Abstract
Cinnamomum insularimontanum Hayata (Lauraceae) is an insect-pollinated, broad-leaved evergreen tree with bird-dispersed seeds. We used allozyme loci, Wright's fixation index, spatial autocorrelation statistics (Moran's I), and coancestry measures to examine changes in genetic structure among four age-classes within a recently founded study population (60Ã100âm area) in southern Korea. There were no significant differences in expected heterozygosity among age classes. However, significant genetic differentiation among age classes was detected (P<0.0001). Fixation indices within age classes showed significant deficits of observed heterozygosity, which may be caused by partial selfing. The homogeneity of genetic structure among four age-classes may reflect similar spatial patterns of seed immigration from surrounding populations occurring year after year. Finally, the average Moran's I and coancestry estimates indicated essentially random spatial distributions of alleles for each of the four age-classes and between seedlings and 2â4âyear juveniles vs adult trees. These findings are very similar to those observed in the same study area for another member of the Lauraceae, Neolitsea sericea, which has a very similar life history and ecological characteristics (ie, bird-dispersed fruits, insect pollination, and a similar age structure). Together, these results suggest that the fleshy drupes of lauraceous species represent an adaptation to aid in the independent dispersal of seed by birds, which in turn may increase the genetic diversity of founders colonizing new habitats.
Similar content being viewed by others
Introduction
The spatial genetic structure among sexually reproduced individuals within plant populations is determined in large part by gene movement via pollen and seed dispersal, that is, gene movement patterns may be a key factor in the internal spatial genetic structuring of plant populations. Although limits to pollen and seed dispersal distances can both contribute to the buildup of genetic isolation-by-distance, the two do not have the same effects. First, in seed plants, pollen dispersal is solely migration, whereas seed or fruit dispersal could result in a migration or a colonization event (Hamrick and Nason, 1996), a distinction that is particularly important in metapopulations (Wade and McCauley, 1988). Second, gene movement in seed plants is sequential (a two-step process, via pollen then by seed), and the development of genetic structure within populations may be more strongly influenced by seed dispersal than by pollen movement. On the one hand, even if pollen disperses long distances, limited seed movement will result in genetically distinct seed shadows among maternal plants and the spatial clustering of half-sibs (eg, Berg and Hamrick, 1994). If pollen dispersal is also restricted, this will result in inbreeding and could eventually cause the buildup of more intense genetic structure within populations over generations (Sokal and Wartenberg, 1983; Hamrick and Nason, 1996). In contrast, if seeds are widely and independently dispersed within populations, only weak or ânear-randomâ spatial genetic structure will result (eg, Dewey and Heywood, 1988; Loiselle et al, 1995; Chung et al, 2000). Even if pollen disperses only short distances, long-distance seed dispersal will effectively randomize the spatial distribution of genetic variation.
Long-lived woody plants allow the study of both long-term temporal and spatial variation in population structure. Age-structured data may allow inferences on historical reproductive episodes, and the extent and patterns of seed dispersal. Recently, studies of genetic structure in terms of demography have revealed sometimes striking changes in genetic structure owing to such demographic processes as changes in density (eg, thinning processes independent of genotypes) and the effects of past major reproductive events by different source trees (eg, Epperson and Alvarez-Buylla, 1997; Schnabel et al, 1998; Parker et al, 2001). Nonetheless, there are still few studies on the spatial distribution of genetic variation in different age or size classes of woody plants so that stage-class-specific changes in fine-scale spatial genetic structure and associated causal processes are not well understood.
Cinnamomum insularimontanum Hyata (Lauraceae) exhibits a metapopulation structure with important implications for stage-class-specific genetic structure. A number of field observations have been made in Hakdongri Reserve on Geoje Island in southern Korea to obtain information on where and how populations of the species get established (MY Chung and MG Chung, personal observation). Gap populations in early ecological stages are commonly found near older mature populations of C. insularimontanum in Hakdongri Reserve. These early seral stages appeared to be characterized by high light environments and many seedlings and juveniles but few or no adults. Older populations, in contrast, occur in closed canopy broad-leaved evergreen forests, are at a successional stage in which there is little or no regeneration, and thus seem to be in the process of becoming locally extinct. As a result, C. insularimontanum would exist only for a few generations at any locality and has to establish new populations continually. Seedlings and juveniles (most with ages <5 years) are commonly found scattered several hundred metres from possible maternal trees in natural habitats on a few islands in southern Korea, indicating that seeds can be dispersed long distances, probably by birds. Many bird species in northeastern Asia consume the fleshy fruits of the members of Lauraceae (Neolitsea spp, Litsea spp, and Persea spp) along coastal areas (Kitamura and Murata, 1987), and have been observed feeding on C. insularimontanum in southern Korea (MY Chung and MG Chung, personal observation). Since the fruits of these species are often consumed and dispersed singly or a few at a time, this type of seed dispersal may serve to weaken spatial genetic structuring within and among populations (Hamrick and Loveless, 1986).
