Am. J. Bot. Join the BSA
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (12)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Kang, S. S.
Right arrow Articles by Chung, M. G.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Kang, S. S.
Right arrow Articles by Chung, M. G.
Agricola
Right arrow Articles by Kang, S. S.
Right arrow Articles by Chung, M. G.
(American Journal of Botany. 2000;87:1634-1646.)
© 2000 Botanical Society of America, Inc.

High levels of allozyme variation within populations and low allozyme divergence within and among species of Hemerocallis (Liliaceae)1

Soon Suk Kang2,0 and Myong Gi Chung3,0

0 Department of Biology, Gyeongsang National University, Chinju 660-701, The Republic of Korea

Received for publication July 13, 1999. Accepted for publication January 18, 2000.


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 LITERATURE CITED
 
Thirty populations from five species of Hemerocallis in Korea were analyzed by starch gel electrophoresis to measure genetic diversity and to determine genetic population structure and the amount of genetic divergence within and between species at 12 isozyme loci. In addition, Moran's I spatial autocorrelation statistics were used to examine the spatial distribution of allozyme polymorphisms in populations of H. thunbergii and H. hakuunensis. Populations of five Korean species maintain high levels of genetic variation and little differentiation among populations and species. Mean expected heterozygosities range from 0.165 in H. hongdoensis, an island endemic, to 0.265 in H. taeanensis, and a total of 81 alleles across the 12 loci were detected in the five species. GST values for each of the five species were low, ranging from 0.051 in H. taeanensis to 0.078 in H. hakuunensis. Mean intraspecific Nei's genetic identities (I) between populations of the five species were all above 0.97. However, a considerable level of heterozygote deficiencies within populations was detected, ranging from 0.242 to 0.411 measured as FIS statistics. This deficiencies may be due to inbreeding, limited pollen and seed dispersal, or from the pooling of subpopulations that differ in allele frequencies. A small spatial scale population substructuring (<12 m) was found in H. thunbergii and H. hakuunensis. A group of populations from each of the five previously designated Hemerocallis species (based on their morphology, ecology, and phenology) agrees with our allozyme data, though pairwise comparisons among species had high I values (from 0.862, H. middendorffii vs. H. hongdoensis, to 0.969, H. thunbergii vs. H. taeanensis). This is attributed to the presence of the same high-frequency alleles in different species at seven loci. In addition, no "diagnostic allele" that appears in all populations of one species, but is absent in other species, was detected at the 12 isozyme loci. These all suggest that species of Hemerocallis in Korea may have recently derived from an ancestor or progenitor harboring high levels of genetic diversity.

Key Words: allozyme divergence • daylily • genetic diversity • genetic structure • Hemerocallis • Liliaceae • spatial autocorrelation


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 LITERATURE CITED
 
Spatial distribution of genetic variation in plant populations is primarily determined by seed and pollen dispersal, habitat distribution, microenvironmental selection, and genetic drift (Levin and Kerster, 1974; Epperson, 1993 ). The degree to which selection and drift affects patterns of genetic variation is known to depend on the dispersal ability (gene flow via pollen and seed dispersal) of plant species (Hamrick and Nason, 1996 ). For example, if gene flow is limited to a small spatial scale, populations will have more inbreeding and are, as a result, more likely to differentiate in response to local selective forces or to genetic drift. Thus, it is important to analyze spatial genetic structure in plant populations, which can be quantified using spatial autocorrelation analysis (Sokal and Oden, 1978 ).

Estimates of levels and apportionment of genetic diversity in plant species could be used as a yardstick for sound management of rare or endemic species (Hamrick et al., 1991 ). Recent reviews of plant allozyme literature have shown that rare, endemic, or narrowly distributed plants tend to contain less allozyme variation than species with a widespread distribution (Hamrick and Godt, 1989 ; Hamrick, Godt, and Sherman-Broyles, 1992 ). However, several studies reported that geographic range is not always a good predictor of the genetic structure of species (Soltis and Soltis, 1991 ; Chung, 1994a ; Qiu and Parks, 1994 ; Lewis and Crawford, 1995 ; Kang and Chung, 1997 a; Kang et al., 1998 ).

In addition, allozymes have proven useful in solving systematic problems at or below the species level and in elucidating the evolutionary history of a variety of plant groups (Gottlieb, 1981 ; Crawford, 1983, 1989 ). For example, agreements between the phenetic relationships based on the morphometric analyses among congeners and those based on allozyme data can be identified (e.g., Chung et al., 1991 ; Whitkus, 1992 ; Mayer, Soltis, and Soltis, 1994 ; McClintock and Waterway, 1994 ; Ford et al., 1998 ; Lee et al., 2000 ).

