GENETIC STRUCTURE AND EVOLUTION OF A FIRE ANT HYBRID ZONE

Two introduced fire ants, Solenopsis invicta and S. richteri, hybridize over an extensive area in the United States spanning central Mississippi, Alabama, and western Georgia. We studied a portion of this hybrid zone in northwestern Mississippi in detail by sampling ants at many sites along two transects extending across the zone and examining gene frequency and size distributions at a large number of genetic and morphological markers. The distributional patterns at these markers are most consistent with the mosaic hybrid zone model, whereby the distribution of various fire ant genotypes is determined initially by the historical patterns of colonization of newly available habitats. However, these distributional patterns probably do not reflect the equilibrium state of interactions because of the very recent secondary contact of the species (< 60 yr) and the dynamic nature of available nesting habitats in this area. Our data suggest that, with prolonged contact and interaction, differential fitness of various hybrid genotypes due to intrinsic and extrinsic selective factors is important in structuring the hybrid zone. For instance, consistent differential introgression of morphological and genetic markers, combined with previous evidence of differences in developmental stability among genotypes, suggest reduced fitness of hybrids relative to parentals due to intrinsic selection (as may be caused by breakup of parental gene complexes). Furthermore, marked reductions in the occurrence of parental-like hybrids in areas where the similar parental species is common suggest reduced fitness of these parental-like hybrids in competition with the parentals (i.e., extrinsic selection). Because the relative roles of such deterministic as well as stochastic forces apparently vary both spatially and temporally, the eventual distribution of the various fire ant genotypes and the fate of the hybrid zone in the United States is difficult to predict.     

MITOCHONDRIAL INHERITANCE PATTERNS ACROSS A COTTONWOOD HYBRID ZONE: CYTONUCLEAR DISEQUILIBRIA AND HYBRID ZONE DYNAMICS

In this study we examine the cytoplasmic inheritance patterns of an interspecific hybridizing population of Fremont and narrowleaf cottonwoods, using mitochondrial DNA. Three mitochondrial probes showing polymorphisms were used to distinguish between trees of known nuclear inheritance. Every tree screened had only one cytoplasmic genotype, either Fremont or narrowleaf. Thus, these results demonstrate that mitochondria are uniparentally inherited in these trees. Previous studies of the nuclear inheritance of this interspecific hybridizing population of cottonwood trees indicated an asymmetry in the frequency of parental genes. Using mitochondrial markers we tested one hypothesis potentially responsible for this asymmetric distribution (i.e., trees of mixed genotypes will be sterile or will not survive if their cytoplasm is derived from one or the other parent). Our results, however, show that both Fremont and narrowleaf mitochondrial markers are found in trees with mixed nuclear genotypes. Thus, nuclear-cytoplasmic incompatibilities do not appear to account for the asymmetric distribution of nuclear genotypes within the hybrid swarm. An alternative explanation for the observed asymmetric distribution of nuclear genotypes is advanced. Although nuclear-cytoplasmic incompatibilities do not appear to explain the asymmetric distribution of nuclear alleles within the hybrid zone, nonrandom associations between nuclear and cytoplasmic genotypes do exist. For example, all F₁ hybrids had Fremont mitochondrial genotypes. Furthermore, backcrosses between F₁ hybrid and narrowleaf trees have a higher than expected proportion of heterozygous loci and a higher than expected proportion of Fremont mitochondria. We propose that seeds, seedlings, or trees with high proportions of heterozygous loci are at a disadvantage unless they also have the Fremont mitochondrial genotype. While it is generally difficult to infer dynamic processes from static patterns, studies such as ours enable one to gain new insights to the dynamics of plant hybrid zones. A hybridization pattern of decreasingly complex backcrosses as one proceeds from higher to lower elevation within the hybrid swarm, a residue of Fremont cytoplasmic DNA within the pure narrowleaf population, and the unidirectional nature of these crosses suggest that the narrowleaf population may be spreading down the canyon and the Fremont population receding. The eventual fate of the hybrid zone, in relation to these processes, is discussed.     