One important question arising from the metapopulation structure of C. insularimontanum concerns how the founding of populations influences internal spatial genetic structure across age classes. If newly founded populations are populated by the long-distance dispersal of immigrants from adjacent source populations year after year, then the internal spatial genetic structure of young populations is expected to be weak both within- and between-age classes. On the other hand, if population expansion results from local seed production and recruitment in the absence of local adaptational differentiation, then fine-scale spatial genetic structure among seedlings and young individuals is expected to be significant and stronger than among older individuals. To test for evidence of these alternative demographic and genetic processes, in this study, multilocus allozyme genotypes were sampled and mapped from a young local population of C. insularimontanum occurring in the undisturbed Hakdongri Reserve on Geoje Island in southern Korea. Wright's fixation index and spatial autocorrelation statistics were calculated and analyzed to examine the dynamics of genetic and spatial genetic structure across age classes in the study population. Finally, the results of the present study were compared with other insect-pollinated woody species with bird-dispersed seeds, in particular Neolitsea sericea for which age-class-specific spatial genetic structure data are available for the same study area (Chung et al, 2000).
Materials and methods
Study species
C. insularimontanum is distributed in southern China and Japan (western region to Chiba Pref. of Honshu, Shikoku, Kyushu, and the Ryukyu Islands), but on the Korean Peninsula it occurs only in a few isolated populations on the southern coast. These populations have areas ranging from ca 500âm2 to several hectares on talus slopes with well-drained, rocky and humus soils (MY Chung and MG Chung, personal observation). C. insularimontanum has a mass-flowering pattern from mid-May to mid-June, and diverse insect pollinators such as bees, wasps, flies, and beetles visit its perfect flowers (Yumoto, 1987). Adults typically produce hundreds of single-seeded fruits becoming fully mature about 6 months after flowering. During winter, the black drupes (diameter ca 12âmm) are consumed by birds one or a few at a time and dispersed away from the maternal trees (MY Chung and MG Chung, personal observation).
Population samples
All 423 individuals were mapped and leaf samples were collected within a 60âmÃ100âm area (altitude 15â40âm above sea level, 6% grade; Figure 1) in a broad-leaved evergreen forest with a low density of Pinus thunbergii and other deciduous trees (eg, oaks and maples). The study population has no recorded history of fire disturbance, there is no evidence of trees having been planted, and eight other long-lived natural evergreen woody plants coexist at the site. Ages of individuals were determined by counting annual rings (eg, based on cores taken by an increment borer for adults and thin transverse sections of stems and leaf scars for juveniles), and individuals were classified into four age-classes: age class I (N=85), seedlings and 2â4âyear juveniles; age class II (N=122), 5â10âyear juveniles; age class III (N=134), 11â25âyear juveniles; and age class IV (N=82), ca 30â40âyear juveniles and nine adults. Since C. insularimontanum is an early successional species that fruits annually upon reaching maturity, adults were identified as any individuals observed fruiting over a 2-year period (1998â1999). One leaf per individual was stored at 4°C until protein extraction.
Enzyme extraction and electrophoresis
Leaves were cut into small pieces and crushed with a mortar and pestle. A potassium phosphate extraction buffer (Mitton et al, 1979) was added and the crushed extract was absorbed onto 4âmmÃ6âmm Whatman 3MM chromatography paper wicks, which were stored at â70°C until needed for electrophoretic analysis. Electrophoresis was performed using 11.5% starch gels. Totally, 9 putative loci for C. insularimontanum from five enzyme systems were resolved using two electrophoretic buffer systems. A Poulik buffer system of Soltis et al's (1983) âsystem 6â was used to resolve diaphorase (Dia-1, Dia-2), phosphoglucomutase (Pgm), and triosephosphate isomerase (Tpi-1, Tpi-2). A discontinuous histidineâcitrate buffer system, a modification (Chung and Kang, 1994) of Soltis et al's (1983) âsystem 11â, was used to resolve 6-phosphogluconate dehydrogenase (Pgd-1, Pgd-2) and phosphoglucoisomerase (Pgi-1, Pgi-2). Stain recipes were taken from Soltis et al (1983), except for diaphorase (Cheliak and Pitel, 1984). The genetic basis of allozyme banding patterns was inferred from segregation patterns with reference to typical subunit structure (Weeden and Wendel, 1989; Wendel and Weeden, 1989) and conceptual methods described in Gillet (1998). Putative loci were designated sequentially, with the most anodally migrating isozyme designated 1, the next 2, and so on. Similarly, alleles were designated sequentially with the most anodally migrating alleles designated with superscript a.
Data analysis
For all data analyses, each of the four age-classes was considered separately as a subgroup (eg, Bacilieri et al, 1994; Berg and Hamrick, 1995; Hossaert-McKey et al, 1996). A locus was considered polymorphic if the frequency of the most common allele did not exceed 0.95. A total of 4 genetic diversity parameters were estimated using the program POPGENE (Yeh et al, 1999): percent polymorphic loci (%P); mean number of alleles per locus (A); mean observed heterozygosity (Ho); and mean unbiased estimates of expected heterozygosity per locus (He). Standard errors for Ho and He were calculated over loci. To summarize genetic diversity over age classes, all individuals were pooled into a single population for which these four genetic diversity statistics were calculated. To make comparisons of He across age classes, a complete randomized block ANOVA, with loci as blocks, was conducted because the same loci were sampled from each age class.