The genus Hemerocallis L. (Liliaceae) is a good system for addressing these issues (i.e., general evolutionary and population genetic processes) with a multipopulation and multispecies data set. Hemerocallis includes ~30 species of herbaceous perennials from Japan, Korea, and China (Matsuoka and Hotta, 1966 ). Many species and ~25 000 cultivars of the genus are widely grown in gardens in Asia, Europe, and North America for their attractive flowers (Stout, 1934 ; Cohen, 1986 ). In spite of their economic importance, numerous nomenclatural and taxonomic problems exist within the genus (Matsuoka and Hotta, 1966 ). The taxonomic difficulties have been attributed to the fact that many species (e.g., H. aurantiaca Baker, H. dumortieri Morren, H. flava L., H. fulva L., H. sulphurea Nakai, and H. thunbergii Baker, etc.) were described from cultivated plants of unknown origin (Matsuoka and Hotta, 1966 ) and the extreme differences in appearance between living plants and dried herbarium specimens (Matsuoka and Hotta, 1986 ). In addition, many species of Hemerocallis are so variable ecologically and morphologically that a proper species concept requires morphological, ecological, and biosystematic studies. Based on field surveys and morphological analysis, Hotta and his associates (Matsuoka and Hotta, 1966 ; Hotta, Ito, and Okata, 1984, 1985 ; Hotta, 1986 ) and Noguchi (1986) have published several papers on Hemerocallis taxonomy mainly from Japan during the past 30 yr. However, there has been little improvement in taxonomic status and understanding of the evolutionary history of several species of the genus such as H. aurantiaca, H. littorea Makino, the H. middendorffii Tr. et Mey. complex, and native species in the Korean Peninsula.

In an attempt to improve our knowledge of the genus, we have collected samples from 75 locations in Korea during the past 10 yr to determine the geographical distribution patterns for the Korean species and to characterize morphological variation within and among species (Chung and Kang, 1994a, b ; Kang and Chung, 1994, 1997b ). Based on our previous studies, we recognize five Hemerocallis species native to Korea: H. hakuunensis Nakai (= H. micrantha Nakai), H. thunbergii (= H. coreana Nakai), H. middendorffii, H. hongdoensis M. Chung & S. Kang, and H. taeanensis S. Kang & M. Chung. The geographic distributions and ecological traits of each species are described in detail here and are also available in Chung and Kang (1994a, b) and Kang and Chung (1997b) .

Flowers of members in sections Fulvae Nakai and Capitatae Nakai in the genus Hemerocallis are orange-yellow and usually visited by bees and butterflies. On the other hand, fragrant lemon-yellow flowers in section Hemerocallis are pollinated by moths (Hotta, Ito, and Okata, 1984 ; Kang and Chung, personal observations). Members in the genus have no specialized mechanisms for seed dispersal, and many seedlings are found near maternal plants in natural populations (Kang and Chung, personal observations). Considering this, a small spatial scale of genetic substructuring is expected in local populations of a species.

It might be of interest to compare restricted or island endemic plants to more widespread Hemerocallis species. When we consider general trends in the plant allozyme literature, we might predict that populations of H. hakuunensis, growing on diverse habitats in southern, central, and northeastern Korea, H. taeanensis, found only on the southwestern Korean coast, and H. hongdoensis, found on Cheju, Hong, Taehuksan, Sohuksan Islands in Korea and Tsushima Island in Japan, may maintain a lower allozyme diversity than those of H. thunbergii, growing on the grasslands of southwestern and central, Korea and from central China to the central part of Japan, and H. middendorffii, which is distributed over a diverse range of ecological conditions in northeastern China, central and northern Japan, and the northeastern Korean Peninsula.

Finally, if a species group has diverged only recently and there simply has not been sufficient time to differentiate through drift, a high Nei's genetic identity among species pairs is also expected (Crawford, 1989 ). Preliminary allozyme studies in Japanese populations of H. thunbergii, H. middendorffii, and H. exalata Stout revealed a high degree of within-population variation and a close relationship among the three species (Kang et al., 1998 ). This study is an extension of this work, which examines the allozyme variation of the five Hemerocallis species in Korea, addressing an important question, namely, how does population genetics influence taxonomic relationships among species? Thus, the purposes of this study were to: (1) assess genetic diversity and structure in species in terms of population size, population substructuring, and the ecological and life-history characteristics; (2) gain an understanding of evolutionary processes operating in the five morphological species; and (3) evaluate the taxonomic treatments proposed by Kang and Chung (1997b) .


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 LITERATURE CITED
 
Population samples
From 1992 to 1996, a total of 1410 samples were collected from 30 populations of five Hemerocallis species in South Korea in order to assess genetic diversity within species and genetic relationships among species (Fig. 1). The five species were sampled proportionally to their abundance in South Korea. Number of individuals per each population (number per square metre) were counted, and habitat types were recorded (see Table 3). Collections of leaf samples from each site were randomly made. For the analysis of spatial distribution of genetic variation in populations of H. thunbergii and H. hakuunensis, in August 1997, all visually identified H. thunbergii (N = 205) and H. hakuunensis (N = 75) individuals were mapped, and leaf samples were collected within a 40 x 60 m plot (altitude 240 m above sea level, 5% gradient facing south-southwest) in Gushi Ri, Hyunsan Myeon, Haenam Gun, Chollanam Province (hereafter referred to as HAN), and within a 40 x 50 m plot (altitude 340 m above sea level, 15% gradient facing south-southeast) in Dongsan Ri, Sangri Myeon, Gosung Gun, Gyeongsangnam Province (KOS). Individuals in HAN grow on an open habitat (grasslands), whereas those in KOS are found under a Pinus densiflora overstory. Leaf samples were placed in plastic bags wrapped with a wet paper towel and stored on ice for transportation to the laboratory. Samples were then stored at 4°C until protein extraction. Voucher specimens of all collections were deposited at the Gyeongsang National University Herbarium (GNUC).



