitochondrial DNA 4 exhibits rate variation between different lineages non-family corrected (Rand 1994), and is generally only useful for lin- 3 eages separated by < 10 myr because of saturation properties (Meyer 1994). These caveats are 2 considered further in the Discussion. A further 1 caveat that may be only occasionally important Post-zygotic isolation index is that interspeciﬁ c introgessive hybridisation 0 05 1015202530 involving mitochondrial DNA may lead to under- Cytochrome b sequence divergence (%) estimation of overall genetic divergence (Smith 1992). Fig. 1. Plot of intrinsic post-zygotic reproductive isola- In ensuring phylogenetic independence of the tion against mitochondrial cytochrome b distances (Kimura-2-parameter corrected). Isolation index is as cross data crosses were not averaged across (e.g., follows: 0, both sexes fertile; 0.5, one sex fertile, the by following Felsensteinʼs (1985) phylogeneti- other sex some individuals recorded as fertile; 1, one cally independent contrasts method) because of sex fertile, one sex viable but infertile; 1.5, one sex current uncertainties in phylogenetic relation- sometimes fertile, one sex viable but infertile; 2, both ships amongst higher-order taxa. Instead, two sexes viable but infertile; 2.5, one sex viable but infertile, one sex sometimes viable; 3, one sex viable, one alternative approaches were adopted: (1) follow- sex missing; 3.5, one sex sometimes viable, one sex ing Price and Bouvier (2002), a list of crosses missing; 4, both sexes missing. All data derives from was prepared wherein each species was repre- method 1 for controlling phylogenetic dependence, sented in only one cross; (2) in a more robust wherein species were only considered in one cross, analysis based on the database generated by (1), irrespective of familial membership. The family-corrected subclass considers only one cross per family only one cross per family was considered. Fol- (method 2). The line of best ﬁ t through the family-cor- lowing phylogenetic corrections, 37 (method 1) rected data is a second-order polynomial (y = 0.0081x2 and 17 (method 2) crosses from 17 families were – 0.007). available for comparative analysis (Appendix). Much ancillary information exists regarding the maximum genetic distance allowing partial genetic and environmental determinants of most reproductive compatibility. This data derives traditional morphological markers constrain their from molecular genetic analyses of introgres- reliability in detecting post-F1 hybrids (Campton sive hybridisation involving wild populations 1987).

(where post-F1 hybrids are formed), backcrossing experiments that examine the fertility of only one sex from the F1 generation, and from studies Results that certifi ed F1 fertility but were excluded from the comparative analysis because of phyloge- In the comparative analysis only six out of 17 netic non-independence (using method 1). Such congeneric crosses showed F1 fi tness reduction data was examined to determine if incomplete (see Appendix). But, most heterogeneric crosses isolation extended beyond the range detected in (13 out of 19) showed fi tness reduction. Post- the comparative analysis. 32 additional crosses zygotic isolation only became evident after a were compiled for this analysis from Verspoor sequence divergence of 7%, and the maximum and Hammar (1991), Argue and Dunham (1999) divergence at which no isolation was reported and Sakaizumi et al. (1992) (data available from was 20.6%. The magnitude of unfi tness was author upon request). Again, genetic distances positively related to sequence divergence were based on cytochrome b sequences obtained (Fig. 1). Moreover, the rate of increase of isola- from GenBank. Morphological analyses of tion appeared to increase gradually with time; introgressive hybridisation were not consid- for instance, the line of best fi t for the data ered because of occasional inconsistencies in from method 2 (only one cross considered per their fi ndings with molecular work (Verspoor family) was a second-order polynomial forced & Hammar 1991), and because the unknown through the origin (Fig. 1). Those crosses that 324 Russell • ANN. ZOOL. FENNICI Vol. 40