For each polymorphic locus in each age class, observed heterozygosity was compared to HardyâWeinberg (HâW) expected values using fixation index (fW) of Wright (1922). The proportion (or probability, P-value) of randomized data sets that gave larger or smaller fW values than that observed was estimated for each locus and age class, based on 3200 randomizations of alleles among individuals within age classes. A Bonferroni adjustment was used to achieve an experiment-wide type I error (α=0.05) for tests of loci and populations. For each age class, a mean fW value across polymorphic loci and a 95% bootstrap confidence interval (CI) was calculated using the program GDA (Lewis and Zaykin, 2001). We also calculated the fixation index (and its confidence limits) for all individuals (fT) by combining age classes into a single, pooled population.
We tested for genetic differentiation among age classes using the exact test of Raymond and Rousset (1995). This test is analogous to Fisher's exact test but uses a Markov chain to explore all potential states of an rÃk contingency table based on r age-classes and k genotypes. This test was conducted using the program ARLEQUIN (Schneider et al, 2000) and 10â000 Markov steps.
An analysis of the spatial distribution of individuals in each age class is helpful for understanding whether self-thinning changes the spatial relationships of individuals, and thus genotypes. To assess the spatial distribution of C. insularimontanum individuals in each age class, Morisita's (1959) index of dispersion (Iδ) was calculated following the methods of Sakai and Oden (1983). If individuals are randomly distributed per plot, Iδ=1; if individuals are distributed in patches, Iδ>1, and if individuals are uniformly distributed (hyperdispersed), Iδ<1. The significance of departures from randomness was assessed by F tests (Morisita, 1959). Further, we used permutation methods (999 times) to test, for each age class, whether the slope of Iδ changes systematically with quadrat size (m2) (Table 2) using the program Permute! Version 3.4 alpha (Casgrain, 2001) to construct a distribution of the slope under the null hypothesis β=0. Under the one-tailed test of the hypothesis that slope of Iδ decreases with increasing quadrat size, we reject the null if there are fewer than 50 random values at least as large as the actual observed β value. Significant level of type I error was corrected for multiple tests by the method of Šidák's (1967). In addition, in order to test the hypothesis that Iδ decays with increasing age, we regressed β for Iδ on age class.
The spatial distributions of allozyme polymorphisms were analyzed separately for age classes IâIV using both Moran's I-statistics as correlations of allele frequencies (Sokal and Oden, 1978) and a pairwise estimate of genetic correlation, fij (âcoancestryâ coefficient; see Kalisz et al, 2001), which measures the correlation between the frequency of a random allele from one individual with that of a random allele from another (Cockerham, 1969). Coancestry coefficient has been used in a number of recent studies (eg, Loiselle et al, 1995; Kalisz et al, 2001; Parker et al, 2001; Chung et al, 2002).
For Moran's I-statistics, only one allele was considered at diallelic loci, because in this case the second allele contributes identical information. For a locus having more than two alleles, all alleles at that locus, regardless of their frequencies, were used for the spatial analyses. However, alleles that were represented by less than five copies (frequencies <2.9, 2.1, 1.9, and 3.0% for the respective age classes) were excluded as non-informative for spatial analysis. Every possible pair of individuals was considered as a join or a connection between two individuals and was assigned to one of the 18 distance classes according to the distance separating the two individuals. Since measures of small-scale autocorrelation more accurately represent spatial structure, a 5âm interval was selected. Moran's I-statistic (Sokal and Oden, 1978) was calculated for each of the 18 distance classes. Each I-value was used to test for significant deviations from the expected values, E(I)=â1/(Nâ1) (Cliff and Ord, 1981), where N is the sample size. A significant positive value of I indicates that the pairs of individuals within a given distance class have more alleles in common than would be expected by chance, whereas a significant negative value indicates that such individuals have fewer alleles in common than expected. Overall significance of individual correlograms was tested using Bonferroni's criteria (Sakai and Oden, 1983). All spatial calculations and statistical analyses were performed using the SAAP program (ver. 4.3) written by D Wartenberg.
The coancestry (fij) was estimated between all pairs of individuals within each age class at a given distance interval and for the total sample (423 individuals), following the methods of Loiselle et al (1995) and Kalisz et al (2001). Mean values of fij were obtained for distance intervals (lags) of 5âm (as done for Moran's I-statistics) by averaging over all pairs of individuals located within that interval. This measure of coancestry should be near zero when there is no genetic correlation between the frequencies of alleles in individuals at the spatial scale of interest, fij>0 when individuals in a given distance class are more closely related than expected by chance, and fij<0 when individuals within a given distance class are less related than expected by chance. Assessment of statistical significance for each fij estimate per given distance was conducted by the randomization procedures described by Kalisz et al (2001).