View larger version (28K):
[in this window]
[in a new window]
  		Fig. 1. Geographic distributions of Hemerocallis in Korea and on Tsushima Island. Alphanumeric codes are from Table 3 , and arrows indicate locations of populations from which samples were collected. The spatial autocorrelation analysis was conducted in populations HAN (H. thunbergii, N = 205) and KOS (H. hakuunensis, N = 75)

 

View this table:
[in this window]
[in a new window]
  		Table 3. Genetic diversity within populations of Hemerocallis: proportion of loci polymorphic (%P), mean number of alleles per polymorphic locus (AP), mean number of alleles per locus (A), mean effective number of alleles per locus (Ae), average observed heterozygosity (HEo), average expected heterozygosity (HEe), standard errors (SE), and FIS for ten polymorphic loci (numbers in parentheses indicate statistically significant number of loci, P < 0.05)

 
Enzyme extraction and electrophoresis
Leaf samples were cut finely and crushed with a mortar and pestle. A phosphate-polyvinylpyrrolidone extraction buffer (Mitton et al., 1979 ) was added to the leaf samples to facilitate crushing and to aid enzyme stabilization. Enzyme extracts were absorbed onto 4 x 6 mm wicks cut from Whatman 3MM chromatography paper, which were then stored at -70°C until needed. Electrophoresis was performed using 11% starch gels. Fifteen putative loci were resolved from eight enzyme systems using three gel/electrode buffer combinations. Two Poulik buffer systems were used: a modification (Haufler, 1985 ) of Soltis et al.'s (1983) "system 6" was used to resolve leucine aminopeptidase (LAP) and fluorescent esterase (FE) and a modification (electrode buffer of pH 8.6) of Soltis et al. (1983) resolved diaphorase (DIA), alcohol dehydrogenase (ADH), and ß-galactosidase (ß-Gal). A morpholine citrate buffer system, a modification (Chung and Kang, 1994c ) of that of Clayton and Tretiak (1972) , was used to resolve malate dehydrogenase (MDH), phosphoglucoisomerase (PGI), and 6-phosphogluconate dehydrogenase (6PGD). Stain recipes were taken from Soltis et al. (1983) , except for the DIA and ß-Gal, which were taken from Cheliak and Pitel (1984) . The genetic basis of enzyme banding patterns was inferred from observed segregation patterns in light of typical subunit structure and subcellular compartmentalization (Weeden and Wendel, 1989 ). Putative loci were designated sequentially, with the most anodally migrating isozyme designated 1, the next 2, etc. Likewise, alleles were designated sequentially with the most anodally migrating allele designated superscript a.

Data analysis
For spatial autocorrelation analysis, the genotypic data were coded so that allele frequency values of 1.0, 0.5, or 0.0 were assigned to each individual being homozygous for a given allele, heterozygous for that allele, or genotypes with no copies of that allele, respectively, for each polymorphic locus (Sokal and Oden, 1978 ). For diallelic loci, only those with allele frequencies <0.95 and >0.05 were used, but only one allele was considered because both alleles provide the same information. For loci having more than two alleles, all alleles were used for the spatial analysis regardless of their frequencies. Every possible pair of individuals was considered as a join or a connection between two individuals and was assigned to one of the ten distance classes according to the distance separating the two individuals. Because measures of small-scale autocorrelation more accurately represent the spatial structure, the first of the ten distance classes was designed based on an estimate of the average distance that separates nearest neighbor individuals. Moran's I statistic (Sokal and Oden, 1978 ) was calculated for each of ten distance classes by interplant distance classes by

where N is number of individuals; the Wij are elements of the weighting matrix, such that Wij equals 1.0 if ith and jth individuals are in the distance class k and zero otherwise; Zi = Xi - X, Zj = Xj - X; the variables Xi and Xj are the genotypic scores for ith and jth individuals, respectively; X is the mean score for all individuals; and W is the sum of all Wij for the distance class. Under null hypothesis, Ho, of a random distribution, Ik has expected value, µ1, equal to -1/(N - 1) and expected variance, µ2 (computational formula given by Sokal and Oden [1978] and Cliff and Ord [1981] ). Thus, the statistic, SNDk = (Ik - µ1)/µ2½, is a two-tailed test statistic of Ho, and it has an approximate standard normal distribution (Cliff and Ord, 1981 ) under Ho. Thus, absolute values of SNDk greater than 1.96 and 2.58 are significant at levels 0.05 and 0.01, respectively. A significant positive value of SNDk indicates that the neighboring individuals have more alleles in common within distance class k than would be expected by chance, whereas a significant negative value suggests that distant individuals have fewer alleles in common. Overall significance of individual correlograms was tested using Bonferroni's criteria (Sakai and Oden, 1983 ). All calculations and statistical analyses were performed using the SAAP program (version 4.3) written by D. Wartenberg.

For calculation of Wright's (1965) FIS, genetic diversity within populations, and genetic divergence among populations and species, a locus was considered polymorphic if two or more alleles were detected, regardless of their frequencies. Wright's FIS values on each polymorphic locus were calculated to measure relative excesses of homozygotes or heterozygotes compared with panmictic expectations within populations of the five species. Deviations of FIS from zero were tested using a chi-square test (Li and Horvitz, 1953 ).