resulted in F1s with reduced sterility (isolation cally, these analyses indicate that the severity of index = 0.5–2) were between taxa with an aver- hybrid unfi tness should increase as the square age sequence divergence of 14.3% (method 1) of time separating two taxa; Orr 1995, Orr & and 14.0% (method 2). The equivalent averages Turelli 2001). Equivalent comparative studies between taxa that produced inviable F1s (index = of other organisms have not detected this pat- 2.5–4) were 23.1% and 21.2% (methods 1 and 2 tern, and have instead generally inferred linear respectively). Means for inviability and sterility relationships. This difference is largely due to were signifi cantly different using Mann-Whit- the delayed acquisition of RI in fi sh, which only ney tests (method 1: U = 48, n = 19; p < 0.001; becomes evident after a sequence divergence of method 2: w = 10, n = 8; p < 0.05). 7%; in contrast, RI in other groups is apparent Analysis of the ancillary information indi- relatively much sooner. This disparity has sev- cated that the number of instances of partial eral possible explanations, including differences reproductive compatibility declined rapidly with in the relative importance of speciation mecha- sequence divergence (data not shown). Partial nisms (mechanisms that typically induce spe- reproductive compatibility occurred at a maxi- ciation instantaneously may be more prevalent in mum sequence divergence of 18.9% (Oryzias non-fi sh groups). Alternatively, my data may be curvinotus ¥ O. latipes; Sakaizumi et al. 1992). biased by non-random sampling introduced by Thus, incomplete isolation does not extend aquacultural practices (see below). beyond the range detected in the comparative Hybrid sterility evolves more rapidly than analysis (20.6%; Appendix). inviability in fi sh as well. These forms of unfi t- ness have different genetic bases as suggested by the fact that hybrid sterility mutations tend to Discussion affect one sex only, whereas most lethal mutations kill both sexes (Wu & Davis 1993). Male sex- and This comparative analysis has ascertained that reproduction related genes often evolve more several patterns of RI evolution that have been quickly than those of females (e.g., Hollocher & characterized in various taxa exist in fi sh too. Wu 1996); a pattern evident in the more rapid evo- For instance, as in Drosophila (Coyne & Orr lution of male sterility in male-heterogametic taxa 1989, 1997) and Lepidoptera (Presgraves 2002), (Orr 1997). Nevertheless the evolutionary rate of intrinsic post-zygotic isolation increases gradu- such genes in both sexes has been suffi cient on ally. This pattern is consistent with the relatively average to induce complete hybrid sterility before greater importance of the D-M model in describ- complete inviability in Drosophila (Coyne & Orr ing the evolution of RI. Otherwise, alternative 1997) and fi sh (method 1: Mann-Whitney test: U models of RI evolution (e.g., frequent microbe- = 92, n = 14, p < 0.05; insuffi cient crosses from induced speciation, polyploid speciation, etc.) method 2 exist for a comparable analysis). would have regularly generated signifi cant Because fi sh sex chromosomes are typically isolation at low sequence divergences; a pattern homorphic (morphologically similar) (Ohno not observed here. Interestingly, the observa- 1974), it is frequently diffi cult to characterise tion that the rate of increase in hybrid unfi tness fi sh sex-determination systems. Thus, although seems to increase with time is predicted from ~1700 fi sh species have been cytogenetically theoretical treatments of the D-M model (specifi - characterised, only ~10% possess cytogeneti-

Table 1. Crosses involving male (XY) or female (ZW) heterogametic species.