It is possible that seeds could be produced by adults within the study population and then locally dispersed. In this case, we would expect a significant positive autocorrelation between seedlings and 2â4 year juveniles and adults. To test this hypothesis, an analysis of coancestry between these juveniles and adult trees was conducted using fij and 5âm lags. Failure to reject the null hypothesis was interpreted to mean either that seedlings are immigrants or are produced locally but randomly dispersed. All calculations and simulations for fij were performed using a program developed by J Nason.
Finally, to test whether the slope (β) of the correlograms obtained using Moran's I and fij for each age class was statistically significant, I and fij estimates were permuted (999 times) with respect to upper bound (m) of each distance class using the program Permute! Version 3.4 alpha to construct a distribution of the slope under the null hypothesis β=0. Significant level of type I error was corrected by the method of Šidák (1967).
Results
Genetic diversity and structure in age classes
Of the nine loci examined, seven were polymorphic and reliably scored. The locus Pgi-1 was monomorphic and allele Pgd-1a was detected only in age classes I (allele frequency=0.006) and II (0.008); thus Pgd-1 was deleted for data analysis. Levels of genetic diversity averaged across polymorphic loci were high and similar across age classes ( Table 1): there was no significant difference among age classes for He (F=1.563, P=0.228).
Values of the fixation index (fW), calculated for each polymorphic locus in each age class, varied among both age classes and loci but showed no clear pattern with respect to age classes (data not shown). Linear regression of fW on age class for each polymorphic locus was not significant. Only one (fW=0.337, Pgi-2 in age class II) of the 28 single locus estimates was statistically significant when type I error was adjusted for multiple tests (P=0.0006, αâ²=0.0018). Bootstrap confidence limits (95% CI) showed a significant deficit of heterozygosity across polymorphic loci in age classes IIâIV but not in age class I (Table 1). When individuals were pooled over age classes into a single population, fT was significant, indicating deficits of heterozygosity compared to HâW expectations (pooled fT=0.081; Table 1). Assuming that the estimated fixation index is entirely because of selfing in the mixed mating model at equilibrium, a crude estimate of outcrossing rate (tm) was 0.85. The exact test of genetic differentiation among age classes was highly significant (P<0.0001).
Spatial genetic structuring within age classes
As shown by Morisita's index of dispersion, Iδ (Table 2), individuals in each age class were significantly clumped in all but the largest quadrat size (50Ã60âm). The regression of Iδ on quadrat size was significantly negative for all but age class II (â0.732<β<â0.920, 0.001<P<0.002, 0.536<R2<0.847; αâ²=0.0128), and reveals that Iδ decays to a value of 1.00 as quadrat size increases. The correlation coefficient between β with respect to each age class was not significantly positive, indicating that Iδ does not significantly decay with increasing age.
According to the criteria described above, nine (age classes III and IV), 10 (age class I), and 11 (age class II) alleles were used for Moran's I-statistics, analyzed separately for four age-classes. Moran's I-values were significantly different from the expected values in 14 (14.0%, age class I), 13 (11.8%, age class II), 10 (11.1%, age class III), and six (8.8%, age class IV) (data not shown). For distance one (<5âm), only three significantly positive cases were observed, in age classes I (Dia-1a and Pgi-2a) and III (Pgd-2a), and this indicates that spatial genetic structuring in the study population is weak at most. The number of significant overall correlograms was only two in age classes I and III, and one in age class II, and none was significant in age class IV, again indicating essentially random distributions of alleles (data not shown). Moreover, patterns of spatial genetic distributions observed separately for age classes IâIV were very similar, and the overall slope of the correlograms of mean Moran's I values was not significantly negative in each of four age-classes (Figure 2).
Based on 99% confidence limits, pairwise estimates of coancestry were not significantly different from zero within age classes, whereas for the total sample two significant but weak values were detected at distance 5âm (0.0205) and 40âm (â0.0086) (Figure 2). At the 95% level, a significant but weak negative value was also observed at 60âm in age class IV. The significance of values in the total sample is probably the result of increased power to reject the null hypothesis resulting from a larger sample size because the mean fij value at 5âm for the total sample is similar to that for age classes I and III (Figure 2). As with the Moran's I values, the overall slope of the correlograms of fij estimates was not significantly negative in four age-classes or the total sample.
Spatial genetic correlation between juveniles and adults
Estimates of coancestry between seedlings and 2â4 year juveniles vs adult trees did not reveal significant evidence of spatial autocorrelation. More specifically, none of the estimates for specific distance intervals was significant, nor was the slope of the computed coancestry values on distance significantly less than (or greater than) zero (results not shown).