Five standard genetic parameters were estimated using a computer program developed by M. D. Loveless and A. Schnabel (personal communication): percentage polymorphic loci (%P), mean number of alleles per polymorphic loci (AP), mean number of alleles per locus (A), effective number of alleles per locus (Ae), and genetic diversity or expected heterozygosity (HEe) (Hamrick, Godt, and Sherman-Broyles, 1992 ). These parameters were estimated at both population and samples as a whole (species level).

Nei's (1973, 1977) statistics of gene diversity formulae (HT, HS, DST, and GST) were used to evaluate the distribution of genetic diversity within and among populations of each species. Statistical significance of each GST value was calculated by a chi-square test (Workman and Niswander, 1970 ). GST values were averaged across polymorphic loci to obtain an estimate of overall population divergence.

Finally, Nei's (1972) genetic distance and identity measures were calculated for all pairs of populations to estimate genetic divergence among populations and species. A cluster analysis on genetic distance values via unweighted pairwise groups method using arithmetic average (UPGMA) was generated using NTSYS (Rohlf, 1988 ) to examine genetic associations among species.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 LITERATURE CITED
 
Fifteen putative loci were resolved from eight enzyme systems. These were Lap-1, Lap-2, Fe-1, Fe-2, Adh, Dia-1, Dia-2, Mdh-1, Mdh-2, Pgd-1, Pgd-2, Pgi-1, Pgi-2, Pgi-3, and ß-Gal. The Lap-2, Dia-2, and Pgd-2 loci were not included in this study due to faint or inconsistent staining. The Pgi-1 and ß-Gal loci were monomorphic in all populations. A total of 84 alleles were observed. The table of allele frequencies for each population is too lengthy to include here, but a copy of this is available in Kang (1997) or upon request from the authors.

Population genetic structure
Chosen according to the criteria described in the Materials and Methods, 37 and 36 alleles were used for spatial autocorrelation analysis for populations of H. thunbergii and H. hakuunensis, respectively. The spatial autocorrelation coefficients, Moran's I, for populations HAN and KOS of H. thunbergii and H. hakuunensis are presented in Tables 1 and 2. For population HAN of H. thunbergii, Moran's I values were significantly different from the expected value (E[I] = -0.005) in 91 (24.6%) of 370 cases, and the overall correlogram was significant for 20 (54.1%) of 37 alleles (Table 1). For distance classes 1–3 (0 to <12 m), 38 significantly positive cases were observed, whereas only two significantly negative cases were detected in the distance classes. This indicates that most of the genetic similarity is shared among individuals separated by <12 m. Similar results were observed in KOS of H. hakuunensis. Moran's I values were significantly different from the expected value (E[I] = -0.014) in 81 (22.5%) of 360 cases, and the overall correlogram was significant for 12 (33.3%) of 36 alleles (Table 2). For distance classes 1–3 (0 to <10 m), 25 positive vs. eight negative values were statistically significant. The distance at which the mean Moran's I values first intercepts the E (I) value may represent the shortest length of an irregularly shaped patch size (Sokal, 1979 ). The mean correlograms of populations HAN and KOS indicate that the minimum patch widths were ~10–12 m (Fig. 2).


View this table:
[in this window]
[in a new window]
  		Table 1. Spatial autocorrelation coefficients (Moran's I) of 37 alleles in Haenam population (HAN) of Hemerocallis thunbergii for ten distance classes. The expected value, {µ}1, is -0.005

 

View this table:
[in this window]
[in a new window]
  		Table 2. Spatial autocorrelation coefficients (Moran's I) of 36 alleles in Kosung population (KOS) of Hemerocallis hakuunensis for ten distance classes. The expected value, {µ}1, is -0.013

 


View larger version (15K):
[in this window]
[in a new window]
  		Fig. 2. Correlograms for populations HAN (Hemerocallis thunbergii) and KOS (H. hakuunensis) using mean values of Moran's I and ten distance intervals. The distance intervals are in Tables 1 and 2

 
In a good agreement with the spatial autocorrelation statistics, mean FIS values showed that significant deficiencies of heterozygotes existed for most loci within populations for each of the five species (Table 3). The FIS values ranged from 0.242 in H. middendorffii to 0.411 in H. taeanensis, indicating a substantial population substructuring.

Allozyme diversity within and among populations
Hemerocallis species in Korea maintain high levels of genetic variation within populations and samples as a whole (Table 3). The genetic diversity (HEe) was significantly correlated with population size (Spearman's rank-correlation coefficient, rs = 0.72, P < 0.001), i.e., genetic diversity was lower in small populations than in large ones. In addition, individuals on open habitats maintain higher levels of allozyme diversity than on other habitats (e.g., forest margins, disturbed sites, etc.) (unpaired t test, t = 5.22, P < 0.001).

Hemerocallis thunbergii, H. hakuunensis, H. middendorffii, and H. taeanensis had high genetic diversity. On the other hand, three populations of H. hongdoensis maintained significantly (P < 0.05) lower mean values of HEe, %P, A, and Ae than those for the other four Hemerocallis species. Hemerocallis hongdoensis also had the lowest number of alleles (39; Table 3). The low values of Ae demonstrated that many of the alleles were present at very low frequencies (Table 3).