AN UNUSUAL PATTERN OF GENE FLOW BETWEEN THE TWO SOCIAL FORMS OF THE FIRE ANT SOLENOPSIS INVICTA

Uncertainty over the role of shifts in social behavior in the process of speciation in social insects has stimulated interest in determining the extent of gene flow between conspecific populations differing in colony social organization. Allele and genotype frequencies at 12 neutral polymorphic protein markers, as well as the numbers of alleles at the sex-determining locus (loci), are shown here to be consistent with significant ongoing gene flow between two geographically adjacent populations of Solenopsis invicta that differ in colony queen number. Data from a thirteenth protein marker that is under strong differential selection in the two social forms confirm that such gene flow occurs. Data from this selected locus, combined with knowledge of the reproductive biology of the two social forms, further suggest that interform gene flow is largely unidirectional and mediated through males only. This unusual pattern of gene flow results from the influence of the unique social enviroments of the two forms on the behavior of workers and on the reproductive physiology of sexuals.     

A TEST OF QUEEN RECRUITMENT MODELS USING NUCLEAR AND MITOCHONDRIAL MARKERS IN THE FIRE ANT SOLENOPSIS INVICTA

We assess nestmate queen relatedness and the genetic similarity of neighboring nests in the polygyne (multiple-queen) social form of the introduced fire ant Solenopsis invicta using both nuclear and mitochondrial markers. We find that estimates of queen relatedness calculated with both types of markers do not differ statistically from zero. Furthermore, there is no significant relationship between the genetic similarity and geographic proximity of nests in each of six study sites. In contrast to these findings, sites show strong mitochondrial, but no nuclear, genetic differentiation. Our results suggest that nonnestmate queen recruitment occurs at a high frequency in introduced populations of this species. Moreover, queens within nests seem to represent a random sample of the queens within the site in which they reside. Therefore, kin selection models that rely on the recruitment of only nestmate queens to explain the persistence of polygyny in ants do not apply to polygyne S. invicta in its introduced range.     

Development of RAPD markers in two introduced fire ants, Solenopsis invicta and S. richteri, and their application to the study of a hybrid zone

We developed RAPD DNA markers useful in distinguishing the fire ants Solenopsis invicta and S. richteri. An initial survey of 200 primers revealed seven informative markers; family studies allowed us to determine expression patterns and to confirm Mendelian inheritance of these markers. The seven RAPD markers, one of which is inherited as a codominant marker, were employed along with three allozyme markers to describe the structure of a hybrid zone that has formed between the two species in the USA, where they have been introduced. The data suggest minor introgression of alleles from one parental species ( S. richteri ) into the other ( S. invicta ), which most likely reflects the documented recent movement of this hybrid zone. This pattern is interpreted as interspecific introgression rather than shared intraspecific polymorphism on the basis of comparisons with samples from native, non-hybridizing populations in South America that lack such polymorphism. The data further reveal that the structure of the hybrid zone in the USA varies geographically. One parental species ( S. invicta ) and the hybrids exhibit a mosaic distribution in the east; a gradual transition between the parental species occurs in the centre, with a large intervening area of hybrid genotypes only; and there is apparent contact between parental populations with a small or no intervening zone of hybridization in the west. These differing patterns in the structure of the hybrid zone presumably reflect the unique histories of colonization in different parts of the range of introduced fire ants.     

RESTRICTED GENE FLOW IN A MOVING HYBRID ZONE OF THE NEWTS TRITURUS CRISTATUS AND T. MARMORATUS IN WESTERN FRANCE

Two hybridizing species of newts, Triturus cristatus and T. marmoratus, with overlapping distributions show a parapatric distribution when surveyed in detail. The factors that govern the distribution of cristatus vs. marmoratus in the département (province) of Mayenne in western France are identified as forestation and relief. The parapatric hybrid zone running through Mayenne is narrow but widens to approximately 20 km in an area with mixed habitat. In this area most breeding sites are shared and F₁ hybrids form about 4% of the total population. Analysis of survey data collected about 30 years previously also shows an essentially parapatric distribution. Comparison of past and present distribution maps reveals that cristatus has superseded marmoratus over large areas in the south of Mayenne. An area where marmoratus replaced cristatus also exists, but it is more limited in size. Gene flow between cristatus and marmoratus is analyzed using 10 diagnostic genetic markers [9 protein loci and mitochondrial (mt) DNA]. In syntopic populations nuclear gene flow is bidirectional with a mean frequency of introgressed alleles (f) of 0.3%. In allotopic populations of cristatus and marmoratus gene flow is present in areas of species replacement (f = 0.3%), while gene flow appears to be absent in those areas that have been continuously occupied by a single species. At the biogeographic level, the presence or absence of introgression is paralleled by the persistence or absence, respectively, of pockets of cristatus–marmoratus syntopy. All F₁ hybrids possess the cristatus type mtDNA. This may be due to asymmetric interspecific mate choice and would explain the observed absence of introgression of the maternally inherited mtDNA genome in areas where cristatus replaced marmoratus. The cristatus–marmoratus hybrid zone bears characteristics of both the clinal (parapatric) hybrid zone model and the mosaic hybrid zone model. Such a mixed model—for which we propose the term “reticulate hybrid zone”—can be appreciated only if studied over a two-dimensional geographic area and also through time.     