Discussion
The coastal forests of northeast Asia possess broad-leaved evergreen woody species such as Camellia japonica, Castanopsis spp, Eurya spp, Ficus spp, Ligustrum spp, Litsea spp, Neolitsea spp, Persea spp, and Quercus spp. Some of these species are also important members of coastal forest vegetation in Japan and Korea, and through a series of studies of population dynamics (eg, Yamamoto, 1992), pollination ecology (eg, Yumoto, 1987), allozyme variation in local populations (eg, Wendel and Parks, 1985; Chung and Kang, 1994), and population genetic structure (eg, Chung et al, 2000; Ueno et al, 2000; Chung et al, 2003), the dynamics of these populations are beginning to become elucidated. As mentioned in the Introduction, parts of this study were undertaken to address how the relationships between trees and birds affect the internal spatial genetic structure of newly founded plant populations. Many fruit-eating birds (eg, Zosterops palpebrosa insularis) living in warm temperature broad-leaved evergreen forests in southern Korea and Japan feed on insects from spring to autumn. When insect food is scarce in winter, the birds collect fruits (drupes) from members of Lauraceae and also nectar from flowers of understory trees (eg, Camellia japonica [Theaceae]) (Yumoto, 1987; MY Chung and MG Chung, personal observation). This situation favors mutualism between the canopy trees (with their drupes having an edible portion) and the associated birds (endozoochory). In the sections that follow, we try to relate internal genetic structure among age classes in the study population to seed dispersal by birds.
Genetic diversity and structure in age classes
Estimates of genetic diversity (He) within the study population of C. insularimontanum were not significantly different among age classes. If most of the seedlings and juveniles studied are the products of repeated long-distance dispersal of seeds by birds, from other nearby older populations (probably the same source populations), there may be no reason to expect genetic differentiation among the four age-classes. However, significant genetic differentiation among age classes was detected, suggesting temporally varying demographic processes and reproductive events might have operated in the study population.
The estimate of the fixation index (fT=0.081) indicates a significant deficit of heterozygosity for the total sample, which may be indicative of inbreeding (a rough estimate of tm=0.85). C. insularimontanum is self-compatible (T Yahara, Kyushu University, personal communication), but the flowers show protandry and bagged inflorescences do not bear fruits (T Yumoto, Kyoto University, personal communication). Although selfing within a flower is not likely, selfing within a tree is possible, because different flowers on the same tree can be found at different stages of development (Yumoto, 1987). Unfortunately, we do not have any direct information on the level of geitonogamy in this species.
Alternatively, the significant deficit of heterozygosity could be a consequence of spatial or temporal Wahlund effects. A spatial Wahlund effect would be generated by spatial genetic structuring within the study population. Moran's I and coancestry (fij) spatial autocorrelation analyses, however, indicate little genetic substructuring. In contrast, evidence of a temporal Wahlund effect would be indicated by significant differentiation between age classes (eg, Tonsor et al, 1993). As noted above, there is a significant differentiation of genotype frequencies among age classes of C. insularimontanum, consistent with a temporal Wahlund effect. In this context, genetic differentiation would be attributable to allele frequency differences among source populations which can be measured hierarchically within- and between-age classes in terms of Wright's (1965) F-statistics (FIS, FST, and FIT). Using the Weir and Cockerham (1984) estimators of these parameters, FIS=0.066, FST=0.019, and FIT=0.084 (calculated using the program FSTAT (ver. 2.9.3.1) by Goudet, 2001). FST is significant (indicating allele frequency differences among age classes) but small, indicating that the overall deficit of heterozygosity (and similarity of FIS and FIT) must be explained by a process other than a temporal Wahlund effect, such as partial selfing. Further studies on mating system and the degree of population differentiation of C. insularimontanum are needed to infer causes of heterozygote deficiencies within local populations.
Spatial genetic structuring within age classes
There are five mature, potential source populations of C. insularimontanum within the Hakdongri Reserve (ca 50âha) on Geoje Island where hundreds of adults exist. Since only nine adults exist in the study population and the range of ages is broad, the study population appeared to have been initially established and then sequentially enlarged by repeated seed movements from maternal trees in nearby older stands (source populations).
As revealed by Moran's I-statistics and correlograms of estimated coancestry (fij), there was very little or essentially no spatial autocorrelation of genotypes for four age-classes in the population of C. insularimontanum (Figure 2). Moreover, although no statistical tests for differences between correlograms have been developed (Sokal and Wartenberg, 1983), the results of this study indicate that the spatial genetic distributions in the four age-classes are similar. One important aspect of seed movement is its role in the initial founding of a population. When populations are founded by long-distance, wind- or bird-dispersed seeds, in many cases the initial spatial distribution of genetic variation may be nearly random, because at long distances the dispersal function relative to distance is âflatâ; hence the relative resting places of seeds become nearly independent of their relative starting places (Epperson, 1990). Our results support these theoretical expectations. The age of a population (ie, how many cycles of regeneration have occurred since population establishment) is also important. Established populations can have much stronger spatial genetic structure because of the opportunity for localized seed dispersal and in older populations increased genetic isolation-by-distance over cycles of regeneration (eg, Epperson, 1993).