Family Species 1 Species 2 Sex-determination system Notes on F1s

Cyprinidae Carassius auratus Cyprinus carpio XX-XY Males sterile; females sterile Rhodeus ocellatus Tanakia limbata ZZ-ZW Male only broods Rhodeus ocellatus Tanakia lanceolata ZZ-ZW Male only broods Poeciliidae Poecilia sphenops P. latipinna ZZ-ZW No hybrid unfi tness ANN. ZOOL. FENNICI Vol. 40 • Evolution of post-zygotic isolation in fi sh 325 cally distinct sex chromosomes (Devlin & Naga- Here, I examined the evolution of intrinsic hama 2002, Arkhipchuk 1995). Of this fraction, post-zygotic isolation. However, because sexual both male (XY) and female (ZW) heterogametic isolation would act fi rst to isolate taxa, the gen- systems have been detected, although the addi- eral relevance of post-zygotic isolation to specia- tional presence of autosomal infl uences in many tion remains somewhat controversial. To evalu- such systems means that the number that can be ate its relevance one must examine the number directly considered in relation to Haldaneʼs rule of species pairs that experience natural range is reduced. I know of only four crosses involv- overlap and suffer intrinsic hybrid unfi tness. ing species that possess the same heterogametic From my data set, two out of the 19 species pairs system (Table 1). Of these, three conform to possessing an intrinsic isolation index ≥ 0.5 have Haldaneʼs rule. These numbers are too small to naturally overlapping ranges (Barbus meridiona- evaluate the generality of Haldaneʼs rule in fi sh, lis–B. barbus, and Alburnus alburnus–Leuciscus which is a nearly ubiquitous pattern in the early cephalus) (Scribner et al. 2001). However, the stages of RI evolution in other groups. So, more proportion of species pairs likely to satisfy these cytogenetic work in fi sh is clearly advantageous. criteria is probably greater because I have not Introgressive hybridisation is prevalent amongst considered extrinsic post-zygotic unfi tness that interspecifi c fi sh populations (Scribner et al. becomes apparent under natural environments 2001), leading Thorgaard and Allendorf (1988) (for example, between different sympatric eco- to suggest that hybrid fi sh may be less suscep- types of the three-spined stickleback, Gasteros- tible to severe developmental incompatibilities teus aculeatus; Hatfi eld and Schluter (1999)). than interspecifi c hybrids from other vertebrate Thus, intrinsic postzygotic RI is probably likely classes showing comparable levels of genetic to play an important role in isolating natural fi sh divergence. The mean divergence of completely populations. Similar conclusions were drawn for isolated taxa in this study was 23.1%, although Lepidoptera (Presgraves 2002) and Drosophila this may be an underestimate because saturation (Coyne & Orr 1997). of cytochrome b sequences can become evident My results are unlikely to have much refl ected around 20% divergence (Meyer 1994). Assum- rate variation in mitochondrial DNA between ing a molecular clock of 2% myr–1, this value lineages (e.g., the rate is strongly reduced equates to ~11.6 myr. Similarly, the maximum amongst salmonids; Smith 1992) because similar divergence for partial reproductive compat- trends generally resulted from both methods that ibility was 20.6% (~10.3 myr). Only the former controlled for phylogenetic