The majority (~92–95%) of the total genetic diversity resided within populations of all five species as shown by low GST values (ranging from 0.051 in H. taeanensis to 0.078 in H. hakuunensis; Table 4). However, GST values for all polymorphic loci in each species were significantly greater than zero (P < 0.001).


View this table:
[in this window]
[in a new window]
  		Table 4. Genetic diversity statistics for ten polymorphic loci in five Hemerocallis species: numbers in parentheses indicate statistically significant loci (P < 0.05), which were tested for significance of GST for polymorphic loci. Also shown is the total number of alleles, AT

 
Genetic relationships among species
A summary of allele frequencies pooled over populations for ten polymorphic loci is given in Table 5. All species shared the same highest-frequency allele at seven loci: Lap-1c, Fe-1d, Fe-2c, Mdh-1c, Mdh-2d, Pgd-1c, and Pgi-3e. Hemerocallis hakuunensis and H. middendorffii shared the same high-frequency allele at Adhg, whereas H. thunbergii, H. hongdoensis, and H. taeanensis all shared a different high-frequency allele at this locus (Adhe). Hemerocallis hakuunensis contained three unique alleles (Lap-1a, Fe-1a, and Dia-1f), H. thunbergii had three (Mdh-1e, Pgi-3i, and Pgi-3j), and H. taeanensis exhibited five unique alleles (Lap-1f, Pgd-1e, Pgd-1i, Pgi-3b, and Pgi-3f; Table 5). However, the frequencies of these alleles are low (<5%; Table 5). In addition, no "diagnostic allele" (Wolfe and Elisens, 1993 ) that appears in all populations of one species, but is absent in other species, was detected in this study. Only Pgi-2a ("marker allele"; Wolfe and Elisens, 1993 ) occurred in high frequency in H. hongdoensis (0.667), but was detected in low (<10%) frequency in H. hakuunensis (0.070) and H. thunbergii (0.022) (Table 5). Genetic identities ranged from 0.810 (H6 vs. O3) to 0.997 (A3 vs. A4) (Table 6). Intraspecific mean genetic identities were very similar in each of the five species, ranging from 0.970 in H. middendorffii to 0.986 in H. thunbergii (Table 6). Average genetic identities for pairwise comparisons of the five species were high, ranging from 0.862 (H. hongdoensis vs. H. middendorffii) to 0.969 (H. taeanensis vs. H. thunbergii) (Table 6). Except for two comparisons (H. hongdoensis vs. H. hakuunensis and H. middendorffii), values of pairwise comparisons exceeded 0.90 (0.904–0.969). It is of interest to find that mean interspecies gene differentiation (GST) values among the five species from 30 populations were low (0.135). Only eight private alleles with low frequencies were detected in the 30 populations. These were Lap-1e (H13, allele frequency = 0.030), Fe-1b (A3, 0.019), Adha (M1, 0.032), Mdh-2h (O2, 0.010), Pgi-2c (O3, 0.020), Pgi-2j (A3, 0.020), Pgi-3f (A3, 0.060), and Pgi-3j (T4, 0.020). These results indicate an extremely close relationship among the species.


View this table:
[in this window]
[in a new window]
  		Table 5. A summary of allele frequencies pooled over populations[fj for ten polymorphic loci among five species of Hemerocallis in Korea

 

View this table:
[in this window]
[in a new window]
  		Table 6. Matrix of average Nei's (1972) genetic identity coefficients (ranges) for all pairwise comparisons of sampled populations of the five [cf2]Hemerocallis[cf1] species in Korea

 
Interpopulational relationships among species are summarized as an UPGMA phenogram of Nei's genetic distance (Fig. 3). Except for one population (A2), the five previously designated morphological species (Chung and Kang, 1994a, b ; Kang and Chung, 1994, 1997b ) can be grouped by the allozymes. Although H. hongdoensis is most closely related to H. hakuunensis in its gross morphology (Chung and Kang, 1994a, b ), it is allozymically the most distinct from other Korean Hemerocallis species (Fig. 3).



View larger version (16K):
[in this window]
[in a new window]
  		Fig. 3. UPGMA cluster analysis of 30 populations of Hemerocallis species in Korea based on Nei's (1972) measure of genetic distance. Abbreviations are from Fig. 1 and Table 3 . Hhak, H. hakuunensis; Hmid, H. middendorffii; Hthu, H. thunbergii; Htae, H. taeanensis; and Hhon, H. hongdoensis.

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
LITERATURE CITED
 
Population genetic structure
The considerable level of heterozygote deficiencies revealed by nine and six positively significant values of FIS out of ten and nine polymorphic loci in the five Hemerocallis species can result from inbreeding (Wright, 1978 ) or from the pooling of subpopulations that differ in allele frequency (Wahlund, 1928 ). Self-pollination in garden conditions results in partly sterile or weakly growing offspring, indicating Hemerocallis species are self-compatible (M. Ito, Chiba University, personal communication). Individuals sampled in this study were found to be patchy in distribution partly because of limited seed dispersal. The spatial analysis of individual genotypes within populations of H. thunbergii, H. hakuunensis, and H. taeanensis (Park, Kang, and Chung, 1998 ) indicated that they are distributed in a structured, isolation-by-distance manner and a minimum patch width of the three species is <12 m. A substantial spatial genetic structure was also reported in populations of other liliaceous plants that have similar life history and ecological traits to Hemerocallis species (Chinographis japonica Maxim. var. japonica; Maki and Masuda, 1994 ; two species of Hosta Tratt.; Chung and Park, 1998 ). In this study samples were collected across the entire range in each population (ranging from ~400 to 6000 m2), thus several patches per population were pooled for the electrophoretic analysis. As these aggregates were small subpopulations or demes differing to some extent in allele frequency, their pooling resulted in a Wahlund effect.