Individuals in the study population exhibited a clumped spatial distribution (Table 2), which could have been shaped by limited seed dispersal patterns and habitat heterogeneity. In the case of C. insularimontanum, microsite heterogeneity is probably responsible for the clumped distribution. If seeds fall beneath the maternal plants, individuals in younger age classes should exhibit a clumped spatial distribution. As seedlings and younger individuals grow, competition among individuals within cohorts causes thinning, shifting from a clumped to random distribution with age (eg, Alsesis blackiana and Platypodium elegans in Panama: Hamrick et al, 1993; Camellia japonica in the same study population: Chung et al, 2003). This pattern is not observed for C. insularimontanum. The homogeneity of spatial genetic structure and similar spatial clumping of individuals among the four age-classes suggest that patterns of seed dispersal at the population level are similar over time.
Spatial genetic correlation between juveniles and adults
Spatial genetic analyses revealed no significant autocorrelation between seedlings and 2â4 year juveniles vs adult trees. This allows us to reject the hypothesis that the youngest age class has been produced by local mating and subsequent dispersal close to the maternal tree. We cannot discriminate from these data alone, however, between local mating and spatially random dispersal vs seed production in the surrounding landscape followed by random dispersal into the study population. Evidence for the latter alternative, that is, the role of seed immigration in populating the study site, comes from inspection of the demographic data. In particular, more than half of the population consists of 10â40 year juveniles and sub-adults that are too old to have been produced by the reproductively mature individuals currently at the site. Extrapolating from this pattern of seed immigration over time, the majority of juveniles less than 10 years old are also likely to be immigrants, and local reproduction is likely only beginning to contribute to population growth within the site.
Comparison to other studies
The autocorrelation statistics observed for C. insularimontanum are similar to those found for most other insect-pollinated woody species with bird-dispersed seeds. For example, Dewey and Heywood (1988) found little autocorrelation in a population of the shrub Psychotria nervosa (Rubiaceae) in south Florida. An exception is the study by Loiselle et al (1995) on a population of Psychotria officinalis in Costa Rica, where they found significant coancestry among individuals separated by up to 5âm, but not beyond, indicating that some fruits fall underneath maternal plants. They suggested that although some proportion of seed dispersal may be localized, there is substantial mixing of seed shadows as a result of fruit removal by birds. Since P. officinalis has the potential for vegetative reproduction (JL Hamrick, personal communication), it is highly probable that the observed significant spatial genetic structure within short distance might be caused in part by clones. More recently, in the same forest as the present study, Chung et al (2000) examined changes in spatial genetic structure among five age-classes using four alleles and Moran's I-statistics of Neolitsea sericea, a dioecious tree that is insect pollinated, with red, bird-dispersed drupes. Like the present results, the average Moran's I-value for each of the five age-classes revealed little spatial genetic structuring among individuals within the population. Moran's I-values were significantly different from the expected values in five (12.5%, age classes I and V), six (15%, age class IV), and seven (17.5%, age classes II and III) of 40 cases in which only two significantly positive values were found in <14âm. In addition, patterns of spatial genetic distributions for five age-classes were very similar and the overall slope of the correlograms of mean Moran's I-values was not significantly negative in each of the five age-classes. The authors suggested that extensive, perhaps independent, seed dispersal by birds might cause a lack of genetic structure in an initially regenerative local population. Individual seeds of C. insularimontanum and N. sericea may move more or less independently because the single-seeded fruits of these species are large (ca 12âmm diameter), and birds can harvest only a few at a time (MY Chung and MG Chung, personal observation). This, combined with the inference that most of the trees in the study site resulted from long-distance seed dispersal from surrounding adult populations, means that we should expect genotypes to be essentially randomly distributed spatially within populations. However, in other species, where birds move seeds in groups (say of half-sibs) owing to seed-caching behavior (eg, Pinus albicaulis; Furnier et al, 1987) or the dispersal of intact fruits (eg, Cecropia obtusifolia; Epperson and Alvarez-Buylla, 1997), seedlings may exhibit striking spatial structuring within local populations, even if the seeds or fruits move long distances. These studies provide important implications that the amount of spatial genetic structure among seedlings may depend on the type of fruit and the behavior of individual birds or other animals (Hamrick and Loveless, 1986). Particularly important is whether or not birds cause correlated seed dispersal into a given local area or gap.
In summary, the present study of spatial and demographic genetics revealed a lack of genetic structure in a young local population, and we suggest that this is because of extensive gene flow through widespread, and perhaps nearly independent, dispersal of seed by birds. Thus, birds may act to maximize genetic diversity in newly founded populations, which appear to be continually forming and necessarily so because regeneration is limited in older, adult populations of C. insularimontanum. Overall, this may also help maintain genetic diversity in the metapopulation of C. insularimontanum.
References
Bacilieri R, Labbe T, Kremaer A (1994). Intraspecific genetic structure in a mixed population of Quercus petraea (Matt.) Leibl and Q. robur L. Heredity 73: 130â141.
