Cytonuclear Disequilibrium and Genetic Drift in a Natural Population of Ponderosa Pine

Corresponding author: Robert G. Latta, Department of Biology, Dalhousie University, 1355 Oxford St., Halifax, Nova Scotia B3H 4J1, Canada. E-mail robert.latta@dal.ca

Communicating editor: A. H. D. BROWN

	ABSTRACT

TOP ABSTRACT METHODS RESULTS DISCUSSION LITERATURE CITED

We measured the cytonuclear disequilibrium between 11 nuclear allozyme loci and both mitochondrial and chloroplast DNA haplotypes in a natural population of ponderosa pine (Pinus ponderosa, Laws). Three allozyme loci showed significant associations with mtDNA variation, while two other loci showed significant association with cpDNA. However, the absolute number of individuals involved in any of the associations was small, such that in none of the nuclear-organellar combinations was the difference between observed and expected numbers >11 individuals. Patterns of association were not consistent across loci or organellar genomes, suggesting that they are not the result of mating patterns, which would act uniformly on all loci. This pattern of disequilibria is consistent with the action of genetic drift and with existing knowledge of the structure of this population and thus does not imply the action of other evolutionary processes. The overall magnitude (normalized disequilibrium) of associations was greater for maternally inherited mtDNA than for paternally inherited cpDNA, though this difference was neither large nor significant. Such significant disequilibria involving the paternally inherited organelle indicate that not only are there a limited number of seed parents, but the effective number of pollen parents is also limited.

CYTONUCLEAR disequilibrium is the nonrandom association of alleles or genotypes at a nuclear locus with haplotypes at cytoplasmically inherited organellar DNA (ASMUSSEN et al. 1987 ; SCHNABEL and ASMUSSEN 1989 ). Thus, cytonuclear disequilibrium is analogous to linkage disequilibrium between nuclear loci, but across genomes with different patterns of inheritance. Theories describing the creation and maintenance of nonzero disequilibria have been developed only relatively recently (e.g., ASMUSSEN et al. 1987 , ASMUSSEN et al. 1989 ; SCHNABEL and ASMUSSEN 1989 ; ASMUSSEN and ARNOLD 1991 ; ASMUSSEN and ORIVE 2000 ), because the molecular tools needed to simultaneously assay nuclear and organellar variation are relatively recent. Nevertheless, a substantial body of theoretical work is available that predicts the level of disequilibria expected from a variety of evolutionary forces, including migration and hybridization (ASMUSSEN et al. 1989 ; SCHNABEL and ASMUSSEN 1989 ; ASMUSSEN and ARNOLD 1991 ; ASMUSSEN and ORIVE 2000 ), selection (CRUZAN and ARNOLD 1993 , CRUZAN and ARNOLD 1994 ; BURKE et al. 1998 ; EDMANDS and BURTON 1999 ), and drift (FU and ARNOLD 1992 ). Most reported measures of disequilibrium give only the association between nuclear and maternally inherited organelles. This simply reflects the limitation that in most systems organellar DNA is maternally inherited. However, paternally inherited organelles are known in a few species, including conifers (reviewed in WAGNER 1992 ), mussels (ZOUROS et al. 1992 ; LIU et al. 1996 ), and potentially some species of angiosperms (MOGENSON 1996 ). Patterns of cytonuclear disequilibrium are expected to be controlled by different forces depending upon the mode of organellar inheritance (ASMUSSEN and ORIVE 2000 ). For example, paternally and maternally inherited organelles will be differentially responsive to seed and pollen movement in plants and this can been used to estimate the relative rates of seed and pollen movement through both cytonuclear (ORIVE and ASMUSSEN 2000 ) and other methods (MCCAULEY 1994 ).