dependence. How- value appears substantially larger than equiva- ever, of potentially greater importance are biases lent values for other lineages (excepting birds) in how hybridisation experiments are selected (Table 2). Although necessarily crude, these fi g- and published (Edmands 2002). The aquacul- ures donʼt provide strong support for Thorgaard and Allendorfʼs (1998) hypothesis. Nevertheless, amongst the fi ve organismal groups for which Table 2. Estimated mean age for completely isolated taxa, and maximum ages for partial fertility in various comparative data currently exists, it seems that taxonomic groups. total post-zygotic isolation evolves most slowly in birds and fi sh. Also noteworthy is that the Taxonomic group Mean age of Maximum age for mean divergence time for completely isolated total post-zygotic partial reproductive fi sh taxa (~11.6 myr) substantially exceeds that isolation (myr) compatibility (myr) between 108 putatively allopatric sister species Drosophila1 ~5 ~7.8 pairs in Avise et al. (1998) (~2.05 myr), and the Birds2 ~11.5 ~16.8 mean time for speciation between putatively Lepidoptera3 ~3.5 ~7.5 allopatric clades (~1.1–2.9 myr species–1; data Amphibians4 ~4.2 ~9 recalibrated assuming a molecular clock of 1 –1 Data obtained from comparative studies in: Coyne & 2% myr ) estimated by McCune and Lovejoy Orr 1997; 2Price & Bouvier 2002; 3Presgraves 2002; (1998). Both these latter studies used cyto- 4Sasa et al. 1998. Estimates were generated from allo- chrome b sequences. zyme data using a molecular clock of 2% myr–1. 326 Russell • ANN. ZOOL. FENNICI Vol. 40 tural industry is interested in developing hybrid Arkhipchuk, V. V. 1995: Role of chromosomal and genome strains that show heterosis, and sometimes steril- mutations in the evolution of bony fi shes. — Hydrobiol. J. 31: 55–65. ity (to prevent genetic contamination of natural Avise, J. C., Walker, D. & Johns, J. C. 1998: Speciation populations and to enhance individual growth) durations and Pleistocene effects on vertebrate phylo- and sex-ratio distortions (to regulate stock geography. — Proc. Roy. Soc. Lond. Biol. Series B 265: population growth) (Bartley et al. 2001). Thus, 1707–1712. Bartley, D. M., Rana, K. & Immink, A. J. 2001: The use the aquacultural industry may preferentially of inter-specifi c hybrids in aquaculture and fi sheries. select and publish crossing experiments that — Rev. Fish Biol. Fish. 10: 325–337. yield hybrids with such traits, causing published Campton, D. E. 1987: Natural hybridization and introgres- crosses to be a non-random sample of those pos- sion in fi shes: methods of detection and genetic interpre- sible. However, a large proportion of crosses tations. — In: Ryman, N. & Utter, F. (eds.), Population genetics and fi shery management: 161–192. University produced wholly inviable offspring (e.g., fi ve out of Washington Press, Seattle. of the 17 (29%) crosses used in method 2), sug- Carson, H. L. & Templeton, A. R. 1984: Genetic revolutions gesting that these biases may not be especially in relation to speciation phenomena: the founding of new strong. Because biases may operate on multiple populations. — Annu. Rev. Ecol. Syst. 15: 97–131. hybrid traits (e.g., heterosis, sterility, partial invi- Coyne, J. E. & Orr, H. A. 1989: Patterns of speciation in