Allozyme diversity within and among populations
Hemerocallis species maintain high levels of allozyme variation. A series study of Hosta (Liliaceae) species (Chung, 1994a, b, c, 1995a, b ; Chung and Chung, 1994 ) and a recent review of allozyme literature (Hamrick and Godt, 1989 ) reported that the average genetic diversity statistics for populations of plants having similar life history traits to those of Hemerocallis were comparable to those observed in this study. Chung et al. (1991) and Chung (1995a) suggested that the high levels of genetic diversity found in Hosta species may be due to the predominantly outcrossing breeding system, extensive gene duplication at several enzyme systems, and high chromosome numbers (2n = 60).

There was a significant positive correlation between allozyme diversity and actual population size in the study populations of Hemerocallis. In general, the low levels of genetic diversity were found in small populations on diverse habitats (e.g., forest margins, forest understories, disturbed orchards, and recently established roadsides). Populations with a small number of individuals coupled with reduced gene flow are genetically less variable due to inbreeding and genetic drift (e.g., Schnabel and Hamrick, 1990 ; Aspinwall and Christian, 1992 ; Parker and Hamrick, 1992 ; Dolan, 1994 ). The loss of variation found in small populations of Hemerocallis could be the result of founder effects. It is obvious that the expansion, contraction, extinction, or recolonization have affected the distribution of genotypes in a spatial scale and genetic variation within and among populations of Hemerocallis.

As expected, Korean populations of H. thunbergii and H. middendorffii have high genetic diversity. The two species have a wide geographic range distributed from China to parts of the Korean Peninsula and the Japanese Archipelagos. Most Korean populations of H. thunbergii grow commonly in the open, grasslands on hillsides in the southwestern Korean Peninsula. It has been observed that Korean populations of the species are large and have a relatively continuous distribution. The spatial autocorrelation analysis in population HAN showed that, based on the sample sizes, the observed average value of Moran's I of 0.03 for the first distance class is roughly comparable to that expected from a stable population with Wright's (1943) neighborhood size (a standardized measure of dispersal) of ~115–125 individuals (Epperson, Huang, and Li, 1999 ). Factors such as large populations distributed over a relatively homogeneous habitat and colonization events (in forest gaps due to fire and in disturbed habitats by reforestation in southern Korea) may have contributed to the considerable level of genetic variation in H. thunbergii. On the other hand, populations of H. middendorffii in Japan grow in diverse ecological conditions on Honshu such as alpine meadows, low grasslands, and forest understories etc., showing complex local variation in morphological traits (Noguchi, 1986 ; Chung and Noguchi, 1998 ). Only a few populations of this species are found in the southeastern Korean Peninsula. Sizes of the three populations (M1–3) were relatively small (<250 individuals per population) and geographically isolated. Population M1 grows on the margin of a hillside under a pine-oak overstory (altitude ~300 m), M2 grows on an open habitat (altitude ~760 m), and M3 on a forest margin along stream sides (altitude ~800 m). A previous allozyme study of four Japanese populations of H. middendorffii (mean sample size = 27, HEe = 0.172–0.300; Kang et al., 1998 ) revealed a slightly lower genetic diversity within populations than that for the three Korean populations. Genetic diversity is also strongly determined by genetic history in long-lived organisms that were undoubtedly strongly influenced by Pleistocene events (e.g., gene flow, size of local populations, distribution ranges, and ecological succession of vegetation in the Japanese Archipelagos and the Korean Peninsula; Noguchi, 1986 ; Kang et al., 1998 ). It is suggested that Pleistocene migration events, selective forces and/or genetic drift may have played roles in shaping population genetic structure of H. middendorffii. Thus, the mechanisms for maintaining high levels of allozyme diversity within populations seem to be very different between the two widely distributed species.

As predicted in the introduction, populations of H. hongdoensis, an island endemic (Fig. 1), contain much less genetic variation. Populations of H. hongdoensis were relatively small, isolated, and occurred in rocky and humus soil on talus slopes or rocky cliffs along the sea coast in Taehuksan, Sohuksan, Hong and Seop Islands in Korea, and Tsushima Island in Japan (Chung and Kang, 1994a, b ). The relatively narrow habitats and geographic distribution of H. hongdoensis might in part account for its low levels of genetic diversity (Hamrick and Godt, 1989 ). A very similar level of allozyme diversity was found in two populations of H. exalata, endemic in Sado and Tobishima Islands, off the western coast of Japan (Kang et al., 1998 ).