Migration creates cytonuclear disequilibria primarily when disequilibria exist in the migrant pool (ASMUSSEN and ARNOLD 1991 ; ASMUSSEN and ORIVE 2000 )—a situation most easily found where migrants are derived from two genetically distinct sources. Thus empirical studies of disequilibria have been undertaken most commonly in hybrid zones between two or more taxa exhibiting diagnostic nuclear and cytoplasmic markers (e.g., PAIGE et al. 1991 ; CRUZAN and ARNOLD 1993 , CRUZAN and ARNOLD 1994 ; SCRIBNER and AVISE 1994 ; SCRIBNER et al. 1999 ; see ARNOLD 1993 for a review of these and other studies). In such cases, measures of disequilibria permit detailed analysis of migration into the hybrid zone and assortative mating between the hybridizing taxa (ASMUSSEN et al. 1989 ; ARNOLD 1993 ). In general, strong disequilibria have been observed in hybrid zones studied, and these are consistent across loci, indicating that the evolutionary dynamics of the disequilibria are being driven by the mating patterns during hybridization as opposed to more locus-specific forces. For example, ASMUSSEN et al. 1987 analyzed the data of LAMB and AVISE 1986 to show that, in a zone of hybridization between the tree frogs Hyla cineria and H. gratiosa, cytonuclear disequilibria were consistent across five allozyme loci. All five loci showed (a) strong association between the species-specific allozyme alleles and the corresponding mitochondrial (mt)DNA haplotype, indicating that hybridization was generally limited, and (b) strong association of the heterozygotes at the allozyme loci with the H. gratiosa mtDNA haplotype, indicating unidirectional hybridization between H. cinerea males and H. gratiosa females (ARNOLD 1993 ).

Like linkage disequilibrium between nuclear loci, cytonuclear disequilibria are also affected by genetic drift. Thus, some linkage disequilibrium (HILL and ROBERTSON 1968 ; HILL 1981 ) and cytonuclear disequilibrium (FU and ARNOLD 1992 ) can be expected in any population regardless of other evolutionary forces. This can be intuitively understood by considering that a finite population contains only a sample of all possible combinations of alleles, such that associations will occur merely by chance. In contrast to the situation in hybrid zones, such stochastically created disequilibrium would be randomly distributed over loci, but with a variance that increases with decreasing effective population size (since allelic designations are arbitrary in nonhybrid populations, the average disequilibrium is meaningless). Such stochastically created linkage disequilibria have been demonstrated for nuclear loci. For example, in Scots pine (Pinus sylvestris), MUONA and SZMIDT 1984 demonstrated that the most recently founded population in their study exhibited the highest variance of D among locus pairs, suggesting that this population had experienced a recent bottleneck.

Because many different factors can create cytonuclear associations, it can be difficult, in the absence of external information, to determine which of the several forces are at work in any observed instance of cytonuclear disequilibrium (SCRIBNER et al. 1999 ). Thus, the study of nonhybrid populations may provide a useful baseline from which to interpret data from hybrid zones. This study reports measures of cytonuclear disequilibrium in a natural, nonhybrid population of ponderosa pine, for which maternally, paternally, and biparentally inherited marker loci were available from prior studies. We find that observed cytonuclear disequilibria are consistent with previous studies of seed and pollen dispersal within the population (LATTA et al. 1998 ).

	METHODS

TOP ABSTRACT METHODS RESULTS DISCUSSION LITERATURE CITED

Study population:
The study population consists of 217 permanently marked trees in a natural all-aged stand of ponderosa pine (P. ponderosa Laws. var. scopulorum Engelm., Pinaceae). The population is located at 1740 m above sea level on a south-facing slope near Boulder, Colorado, and has been the site of ongoing research for 20 years (e.g., LINHART et al. 1981 ; MITTON et al. 1981 ; LINHART and MITTON 1985 ; LINHART 1988 ; LATTA et al. 1998 ). The site lies on the eastern edge of the distribution of ponderosa pine, near the pronounced transition between the montane forest of the Rocky Mountain foothills and the grassland ecosystems of the western great plains. Since the European settlement of the region, with attendant fire suppression, ponderosa pine has been expanding eastward onto the prairie (FARRIS and MITTON 1984 ; VEBLEN and LORENZ 1990 ; MAST et al. 1998 ). As a result, the density of juvenile trees in formerly sparse stands has increased dramatically, and isolated trees now occur up to several kilometers east of the former ecotone.

The genetic consequences of this movement may include founding events on a very local scale. The spatial arrangement of genetic markers in this population has been described in LINHART et al. 1981 for nuclear allozymes and by LATTA et al. 1998 for organellar DNA. Briefly, the population has been surveyed for 11 polymorphic allozyme loci (two to four alleles per locus), which show mild, but significant, structuring among six demes within the stand (LINHART et al. 1981 ). Selfing rates in this population are low (t = 0.96; MITTON et al. 1981 ). Maternally inherited mtDNA haplotype frequencies (length polymorphism in the second intron of nad1) are 0.837 and 0.163 for the longer and shorter haplotypes, respectively. Spatial autocorrelation revealed significant and pronounced spatial structuring of the mtDNA haplotypes, which occur in small patches of 6 m diameter on average (P < 0.05; LATTA et al. 1998 ). Moreover, pairs of individuals that have the same haplotype and that occur in the same patch are related on average as half sibs on the basis of similarity of allozyme genotypes (r = 0.266). This indicates that these patches represent maternally related groups that occur near one another because of limited seed dispersal. The paternally inherited cytoplasmic (cp)DNA also exhibits length variation in the spacer region between trnG and trnR, with haplotype frequencies of 0.615 and 0.385 for the longer and shorter fragments. In contrast to the mtDNA, no spatial patterning of cpDNA haplotypes was observed (P > 0.25; LATTA et al. 1998 ). Such lack of spatial structuring is consistent with direct estimates of widespread pollen dispersal in this population. However, FARRIS and MITTON 1984 documented that isolated trees recently established on the prairie have a higher selfing rate than do trees in stands of high density. Thus, trees that are sufficiently isolated by distance from others may represent founder events of both the maternal and paternal lineages.