Berg EE, Hamrick JL (1994). Spatial and genetic structure of two sandhills oaks: Quercus laevis and Quercus margaretta (Fagaceae). Am J Bot 81: 7â14.
Berg EE, Hamrick JL (1995). Fine-scale genetic structure of a turkey oak forest. Evolution 49: 110â120.
Casgrain P (2001). Permute! Version 3.4alpha. Available at http://www.umontreal.ca/casgrain/en/telecharger/index.htm1.
Cheliak WM, Pitel JP (1984). Technique for starch gel electrophoresis of enzyme from forest tree species. Information Report PI-X-42. Petawawa National Forestry Institute, Agriculture Canada, Canadian Forestry Service, Chalk River, Ontario, pp 1â49.
Chung MG, Chung MY, Oh GS, Epperson BK (2000). Spatial genetic structure in a Neolitsea sericea population (Lauraceae). Heredity 85: 490â497.
Chung MG, Kang SS (1994). Genetic variation and population structure in Korean populations of Eurya japonica (Theaceae). Am J Bot 81: 1077â1082.
Chung MY, Epperson BK, Chung MG (2003). Genetic structure of age classes in Camellia japonica (Theaceae). Evolution 57(1): (in press).
Chung MY, Nason J, Chung MG, Kim K-J, Park J-W, Sun B-Y et al (2002). Landscape-level spatial genetic structure in Quercus acutissima (Fagaceae). Am J Bot 89: 1229â1236.
Cliff AD, Ord JK (1981). Spatial Processes â Methods and Applications. Pion: London.
Cockerham CC (1969). Variance of gene frequencies. Evolution 23: 72â84.
Dewey SE, Heywood JS (1988). Spatial genetic structure in a population of Psychotria nervosa. I. Distribution of genotypes. Evolution 42: 834â838.
Epperson BK (1990). Spatial autocorrelation of genotypes under directional selection. Genetics 124: 757â771.
Epperson BK (1993). Recent advances in correlation analysis of spatial patterns of genetic variation. Evol Biol 27: 95â155.
Epperson BK, Alvarez-Buylla ER (1997). Limited seed dispersal and genetic structure in life stages of Cecropia obtusifolia. Evolution 51: 275â282.
Furnier G, Knowles P, Clyde MA, Dancik BP (1987). Effects of avian seed dispersal on the genetic structure of white-bark pine populations. Evolution 41: 607â612.
Gillet E (1998). DIPLOGEN: qualitative inheritance analysis of zymograms and DNA electrophenograms in diploid individuals. Available at http://www.uni-forst.gwdg.de/fg/software.htm.
Goudet J (2001). FSTAT, a program to estimate and test gene diversities and fixation indices (version 2.9.3.1). Available at http://www/unil.ch/izea/softwares/fstat.html.
Hamrick JL, Loveless MD (1986). The influence of seed dispersal mechanisms on the genetic structure of plant populations. In: Estrada A, Fleming TH (eds) Frugivores and Seed Dispersal, Dr. Junk Publications: Dordrecht, The Netherlands. pp 211â223.
Hamrick JL, Murawski DA, Nason JD (1993). The influences of seed dispersal mechanisms on the genetic structure of tropical tree populations. Vegetatio 107/108: 281â297.
Hamrick JL, Nason JD (1996). Consequences of dispersal in plants. In: Rhodes OE Jr, Chesser RK, Smith MH (eds) Population Dynamics in Ecological Space and Time, University of Chicago Press: Chicago, IL. pp 203â206.
Hossaert-McKey M, Valero M, Magda D, Jarry M, Cuguen J, Vernet P. (1996). The evolving genetic history of a population of Lathyrus sylvestris: evidence from temporal and spatial genetic structure. Evolution 50: 1808â1821.
Kalisz S, Nason JD, Hanzawa FA, Tonsor SJ (2001). Spatial population genetic structure in Trillium grandiflorum: the role of dispersal, mating, history, and selection. Evolution 55: 1560â1568.
Kitamura S, Murata G (1987). Colored Illustrations of Woody Plants of Japan. Hoikusha Publishing Co.: Japan.
Lewis PO, Zaykin D (2001). Genetic Data Analysis: Computer Program for the Analysis of Allelic Data. Version 1.0 [d16c]. Available at http://lewis.eeb.uconn.edu/lewishome/software.html.
Loiselle BA, Sork VL, Nason J, Graham C (1995). Spatial genetic structure of a tropical understory shrub, Psychotria officinalis (Rubiaceae). Am J Bot 82: 1420â1425.
Mitton JB, Linhart YB, Sturgeon KB, Hamrick JL (1979). Allozyme polymorphisms detected in mature needle tissue of Ponderosa pine. J Hered 70: 86â89.
Morisita M (1959). Measuring of the dispersion of individuals and analysis of the distributional patterns. Mem Fac Sci Kyushu Univ Ser E Biol 2: 215â235.