Dro- sophila. — Evolution 43: 362–381. ability), effects of their interactions on the com- Coyne, J. E. & Orr, H. A. 1997: “Patterns of speciation in parative analysis are potentially complex, but not Drosophila” revisited. — Evolution 51: 295–303. necessarily detrimental. Coyne, J. A., Barton, N. H. & Turelli, M. 1997: A critique In summary, the evolution of intrinsic post- of Sewall Wrightʼs shifting balance theory of evolution. zygotic RI in fi sh shows several phenomema — Evolution 51: 643–670. Devlin, R. H. & Nagahama, Y. 2002: Sex determination and identifi ed in other taxa. Specifi cally, hybrid sex differentiation in fi sh: an overview of genetic, physi- sterility evolves more rapidly than inviability, ological, and environmental infl uences. — Aquaculture and RI evolves gradually, consistent with the 208: 191–364. relatively greater importance of the D-M model Dobzhansky, T. 1936: Studies of hybrid sterility. II. Locali- in causing RI. It is likely that such intrinsic isola- sation of sterility factors in Drosophila pseudoobscura hybrids. — Genetics 21: 113–135. tion plays an important role in isolating natural Edmands, S. 2002: Does parental divergence predict reproduc- populations, either directly or indirectly through tive compatibility? — Trends Ecol. Evol. 17: 520–527. promoting sexual isolation (reinforcement; Noor Felsenstein, J. 1985: Phylogenies and the comparative 1999). Because of general diffi culties in charac- method. — Am. Nat. 125: 1–15. terising fi sh genetic sex-determination systems, Fitzpatrick, B. M. 2002: Molecular correlates of reproductive isolation. — Evolution 56: 191–198. and because of the plethora of systems poten- Haldane, J. B. S. 1922: Sex ratio and unisexual sterility in tially represented in any given cross, too few animal hybrids. — J. Genet. 12: 101–109. crosses are currently available to evaluate the Hatfi eld, T. & Schluter, D. 1999: Ecological speciation in generality of Haldaneʼs rule in fi sh. sticklebacks: environment dependent hybrid fi tness. — Evolution 53: 866–873. Hollocher, H. & Wu, C.-I. 1996: The genetics of reproductive isolation in the Drosophila simulans Clade: x vs. auto- Acknowledgements somal effects and male vs. female effects. — Genetics 143: 1243–1255. The United Kingdom Natural Environment Research Kumar, S., Tamura, K. & Jakobsen, I. B. 2001: MEGA Council provided fi nancial support. The comments of one 2: molecular evolutionary genetic analysis software. of the referees considerably benefi ted the manuscript. Many — Bioinformatics 17: 1244–1245. thanks to those who deposited their sequence data in Gen- Laurie, C. C. 1997: The weaker sex is heterogametic: 75 Bank. years of Haldaneʼs rule. — Genetics 147: 937–951. Mayr, E. 1942: Systematics and the origin of species. — Columbia University Press, New York. McCune, A. R. & Lovejoy, N. R. 1998: The relative rate of References sympatric and allopatric speciation in fi shes: tests using DNA sequence divergence between sister species and Argue, B. J. & Dunham, R. A. 1999: Hybrid fertility, intro- among clades. — In: Howard, D. J. (ed.), Endless forms: gression, and backcrossing in fi sh. — Rev. Fish. Sci. 7: species and speciation: 172–185. Oxford University 137–195. Press, Oxford. ANN. ZOOL. FENNICI Vol. 40 • Evolution of post-zygotic isolation in fi sh 327