Cytonuclear disequilibria:
Because trees in the stand are permanently marked, the composite three-genome genotype of each tree could be reconstructed for 182 of the original 217 trees on the site using the organellar haplotypes and allozyme genotypes determined in the above studies. This represents a complete census of surviving trees on the site. We analyzed these data for disequilibria (i) between allozymes and maternally inherited mtDNA, (ii) between allozymes and paternally inherited cpDNA, (iii) between allozymes and the joint mtDNA-cpDNA cytotype, (iv) between the two organellar genomes, and (v) among nuclear allozyme loci. Methods for the estimation of true three-way disequilibria (i.e., nuclear-mtDNA-cpDNA; SCHNABEL and ASMUSSEN 1989 ) are not yet available, and so these were not calculated. Analysis followed ASMUSSEN and BASTEN 1994 and used the computer program kindly provided by C. Basten. In what follows, we use D to refer to any disequilibrium and follow the notation of SCHNABEL and ASMUSSEN 1989 to refer to specific types of disequilibrium (e.g., D_A/M to refer to the allelic association between allele A and haplotype M). Disequilibria are defined as the difference between the frequency of a cytonuclear genotype and the product of the frequencies of the relevant organellar haplotype and nuclear allele or genotype (e.g., D_A/M = freq[A/M] - freq[A]*freq[M]).

We proceeded in three steps. First, we tested each locus-organelle combination for significant overall departures from random association. Because exact tests are computationally impractical for overall associations (i.e., for all allele-haplotype and genotype-haplotype combinations combined), we employed the Monte Carlo approach of BASTEN and ASMUSSEN 1997 . Second, for those loci that approached statistical significance, we examined each allele-haplotype and genotype-haplotype combination individually, calculating the exact test of significance and the normalized disequilibrium, D* (D* is the disequilibrium D divided by its maximum possible value; ASMUSSEN and BASTEN 1996 ). We then identified the particular genotype-haplotype or allele-haplotype combination(s) that departed from expectation and mapped the physical locations of trees with this combination. Finally, the variance of the disequilibrium across nuclear loci was computed as D²/n (WEIR and HILL 1980 ). We used the normalized disequilibria (D*) to calculate the variance, because these are more directly comparable across loci. All loci were treated as diallelic by pooling all alleles but the most common into a single composite allele. This has two purposes. First, it gives a single D* value for each locus, whereas multiple alleles would have multiple values of D*. Second, it produces the most intermediate allele frequencies, which maximizes the bounds on D (ASMUSSEN and BASTEN 1996 ), thus giving the greatest resolution of D*.

	RESULTS

TOP ABSTRACT METHODS RESULTS DISCUSSION LITERATURE CITED

Linkage disequilibrium was not significant between any pair of nuclear loci (² < 3.84 for all pairs). In addition, there was no association between mtDNA haplotype and cpDNA haplotype (² = 0.016, not significant).

Table 1 lists the significance (P values from the Monte Carlo approximation) of overall association between alleles or genotypes at each nuclear locus and mtDNA, cpDNA, and joint mtDNA-cpDNA cytotype. This represents the test for overall nonrandom association between genotypes or alleles and haplotypes. For most loci, disequilibria are not statistically different from zero. However, 13 significant associations (P < 0.05) are observed where only 3 are expected by random chance. Three loci show associations with mtDNA (Fe, Got, and Udp), while two are significantly associated with cpDNA (Per and Sdh). Significant associations between nuclear loci and the joint mtDNA-cpDNA cytotype were observed only for loci that showed a significant association with either cpDNA or mtDNA alone (Table 1). Thus the joint disequilibrium measures repeat the information contained in the two-way measures and are not discussed further here.