Parker KC, Hamrick JL, Parker AJ, Nason JD (2001). Fine-scale genetic structure in Pinus clausa (Pinaceae) populations: effects of disturbance history. Heredity 87: 99â113.
Raymond M, Rousset F (1995). An exact test for population differentiation. Evolution 49: 1280â1283.
Sakai AK, Oden NL (1983). Spatial pattern of sex expression in silver maple (Acer saccharum L.): Morista's index and spatial autocorrelation. Am Nat 122: 489â508.
Schnabel AR, Nason JD, Hamrick JL (1998). Understanding the population genetic structure of Gleditsia triacanthos L.: seed dispersal and variation in female reproductive success. Mol Ecol 7: 819â832.
Schneider S, Roessli D, Excoffier L (2000). Arlequin ver. 2.000: A Software for Population Genetics Data Analysis. Genetics and Biometry Laboratory, University of Geneva: Switzerland. Available at http://anthro.unige.ch/arlequin.
SËidák Z (1967). Confidence regions for the means of multivariate normal distributions. J Am Stat Assoc 62: 626â633.
Sokal RR, Oden NL (1978). Spatial autocorrelation in biology. 1. Methodology. Biol J Linn Soc 10: 199â249.
Sokal RR, Wartenberg D (1983). A test of spatial autocorrelation analysis using an isolation-by-distance model. Genetics 105: 219â237.
Soltis DE, Haufler CR, Darrow DC, Gastony GJ (1983). Starch gel electrophoresis of ferns: a compilation of grinding buffers, gel and electrode buffers, and staining schedules. Am Fern J 73: 9â27.
Tonsor SJ, Kalisz S, Fisher J (1993). A life-history based study of population structure: seed bank to adults in Plantago lanceolata. Evolution 47: 833â843.
Ueno S, Tomura N, Yoshimura H, Manabe T, Yamamoto S (2000). Genetic structure of Camellia japonica L. in an old-growth evergreen forest, Tsushima, Japan. Mol Ecol 9: 647â656.
Wade MJ, McCauley DE (1988). Extinction and recolonization: Their effects on the genetic differentiation of local populations. Evolution 42: 995â1005.
Weeden NF, Wendel JF (1989). Genetics of plant isozymes. In: Soltis DE, Soltis PS (eds) Isozymes in Plant Biology, Dioscorides Press: Portland, OR. pp 46â72.
Weir BS, Cockerham (1984). Estimating F-statistics for the analysis of population structure. Evolution 38: 1358â1370.
Wendel NF, Parks CR (1985). Genetic diversity and population structure in Camellia japonica L. (Theaceae). Am J Bot 72: 52â65.
Wendel JF, Weeden NF (1989). Visualization and interpretation of plant isozymes. In: Soltis DE, Soltis PS (eds) Isozymes in Plant Biology, Dioscorides Press: Portland, OR. pp 5â45.
Wright S (1922). Coefficients of inbreeding and relationship. Am Nat 56: 330â338.
Wright S (1965). The interpretation of population structure by F-statistics with special regard to systems of mating. Evolution 19: 395â420.
Yamamoto S (1992). Gap characteristics and gap regeneration in primary evergreen broad-leaved forests of Western Japan. Bot Mag Tokyo 105: 29â45.
Yeh FC, Yang RC, Boyle TBJ (1999). POPGENE Version 1.31, Microsoft Window-based Free Ware for Population Genetic Analysis. University of Alberta and Centre for International Forestry Research: Alberta, Canada. Available at http://www.ualberta.ca/~fyeh/index.htm.
Yumoto T (1987). Pollination systems of a warm temperature evergreen broad-leaved forest in Yakushima Island. Ecol Res 2: 133â145.
Acknowledgements
The authors thank NW Lee and SK Lee for providing support and assistance in collecting samples. The authors also thank A Schnabel for many helpful suggestions on the earlier version of the manuscript. This research was supported by a grant from Korea Research Foundation Grant (KRF-99-041-D00373) to MGC.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Chung, M., Nason, J., Epperson, B. et al. Temporal aspects of the fine-scale genetic structure in a population of Cinnamomum insularimontanum (Lauraceae). Heredity 90, 98â106 (2003). https://doi.org/10.1038/sj.hdy.6800187
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1038/sj.hdy.6800187
Keywords
This article is cited by
-
Genetic diversity and population structure of pummelo (Citrus maxima) germplasm in China
Tree Genetics & Genomes (2017)
-
Genetic diversity of flowering dogwood in the Great Smoky Mountains National Park
Tree Genetics & Genomes (2012)
-
Genetic structure at patch level of the terrestrial orchid Cyclopogon luteoalbus (Orchidaceae) in a fragmented cloud forest
Plant Systematics and Evolution (2011)
-
Multiple-scaled spatial genetic structures of Fraxinus mandshurica over a riparianâmountain landscape in Northeast China
Conservation Genetics (2010)
-
Fine-scale spatial genetic structure and allozymic diversity in natural populations of Ocotea catharinensis Mez. (Lauraceae)
Conservation Genetics (2010)