Meyer, A. 1994: Shortcomings of the cytochrome b gene as a cytonuclear methods of biological inference. — Rev. molecular marker. — Trends Ecol. Evol. 9: 278–280. Fish Biol. Fish. 10: 293–323. Muller, H. J. 1940: Bearing of the Drosophila work on sys- Singh, R. S. & Kulathinal, R. J. 2000: Sex gene pool evolu- tematics. — In: Huxley, J. S. (ed.), The new systematics: tion and speciation: a new paradigm. — Genes Genet. 185–268. Clarendon Press, Oxford. Syst. 75: 119–130. Noor, M. A. F. 1999: Reinforcement and other consequences Smith, G. R. 1992: Introgression in fi shes — signifi cance for of sympatry. — Heredity 83: 503–508. paleontology, cladistics and evolutionary rates. — Syst. Ohno, S. 1974: Animal cytogenetics, vol. 4. Chordata 1. Biol. 41: 41–57. — Gebruder Borntraeger, Berlin. Smitherman, R. O., Dunham, R. A. & Whitehead, P. K. 1996: Orr, H. A. 1995: The population genetics of speciation: the Selection, hybridization and genome manipulation in evolution of hybrid incompatibilities. — Genetics 139: Siluroidei. — Aquat. Living Resour. 9: 93–102. 1805–1813. Stebbins, G. L. 1950: Variation and evolution in plants. Orr, H. A. 1997: Haldaneʼs rule. — Annu. Rev. Ecol. Syst. — Columbia University Press, New York. 28: 195–218. Suzuki, R. 1973: On the sterile F1 hybrids between the Orr, H. A. & Turelli, M. 2001: The evolution of postzygotic genera Pseudogobio and Gnathopogon (Cyprinidae). isolation: accumulating Dobzhansky-Muller incompat- — Jap. J. Ichthyology 20: 235–238. ibilities. — Evolution 55: 1085–1094. Thompson, J. D., Higgins, D. G. & Gibson, T. J. 1994: Clus- Presgraves, D. C. 2002: Patterns of postzygotic isolation in tal W: improving the sensitivity of progressive multiple Lepidoptera. — Evolution 56: 1168–1183. alignment through sequence weighting, positions spe- Price, T. D. & Bouvier, M. M. 2002: The evolution of F1 cifi c gap penalties and weight matrix choice. — Nucl. postzygotic incompatibilities in birds. — Evolution 56: Acids Res. 22: 4673–4680. 2083–2089. Thorgaard, G. H. & Allendorf, F. W. 1988: Developmental Rand, D. M. 1994: Thermal habit, metabolic rate and the genetics in fi shes. — In: Malacinski, G. M. (ed.), Devel- evolution of mitochondrial DNA. — Trends Ecol. Evol. opmental genetics of animals and plants: 369–391. Mac- 9: 125–131. Millan, New York. Ryabov, I. N. 1981: Intersubfamiliy hybridization of the Tiffi n, P., Olson, M. S. & Moyle, L. C. 2001: Asymmetrical family Cyprinidae. — J. Ichthyology 19: 57–73. crossing barriers in angiosperms. — Proc. Roy. Soc. Sakaizumi, M., Shimizu, Y. & Hamaguchi, S. 1992: Elec- Lond. Biol. Series B 268: 861–867. trophoretic studies of meiotic segregation in inter- and Turelli, M., Barton, N. H. & Coyne, J. A. 2001: Theory and intraspecifi c hybrids among east Asian species of the speciation. — Trends Ecol. Evol. 16: 330–343. genus Oryzias (Pisces: Oryziatidae). — J. Exp. Zool. Verspoor, E. & Hammar, J. 1991: Introgressive hybridisation 264: 85–92. in fi shes: the biochemical evidence. — J. Fish Biol. 39 Sasa, M. M., Chippendale, P. T. & Johnson, N. A. 1998: Pat- (Supplement A): 309–334. terns of postzygotic isolation in frogs. — Evolution 52: Werren, J. H. 1998: Wolbachia and speciation. — In: 1811–1820. Howard, D. J. (ed.), Endless forms: species and specia- Schwartz, F. J. 1981: World literature to fi sh hybrids with tion: 245–260. Oxford University Press, Oxford. an analysis by family, species, and hybrid: Supplement Wiggins, T. A., Bender, T. R., Mudrak, V. A. & Takacs, M. 1. — NOAA Technical Report NMFS (National Marine A. 1983: Hybridization of yellow perch and walleye. Fisheries Service) SSRF (Special Scientifi c Report Fish- — Progressive Fish-Culturist 45: 131–133. eries) 750. U.S. Department of Commerce, Seattle. Wu, C.-I. & Davis, A. W. 1993: Evolution of postmating Scribner, K. T., Page, K. S. & Bartron, M. L. 2001: Hybridi- reproductive isolation: the composite nature of Haldaneʼs zation in freshwater fi shes: a review of case studies and rule and its genetic basis. — Amer. Nat. 142: 187–212. 328 Russell • ANN. ZOOL. FENNICI Vol. 40