AN EMPIRICAL TEST OF PREDICTIONS OF TWO COMPETING MODELS FOR THE MAINTENANCE AND FATE OF HYBRID ZONES: BOTH MODELS ARE SUPPORTED IN A HARD-CLAM HYBRID ZONE

Two models developed to discern the mode of selection in hybrid zones differ in some predictions. The tension-zone model predicts that selection acts against hybrids and independently of the environment (endogenous selection) and that selection is invariant throughout the hybrid zone. The ecological selection-gradient, or ecotone, model maintains that fitness of different genotypes varies in response to environmental variation (exogenous selection) and thus, that in a region of the zone, fitness of hybrids is at least equal to that of the parental species. Therefore, to assess the predominant mode of selection operating in a hybrid zone, it is fundamental to evaluate whether selection is acting specifically against hybrid individuals, that is, whether hybridity alone is the basis for deficiencies of hybrids, and to evaluate whether the relative fitness of hybrids versus that of pure species varies across the zone. In a hardclam (genus Mercenaria) hybrid zone located in a polyhaline lagoon in east-central Florida, we used age-specific and location-specific analyses to determine that a hybrid deficit occurrs, that the deficit seems to be due to selection against hybrids, and that selection varies across the zone. Various measures of deviation from Hardy-Weinberg equilibrium, linkage disequilibrium analyses, and shifts in allele frequencies at semidiagnostic loci support the idea that selection is strongest in the northern region of the lagoon, the zone of sympatry and hybridization. Southward, into the range of M. mercenaria (the numerically predominant species), the percentage of hybrids remains relatively high and selection against hybrids decreases. For some genetic linkage groups, selection for M. mercenaria alleles seems to be occurring, but selection seems to be acting principally against alleles characteristic of M. mercenaria and, to a lesser degree, for alleles characteristic of M. campechiensis (the rarer species). These findings and others from previous analyses we have done on this hybrid zone demonstrate that selection in the zone is complex, and that characteristics of both the tension-zone and ecotone models are present. Supporting the tension-zone model, selection against hybrids per se clearly occurs, but specific genotypes seem to be at a selective disadvantage, whereas others have a selective advantage, and selection operates differentially on the two parental species within the zone. Supporting the ecotone model, the strength of overall selection varies throughout the zone, and environmentally mediated selection in which each species and hybrids have an advantage in specific habitats occurs, but some selection against hybrids is invariant throughout the zone. Thus, the structure and genetic architecture of this hybrid zone appear to be products of a complicated interaction between both types of selective forces cited in the two competing models.     

THE GENETICS OF CENTRAL AND MARGINAL POPULATIONS OF DROSOPHILA SERRATA. II. HYBRID BREAKDOWN IN FITNESS COMPONENTS AS A CORRELATED RESPONSE TO SELECTION FOR DESICCATION RESISTANCE

Multiple-peak epistasis is one of the four premises that underlie Wright's shifting-balance theory of evolution. A selection experiment was conducted in an attempt to push different geographic populations to different fitness peaks as a correlated response to selection for an additively controlled character (desiccation resistance). Four populations of Drosophila serrata, sampled from central and marginal areas of its distribution along a 3000-km stretch of Australia's east coast, underwent selection for desiccation resistance for 14 generations. After selection had ceased, control lines from each of the populations were crossed to determine the amount of hybrid breakdown that existed before selection and selected lines were crossed to determine the amount of hybrid breakdown after selection. Hybrid breakdown was measured in three fitness traits: developmental time, viability, and fecundity. When the individual crosses were examined, virtually no evidence was found for hybrid breakdown between these populations. However, the level of hybrid breakdown in development time in the control lines increased as the distance between the populations in the field increased. This relationship was lost in the selected lines. Therefore, selection for desiccation resistance influenced the level of hybrid breakdown in a fitness trait, although selection may need to be maintained for longer than 14 generations if a new relationship between hybrid breakdown and distance is to be formed.