Populations of H. hakuunensis are commonly found on the humus or granitic soils and open areas or margins of pine-oak forests on the hillsides of the southern, central, and northwestern Korean Peninsula (Fig. 1; Chung and Kang, 1994a ; Kang and Chung, 1994, 1997b ). On the other hand, H. taeanensis grows only on sandy soil under pine-oak forests near the coast of Taean Gun, a part of the central-western Korean Peninsula (Fig. 1). Although H. hakuunensis has a broader geographic distribution, the two species have a very similar and high genetic diversity at the population and species levels. This is somewhat surprising, if one compares and contrasts the geographical and ecological distributions of these two species. If only the geographical range, the number of populations, and the spectrum of ecological niches were taken into account, one might expect that H. taeanensis would be the least genetically variable of the five species. However, H. taeanensis contains a level of variation comparable to that found in a broad and widespread Hemerocallis species. Although H. taeanensis has restricted distributions, it is locally common with a relatively large population size (Wright's [1943] neighborhood size of ~115 individuals; recalculated from Park, Kang, and Chung, 1998 ), and thus may resemble more widespread congeners such as H. hakuunensis and H. thunbergii in its amount of genetic variation (Holsinger and Gottlieb, 1991 ).

Hamrick and Godt (1989) noted the importance of the evolutionary history of particular species in influencing levels of genetic diversity. For example, species that evolve via hybridization should have enhanced levels of genetic variation (e.g., Brunsfeld, Soltis, and Soltis, 1991 ; Broyles and Wyatt, 1993 ). Most diploid Hemerocallis species easily cross naturally and artificially, and hybrids are highly fertile (Nakao and Yamashita, 1956 ; Kawano, 1961 ; Kawano and Noguchi, 1973 ; S. Kawano and J. Noguchi, Kyoto University, personal communication). Species of Hemerocallis have a typical outcrossing pollination system carried out by bees, flies, butterflies, and moths (Hotta, Ito, and Okata, 1985 ). With this reproductive system, Hemerocallis species have weakly developed genetic isolating mechanisms. For this reason, interspecific natural hybrids have been documented between species that occur in sympatric and parapatric populations (Nakao and Yamashita, 1956 ; Kawano, 1961 ; Hotta, Ito, and Okata, 1985 ). For example, Hotta, Ito, and Okata (1985) presented evidence of gene flow in the past from H. littorea to H. thunbergii and H. hongdoensis (the authors named H. hongdoensis as H. hakuunensis) in a few island populations around Kyushu, Japan, based on floral color, the V-shaped pattern on the perianth, stolon of the root system, and evergreen/deciduous habit. It has been difficult to find hybrid swarms in Korea because Korean Hemerocallis species lack the morphological traits that Hotta and his associates used. In addition, we could not find any diagnostic alleles at the eight enzyme systems for detecting hybridization or introgression among Korean species. Other sources of genetic markers at DNA levels are needed to determine whether introgression has occurred among Korean species.

The GST values for each of the five species are low and very similar. This level of divergence is low compared with that of the mean for short-lived herbaceous perennials (GST = 0.233), outcrossing animal-pollinated species (GST = 0.197), species with seed dispersal mechanism by gravity (GST = 0.277), and species with sexual reproduction (GST = 0.225; Hamrick and Godt, 1989 ). This is somewhat surprising when we consider the fact that no primary special seed dispersal mechanisms were apparent in the five species. Seeds are mainly dispersed by gravity and fall around individual parent plants (Kang and Chung, personal observations). However, we cannot rule out the possibility of the secondary seed movement by rainfall into adjacent areas in the hillsides among the peninsular populations of the species for a moderately long period of time. In addition, Hemerocallis species are predominantly pollinated by insects (bees and butterflies in diurnal species such as H. hakuunensis, H. middendorffii, H. taeanensis, and H. hongdoensis, and moths in nocturnal species of H. thunbergii). Considering these and the small-scale population substructuring found in Hemerocallis species, it is suggested that most populations of each species may have diverged only recently, after being founded by individuals harboring similar genotypes rather than high rates of historical gene flow among populations.