For all loci showing a significant overall association between alleles or genotypes and one of the organelles, the association could be attributed to one particular nuclear genotype or allele (Table 2). For example, the association between Fe genotypes and mtDNA (P < 0.02; Table 1) was entirely attributable to the fact that there were significantly more BB homozygotes carrying mtDNA haplotype m (and fewer with haplotype M) than expected under a random association (P < 0.0025; Table 2). None of the other Fe genotypes showed significant association with the mtDNA haplotypes, although there is no a priori reason why they should not. For all loci, the difference between observed and expected numbers of individuals exhibiting a particular cytonuclear genotype (O - E) was small (Table 2). In the case of the Fe-mtDNA association, we observed only 8 more individuals with the BB-m genotype than the 11 that are expected under random association. Similar results were observed for the allelic associations (not shown).

The spatial arrangement of the organellar haplotypes is presented in Fig 1, a and d. Mitochondrial haplotypes show significant clustering in space (P < 0.05), while cpDNA haplotypes do not (LATTA et al. 1998 ). The mtDNA patches were identified by spatial clustering of the less common of the two mtDNA haplotypes and can be ascribed to limited dispersal of seeds. Those allozyme-mtDNA combinations that are more common than expected under random association correspond closely to matrilineal clusters identified previously (Fig 1B and Fig C). In contrast to mtDNA, however, chloroplast haplotypes show no spatial clustering (Fig 1D), which is attributable to widespread movement of the pollen (for which there is direct evidence; cf. LATTA et al. 1998 ). Similarly, no spatial clustering is observed for the nuclear-cpDNA joint genotypes (Fig 1E and Fig F).

View larger version (38K): In this window In a new window Download PPT slide   Figure 1. Spatial arrangement of genotypes at loci showing significant cytonuclear disequilibrium. (a and d) Spatial arrangement of organellar haplotypes. (b, c, e, and f) Solid triangles indicate the location of individuals with joint nuclear-organellar genotypes that are present in excess of random expectation (cf. Table 2).

Under random genetic drift, some loci will drift to a positive and some to a negative association between allele A and the organellar haplotype M (where allelic labels are arbitrarily assigned). Thus, the variance of D across loci will be inversely related to the effective population size (FU and ARNOLD 1992 ). Both mtDNA and cpDNA show a significant variance of D across allozyme loci (Table 3). Sampling variance of Var(D) for nuclear loci is approximately 1/n (HILL 1981 ) or in this case 0.005, and Var(D) for cytoplasmic disequilibria is typically about twice that for nuclear loci (FU and ARNOLD 1992 ). These values fall outside the 95% confidence limits of Var(D) for both mtDNA and cpDNA. Although the maternally inherited mtDNA shows a higher variance of D than the paternally inherited cpDNA, this difference is not significant (F_10,10 = 2.03, P > 0.1). Data on the male and female reproductive output of each tree in the stand are available for the last 20 years (LINHART and MITTON 1985 ). These data indicate that male and female reproduction is concentrated in a few individuals that account for the majority of the reproductive output. This is only slightly more pronounced for seed than pollen output—the coefficient of variation for female cone production is somewhat greater than that for pollen cone production, but only by 50%.

	DISCUSSION

TOP ABSTRACT METHODS RESULTS DISCUSSION LITERATURE CITED

Disequilibrium and drift:
The nonrandom association of nuclear alleles or genotypes with organellar haplotypes can arise from a number of evolutionary forces that fall into three categories (ASMUSSEN et al. 1987 ): (i) nonrandom mating, including patterns of admixture, migration, and hybridization; (ii) interactive fitness effects across genomes; and (iii) the historical sampling of gametes in finite populations (drift). Given the many factors that may give rise to cytonuclear disequilibrium, SCRIBNER et al. 1999 caution that in the absence of independent information it can be difficult to ascribe a particular evolutionary process to observed patterns. While cytonuclear disequilibria have received the most attention in the context of migration, primarily situations where independent information suggests that hybridization is occurring (ARNOLD 1993 ), our study suggests that cytonuclear disequilibrium can be present in nonhybridizing populations through the action of genetic drift.

Significant associations between nuclear and organellar polymorphisms are observed in this population for both the mitochondrial and chloroplast genomes (Table 1). Five nuclear loci show associations that are significant and for which the normalized disequilibrium D* is quite large (Table 2). However, despite large normalized disequilibria, the absolute magnitude of the difference between the observed and expected number of individuals carrying each combination is quite small, typically 5–10 individuals. Such small deviations (though statistically significant) are consistent with the action of genetic drift, which in this case we ascribe to very localized founder events (see below). Since many populations fluctuate over time, such drift-induced cytonuclear disequilibria are likely to occur in a wide variety of species.