Appendix. Cross data for comparative analysis. Cytochrome b sequences were obtained from GenBank. * = cross used in analyses involving only one cross per family.

Family Species 1 (cyt. b Species 2 (cyt. b Sequence Hybrid Reference accession no.) accession no.) divergence unﬁ tness (%) score

Acipenseriformes: Acipenser ruthensus Huso huso 4.9 0* 1 Acipenseridae (AF283733) (AF283745) Beloniformes: Oryzias latipes O. luzonensis 19.7 4* 2 Adrianichthyidae (AB084753) (AB084755) Cyprinodontiformes: Aphyosemion A. exiguum 5.9 0* 1 Aplocheilidae bualanum (AF002304) (AF002303) Aphyosemion A. christyi 8.3 0 1 cognatum (AF002322) (AF002324) Cypriniformes: Misgurnus Cobitis biwae 18.3 2* 3 Cobitidae anguillicaudatus (AB039347) (AF051868) Cypriniformes: Cyprinus carpio Zacco temminckii 25.0 4* 4 Cyprinidae (Y09469) (AF309084) Richardsonius Mylocheilus 13.5 0.5 1 balteatus caurinus (AY096011) (AF117169) Notemigonus Scardinius 15.1 2 1 crysoleucas erythrophthalmus (U01318) (Y10444) Rutilus rutilus Abramis brama 12.9 0.5 1 (AF090772) (Y10441) Barbus meridionalis B. barbus 8 1 1 (AF112130) (AY013479) Leuciscus schmidti Schizothorax 24.8 4 4 (AY026396) pseudoaksaiensis (AF180827) Gobio gobio Chondrostoma nasus 23.5 4 4 (AJ388431) (AJ388454) Biwia zezera Henigrammocypris 21.7 4 4 (AF309507) rasborella (AF375863) Carassius auratus Zacco platypus 25.3 4 4 (AF045966) (AF309085) Hypophthalmichthys Aristichthys nobilis 8.1 0 1 molitrix (AF051855) (AF051866) Puntius conchonius Rhodeus ocellatus 25.5 4 4 (AY004751) (AF051876) Alburnus alburnus Leuciscus cephalus 17.2 2 1 (AJ388428) (LCE389571) Cyprinella lutrensis C. venusta 15.8 0 1 (U01319) (AF261218) Cyprinodontiformes: Fundulus olivaceus F. notatus 6.4 0* 1 Fundulidae (U77123) (L31598) Cyprinodontiformes: Xiphophorus helleri X. maculatus 4.0 0* 1 Poeciliidae (AF404301) (XMU06515)

Continues ANN. ZOOL. FENNICI Vol. 40 • Evolution of post-zygotic isolation in ﬁ sh 329

Appendix. Continued.

Family Species 1 (cyt. b Species 2 (cyt. b Sequence Hybrid Reference accession no.) accession no.) divergence unﬁ tness (%) score

Perciformes: Lepomis L. cyanellus 20.6 0 1 Centrarchidae macrochirus (AY115973) (AY115976) Pomoxis annularis P. nigromaculatus 2.1 0* 1 (AY115990) (AY115992) Micropteris M. dolomieu 11.8 0 1 salmoides (AY115998) (AY116000) Perciformes: Cichlidae Oreochromis O. niloticus 0.7 0* 1 mossambicus (AF550011) (X81565) Sarotherodon Oreochromis aureus 3.4 0 1 galilaeus (AF375617) (AF375618) Perciformes: Moronidae Morone saxatilis M. chrysops 14.5 1* 1 (AF240746) (AF240745) Perciformes: Percidae Perca ﬂ avescens Stizostedion vitreum 20.7 4* 5 (AF045357) (AF386602) Perciformes: Serranidae Epinephelus E. aeneus 19.5 4* 6 marginatus (AJ420206) (AJ420205) Perciformes: Sparidae Acanthopagus latus Sparidentex hasta 12.8 0* 6 (AF539743) (AF240734) Sparus auratus Pagrus major 21.1 2 1 (AJ319809) (NC_003196) Pleuronectiformes: Pleuronectes Platichthys ﬂ esus 6.1 0* 1 Pleuronectidae platessa (AF113179) (AY164472) Salmoniformes: Oncorhynchus O. clarkii 4.4 0 1 Salmonidae mykiss (D58401) (AY032633) Oncorhynchus keta O. gorbuscha 4.2 0 1 (AF165078) (AF165077) Salmo trutta S. salar 7.0 1* 1 (D58400 ) (AF202032) Oncorhynchus Salvelinus fontinalis 14.4 0 1 masou (D58402) (D58399) Siluriformes: Clariidae Clarias fuscus C. gariepinus 16.0 1.5* 7 (AF416885) (AF475153) Siluriformes: Ictaluridae Ictalurus furcatus I. punctatus 9.5 4* 1 (AF484159) (AB069646)

1 = Argue & Dunham 1995; 2 = Sakaizumi et al. 1992; 3 = Suzuki 1973; 4 = Ryabov 1981; 5 = Wiggins et al. 1983; 6 = Bartley et al. 2001; 7 = Smitherman et al. 1996.