A common feature of cytonuclear disequilibria caused by migration patterns in hybrid zones is consistency in both the magnitude and direction of disequilibrium across all nuclear loci. For the Hyla example (LAMB and AVISE 1986 ; ASMUSSEN et al. 1987 ; ARNOLD 1993 ), all loci showed the same pattern of association with mtDNA haplotypes, indicating that disequilibrium was being created by mating patterns that act in a consistent manner on all loci. By contrast, in this nonhybrid population of ponderosa pine, estimates of disequilibrium are highly heterogeneous across loci (Table 3). The bulk of loci show no associations, while a few loci show significant disequilibrium (Table 1). Moreover, the associations are not consistent across mtDNA and cpDNA. Those loci that are significantly associated with haplotypes at the mitochondrial genome are not those that are associated with cpDNA, and vice versa. If the disequilibria were being created through nonrandom mating patterns, we would predict the same nuclear loci to associate with the paternally inherited cpDNA as with the maternally inherited mtDNA (i.e., cpDNA would be associated with the allele contributed by the pollen parent). The absence of such patterns in this population argues against a pattern of nonrandom mating. This view is strengthened by the lack of association between the two organelles since mt-cp disequilibrium is indicative of assortative mating (SCHNABEL and ASMUSSEN 1989 ). Instead, the observed pattern seems likely to be the result of locus-specific forces such as selection or of a stochastic process such as drift.

The spatial patterns have been superimposed on the disequilibria by the differential dispersal abilities of seed and pollen. The clusters of mtDNA haplotypes presumably reflect recent founding of that patch by a maternal tree that had that haplotype. If the founding tree also carried a relatively uncommon allozyme allele, then the nuclear allele and mtDNA haplotype would become associated because both are present in the matrilineal family that descended from that individual. Moreover, these individuals would be clustered in space due to limited seed movement. Thus the reduction of N_e that produced the drift seems likely to have been due to very localized founder events surrounding the trees that were reproductive at the time the population began to expand and increase in density (100 years ago; VEBLEN and LORENZ 1990 ). These spatial patterns of the joint genotypes are consistent with the hypothesis that drift is the causal agent in this case.

Very few data exist for cytonuclear disequilibria involving a paternally inherited organelle. The data of LI 1995 for jack pine (P. banksiana) were analyzed by BASTEN and ASMUSSEN 1997 but did not reveal significant associations. In this population, by contrast, cpDNA show levels of D similar to mtDNA. The existence of significant associations between allozymes and cpDNA in ponderosa pine suggests that founding events occurred through the paternal lines in much the same way as through the maternal lineages. For nuclear loci, the relationship between effective population size, N_e, and linkage disequilibrium was defined by HILL and ROBERTSON 1968 and HILL 1981 . This makes it theoretically possible to estimate N_e from data on linkage disequilibrium, although such estimates are prone to high sampling error (WAPLES 1991 ). Although no explicit method of inferring effective population size from cytonuclear disequilibria has been described, simulations by FU and ARNOLD 1992 gave Var(D) of 0.065 for N_e = 10 and 0.013 for N_e = 50. The observed value for cpDNA (0.038; Table 3) falls within this range. Since the census size of the population is 217, the effective number of pollen parents appears to be between one-twentieth and one-quarter of the census size.

Thus, the number of pollen founders in this population appears to have been almost as limited as the number of seed founders. We might predict that fewer seed parents contribute to the founding of a new population than pollen parents because of the greater potential for pollen immigration and thus a lower effective population size of females than males. Although chloroplasts do show slightly lower values of normalized disequilibrium and its variance (0.038) than do mitochondria (0.078), this difference is not significant, such that the effective population size of males appears to be at most only slightly greater than that of females. Data on the reproductive output of each tree have been collected over the last 20 years (cf. LINHART and MITTON 1985 ) and accord with the genetic findings. Although there is a higher variance of reproductive output through seed than through pollen (and thus presumably a lower effective number of females), there is still a high variance of reproductive output through pollen (Table 3), and thus the effective number of males in this population is also limited. This accords well with the finding of FARRIS and MITTON 1984 that isolated trees colonizing a new habitat have higher selfing rates than trees in established stands and thus appear to experience a limited number of pollen parents.

Alternatives to drift:
Cytonuclear disequilibria can result from migration primarily where immigrants themselves exhibit cytonuclear associations, or where seed and pollen immigrants are genetically differentiated at nuclear or paternally inherited organellar markers (ORIVE and ASMUSSEN 2000 ). This would most likely happen if there are multiple sources of immigrants that are genetically differentiated for both nuclear and organellar loci. Although conifer populations typically exhibit extensive gene flow (HAMRICK and GODT 1990 ), it is unlikely that immigrant seeds or pollen contain the necessary disequilibrium. Although immigrants might arrive from multiple source populations, there is typically little genetic differentiation among populations of conifers in either nuclear (HAMRICK and GODT 1990 ) or cpDNA markers (DONG and WAGNER 1994 ; LATTA and MITTON 1997 ), although exceptions do exist (e.g., MITTON et al. 1980 ). Indeed, it is precisely because gene flow is extensive in many conifers that immigrants from multiple populations are likely to be undifferentiated and therefore do not show disequilibria in the migrant pool.

Alternatively, selective scenarios can be imagined that would produce the observed level of differentiation, but these, too, require a number of assumptions. In several hybrid zones, fitness interaction has been demonstrated between the nuclear and organellar genomes (e.g., CLARK and LYCKEGAARD 1988 ; CRUZAN and ARNOLD 1994 ; BREEUWER and WERREN 1995 ; BURKE et al. 1998 ), where nuclear loci may serve as markers of larger chromosomal segments. This tends to be more apparent in distant crosses (i.e., hybrids) than within single populations (CLARK and LYCKEGAARD 1988 ). In a nonhybridizing population of conifers, linkage disequilibrium among nuclear loci is typically quite low, such that the association between marker loci and adjacent genes under selection will be weak and variable (STRAUSS et al. 1992 ). Therefore, a selective explanation for the associations seen here must posit that selection acts directly on the interaction between the organellar DNA and the specific allozyme locus involved (rather than the allozymes marking a region of the chromosome that is under selection). This is a much more restrictive case.

We therefore interpret the disequilibria as consistent with existing knowledge of the population (LATTA et al. 1998 ). They are consistent with neutral genetic drift without the need to invoke additional evolutionary processes, such as migration or selection. Moreover, as we have argued, several assumptions must be met before selection or migration can produce the disequilibria observed here. Therefore, while such processes may be acting in this population, the observed levels of cytonuclear disequilibrium do not provide evidence of their action. A far more parsimonious explanation posits that drift has occurred through a limited number of maternal and paternal parents giving rise to the extant population and that this has created occasional nonrandom associations between nuclear and organellar genotypes.

	ACKNOWLEDGMENTS

We thank C. J. Basten for his generous help in providing the computer program that calculated the estimates of disequilibrium. Numerous field assistants helped with the collection of data. The final manuscript was greatly improved by the comments of M. A. Asmussen and two anonymous reviewers. Collection of the allozyme and reproductive data was funded by National Science Foundation grants BSR 8918478 and DEB 9120065 to Y.B.L. and J.B.M., and U.S. Department of Agriculture grant 95-37101 to Y.B.L. Organellar data collection was funded by a National Science Foundation Doctoral Dissertation Improvement Grant to R.G.L.

Manuscript received November 23, 2000; Accepted for publication March 9, 2001.

	LITERATURE CITED

TOP ABSTRACT METHODS RESULTS DISCUSSION LITERATURE CITED

ARNOLD, J., 1993  Cytonuclear disequilibria and hybrid zones. Annu. Rev. Ecol. Syst. 24:521-554.

ASMUSSEN, M. A. and J. ARNOLD, 1991  The effects of admixture and population subdivision on cytonuclear disequilibrium. Theor. Popul. Biol. 39:273-300[Medline].

ASMUSSEN, M. A. and C. J. BASTEN, 1994  Sampling theory for cytonuclear disequilibria. Genetics 138:1351-1363[Abstract].

ASMUSSEN, M. A. and C. J. BASTEN, 1996  Constraints and normalized measures for cytonuclear disequilibria. Heredity 76:207-214.

ASMUSSEN, M. A. and M. E. ORIVE, 2000  The effects of pollen and seed migration on nuclear-dicytoplasmic systems. I. Nonrandom associations and equilibrium structure with both maternal and paternal cytoplasmic inheritance. Genetics 155:813-831[Abstract/Free Full Text].

ASMUSSEN, M. A., J. ARNOLD, and J. C. AVISE, 1987  Definition and properties of disequilibrium statistics for associations between nuclear and cytoplasmic genotypes. Genetics 115:755-768[Abstract/Free Full Text].

ASMUSSEN, M. A., J. ARNOLD, and J. C. AVISE, 1989  The effects of assortative mating and migration on cytonuclear associations in hybrid zones. Genetics 122:923-934[Abstract/Free Full Text].

BASTEN, C. J. and M. A. ASMUSSEN, 1997  The exact test for cytonuclear disequilibria. Genetics 146:1165-1171[Abstract].

BREEUWER, J. A. J. and J. H. WERREN, 1995  Hybrid breakdown between two haplodiploid species: the role of nuclear genes and cytoplasmic genes. Evolution 49:705-717.

BURKE, J. M., T. J. VOSS, and M. L. ARNOLD, 1998  Genetic interactions and natural selection in Louisiana Iris hybrids. Evolution 52:1304-1310.

CLARK, A. G. and E. M. S. LYCKEGAARD, 1988  Natural selection with nuclear and cytoplasmic transmission. III. Joint analysis of segregation and mtDNA in Drosophila melanogaster. Genetics 118:471-481[Abstract/Free Full Text].

CRUZAN, M. B. and M. L. ARNOLD, 1993  Ecological and genetic associations in an Iris hybrid zone. Evolution 47:1432-1445.

CRUZAN, M. B. and M. L. ARNOLD, 1994  Assortative mating and natural selection in an Iris hybrid zone. Evolution 48:1946-1958.

DONG, J. and D. B. WAGNER, 1994  Paternally inherited chloroplast polymorphism in Pinus: estimation of diversity and population subdivision and tests of disequilibrium with a maternally inherited mitochondrial polymorphism. Genetics 136:1187-1194[Abstract].

EDMANDS, S. and R. S. BURTON, 1999  Cytochrome C oxydase activity in interpopulation hybrids of a marine copepod: a test of nuclear-nuclear or nuclear-cytoplasmic coadaptation. Evolution 53:1972-1978.

FARRIS, M. A. and J. B. MITTON, 1984  Population density, outcrossing rate and heterozygote superiority in ponderosa pine. Evolution 38:1151-1154.

FU, Y. X. and J. ARNOLD, 1992  Dynamics of cytonuclear disequilibria in finite populations and comparison with a two-locus nuclear system. Theor. Popul. Biol. 41:1-25[Medline].

HAMRICK, J. L., and M. J. W. GODT, 1990 Allozyme diversity in plant species, pp. 43–63 in Plant Population Genetics, Breeding and Genetic Resources, edited by A. H. D. BROWN, M. T. CLEGG, A. L. KAHLER and B. S. WEIR. Sinauer, Sunderland, MA.

HILL, W. G., 1981  Estimation of effective population size from data on linkage disequilibrium. Genet. Res. 38:209-216.

HILL, W. G. and A. ROBERTSON, 1968  Linkage disequilibrium in finite populations. Theor. Appl. Genet. 38:226-231.

LAMB, T. and J. C. AVISE, 1986  Directional introgression of mitochondrial DNA in a hybrid population of tree frogs: the influence of mating behavior. Proc. Natl. Acad. Sci. USA 83:2526-2530[Abstract/Free Full Text].

LATTA, R. G. and J. B. MITTON, 1997  A comparison of population structure across four classes of gene marker in limber pine. Genetics 146:1153-1163[Abstract].

LATTA, R. G., Y. B. LINHART, D. FLECK, and M. ELLIOT, 1998  Direct and indirect estimates of seed vs. pollen movement within a population of ponderosa pine. Evolution 52:61-67.

LI, T., 1995 Cytonuclear population genetic structure of jack pine (Pinus banksiana Lamb.) and lodgepole pine (Pinus contorta Dougl.). Ph.D. Dissertation, University of Kentucky, Lexington, KY.

LINHART, Y. B., 1988 Ecological and evolutionary studies of ponderosa pine in the rocky mountains, pp. 77–89 in Ponderosa Pine: The Species and its Management, edited by D. M. BAUMGARTNER and J. E. LOTAN. Cooperative Extension, Washington State University, Pullman, WA.

LINHART, Y. B. and J. B. MITTON, 1985  Relationships among reproduction, growth rate, and protein heterozygosity in ponderosa pine. Am. J. Bot. 72:181-184.

LINHART, Y. B., J. B. MITTON, and K. B. STURGEON, 1981  Genetic variation in space and time in a population of ponderosa pine. Heredity 46:407-426.

LIU, H-P., J. B. MITTON, and S-K. WU, 1996  Paternal mitochondrial DNA differentiation far exceeds maternal mitochondrial DNA and allozyme differentiation in the freshwater mussel, Anodonta grandis grandis. Evolution 50:952-957.

MAST, J. M., T. T. VEBLEN, and Y. B. LINHART, 1998  Disturbance and climatic influences on age structure of ponderosa pine at the pine grassland ecotone. Colorado Front Range. J. Biogeogr. 25:743-755.

MCCAULEY, D. E., 1994  Contrasting the distribution of chloroplast DNA and allozyme polymorphism among local populations of Silene alba: implications for studies of gene flow in plants. Proc. Natl. Acad. Sci. USA 91:8127-8131[Abstract/Free Full Text].

MITTON, J. B., K. B. STURGEON, and M. L. DAVIS, 1980  Genetic differentiation in ponderosa pine along a steep elevational transect. Silvae Genet. 29:100-103.

MITTON, J. B., Y. B. LINHART, M. L. DAVIS, and K. B. STURGEON, 1981  Estimation of outcrossing in ponderosa pine, Pinus ponderosa, Laws, from patterns of segregation in protein polymorphisms and from frequencies of albino seedlings. Silvae Genet. 30:117-121.

MOGENSON, H. L., 1996  The hows and whys of cytoplasmic inheritance in seed plants. Am. J. Bot. 83:383-404.

MUONA, O., and A. E. SZMIDT, 1984 A multilocus study of natural populations of Pinus sylvestris, pp. 226–240 in Population Genetics in Forestry, edited by H. R. GREGORIUS. Springer-Verlag, Berlin.

ORIVE, M. E. and M. A. ASMUSSEN, 2000  The effects of pollen and seed migration on nuclear-dicytoplasmic systems. II. A new method for estimating plant gene flow from joint nuclear-cytoplasmic data. Genetics 155:833-854[Abstract/Free Full Text].

PAIGE, K. N., W. C. CAPMAN, and P. JENNETTEN, 1991  Mitochondrial inheritance patterns across a cottonwood hybrid zone: cytonuclear disequilibria and hybrid zone dynamics. Evolution 45:1360-1369.

SCHNABEL, A. and M. A. ASMUSSEN, 1989  Definition and properties of disequilibria within nuclear-mitochondrial-chloroplast and other nuclear-dicytoplasmic systems. Genetics 123:199-215[Abstract/Free Full Text].

SCRIBNER, K. T. and J. C. AVISE, 1994  Population cage experiments with a vertebrate: the temporal demography and cytonuclear genetics of hybridization in Gambusia fishes. Evolution 48:155-171.

SCRIBNER, K. T., S. DATTA, J. ARNOLD, and J. C. AVISE, 1999  Empirical evaluation of cytonuclear models incorporating genetic drift, and tests for neutrality of mtDNA variants: data from experimental Gambusia hybrid zones. Genetica 105:101-108.

STRAUSS, S. H., R. LANDE, and G. NAMKOONG, 1992  Limitations of molecular-marker aided selection in forest tree breeding. Can. J. For. Res. 22:1050-1061.

VEBLEN, T. T., and D. C. LORENZ, 1990 The Colorado Front Range: A Century of Ecological Change. University of Utah Press, Salt Lake City, UT.

WAGNER, D. B., 1992  Nuclear, chloroplast and mitochondrial DNA polymorphisms as biochemical markers in population genetic analyses of forest trees. New For. 6:373-390.

WAPLES, R. S., 1991  Genetic methods for estimating the effective size of cetacean populations. Rep. Int. Whaling Comm. 13:279-300.

WEIR, B. S. and W. G. HILL, 1980  Effect of mating structure on variation in linkage disequilibrium. Genetics 95:477-488[Abstract/Free Full Text].

ZOUROS, E., A. O. BALL, C. SAAVEDRA, and K. R. FREEMAN, 1992  Direct evidence for extensive paternal mitochondrial DNA inheritance in the marine mussel Mytilus. Nature 359:412-414[Medline].

This article has been cited by other articles:

				K. J. Monsen, B. M. Honchak, S. E. Locke, and M. A. Peterson Cytonuclear Disequilibrium in Chrysochus Hybrids Is Not Due to Patterns of Mate Choice J. Hered., July 9, 2007; (2007) esm039v1. [Abstract] [Full Text] [PDF]

				M. B. Hamilton and J. R. Miller Comparing Relative Rates of Pollen and Seed Gene Flow in the Island Model Using Nuclear and Organelle Measures of Population Structure Genetics, December 1, 2002; 162(4): 1897 - 1909. [Abstract] [Full Text] [PDF]