Genetic relationships among species
A group of populations of each of the five Hemerocallis species in Korea based on their morphology and phenology (Chung and Kang, 1994a ) is in accordance with that of our allozyme data (see Fig. 3). Although the five Korean species are morphologically distinct (Matsuoka and Hotta, 1966 ; Chung and Kang, 1994a, 1997a ), they had high genetic identities. The high genetic identities for pairwise comparisons among species of Hemerocallis are attributable to the presence of the same high-frequency alleles in different species at seven loci. In addition, no diagnostic allele was detected at the eight enzyme systems. One may question that the 30 populations examined here could be interpreted as populations, varieties, or subspecies of a highly polymorphic taxon with some populations isolated on islands since the last Ice Age and others connected by clines on the mainland (C. R. Parks, personal communication). This interpretation is less plausible because the five Hemerocallis species could be recognized by several quantitative and qualitative characters (Chung and Kang, 1994a ; Kang and Chung, 1997b ). For example, the flowers of H. thunbergii (section Hemerocallis; Matsuoka and Hotta, 1966 ) are fully open in the late afternoon and throughout the night, and close the following morning ("nocturnal habit"). Also, flowers (5–28 per inflorescence) are bright, lemon-yellow, and fragrant. Petals are fringed, and perianth tubes are significantly longer (mean ± 1 SD = 3.62 ± 0.50 cm) than those for other members of daylilies in Korea (e.g., mean ± 1 SD = 2.23 ± 0.31 cm for H. hakuunensis; Chung and Kang, 1994a ). On the other hand, flowers (5–16 per inflorescence) of H. hakuunensis (section Fulvae Nakai, Matusoka and Hotta, 1966) are orange-yellow and "diurnal." The difference in flowers is the consequence of their adaptation to different pollinators (moths for H. thunbergii vs. bees and butterflies for H. hakuunensis). Nonetheless, allozyme differentiation between the two species is slight (mean I = 0.947 [0.914–0.980]). High genetic identities among congeners have been reported (e.g., Lasthenia—Crawford, Ornduff, and Vasey, 1985 ; Heuchera—Soltis, 1985 ; see Crawford, 1989 , for general discussion; Salix exigua group—Brunsfeld, Soltis, and Soltis, 1991 ; Streptanthus glandulosus complex—Mayer, Soltis, and Soltis, 1994 ). These are considered examples of recent evolution accompanied by little genetic change such as a progenitor-derivative species pair (Gottlieb, 1973 ). On the other hand, the low mean GST value (0.120) obtained from 27 populations, with the exception of three island populations of H. hongdoensis, may indicate that gene movement may have occurred among peninsular species for a long period of time, resulting in high genetic identities among peninsular species. However, this statement is unlikely because genetic substructuring in populations of H. thunbergii and H. hakuunensis was found at a spatial scale of <12 m, partly due to limited pollen and seed dispersal. Instead it is suggested that members of Hemerocallis may have recently derived from an ancestor or progenitor harboring high levels of genetic diversity (e.g., large number of alleles).

Hotta, Ito, and Okata (1985) and Chung and Kang (1994a) suggested that the progenitor (probably peninsular populations of H. hakuunensis) of H. hongdoensis had migrated from the Korean Peninsula into Tsushima Island and/or Kyushu of the southern Japan and Hong, Taehuksan, Sohuksan, and nearby islands during the Ice Age (the glacial Würm). During the Ice Age, the sea was ~100 m lower than at present and land connection existed at the Korean Strait (between Tsushima Island, Japan and Pusan, Korea) and between remote islands (such as Hong, Teahuksan, and Sohuksan Islands) off the southwestern Korean coast and the Korean Peninsula. The glacial remnants after the last Ice Age differentiated and became these islands. Chung and Kang (1994a) proposed this scenario based on paleoclimatic history, similar morphological features between H. hakuunensis and H. hongdoensis (e.g., lack of stolon, deciduous habit, and orange-yellow perianth color, etc.), and close geographic distributions of the two species. The present allozyme data (e.g., Pgi-2a) support this hypothesis (see Table 5, Fig. 4). For example, nine populations of H. hakuunensis have the Pgi-2a allele, ranging from 0.01 in H2 and H12 to 0.34 in H9 (altitude ~1300 m). This allele was also detected in the populations of H. hongdoensis at a relatively high frequency (from 0.48 in O1 to 0.83 in O3). In addition, all 11 individuals from Tsushima Island showed only the Pgi-2a allele (Fig. 4).



View larger version (28K):
[in this window]
[in a new window]
  		Fig. 4. Occurrence of a marker allele, Pgi-2a. Alphanumeric codes are same as in Table 3 and Fig. 3 . Darkened portions of circles represent allele frequencies in the respective populations. Arrows indicate populations of H. hongdoensis.

 
Past events regarding the origin of H. taeanensis cannot be inferred with certainty. We hypothesize that origin and establishment of H. taeanensis may have resulted from geographic isolation of a peripheral population of H. hakuunensis. This statement is based on the morphological and phenological similarity between the two species (e.g., orange-yellow perianth [100% vs. ~95% within populations of H. hakuunensis and H. taeanensis, respectively], a somewhat swollen root system, diurnal flowers, and absence of odor) and close geographic distribution. Our present allozyme data might support this hypothesis because, as presented in Table 5, most alleles of H. taeanensis, with exception of five unique alleles, Fe-2h, and Pgi-2c, are found in H. hakuunensis.

In summary, species of Hemerocallis are characterized by a small-scale genetic substructuring within populations, a high degree of within-population diversity, low genetic differentiation among populations in a species, and an extremely close genetic relationship among the species. A grouping of populations of each of the five Hemerocallis species in Korea based on their morphology and phenology accords with that of the allozyme data, though the populations had high genetic identities. These all suggest that the members of Hemerocallis may have recently derived from an ancestor or progenitor harboring high levels of genetic diversity.


View this table:
[in this window]
[in a new window]
  		Table 3. Continued

 

View this table:
[in this window]
[in a new window]
  		Table 5. Continued

 

   FOOTNOTES
 
1 The authors thank J. M. Chung, N. W. Lee, J. G. Hong, S. J. Yoon, S. G. Chung, S. S. Suh, and M. Y. Chung for providing support and assistance in collecting samples. Special thanks go to J. Nason, C. Parks, A. Schnabel, and E. Myers for helpful comments on earlier drafts of the manuscript. This research was supported in part by a NSF Dissertation Improvement grant BSR-8914430 to MGC and S. B. Jones and Korea Science and Engineering Foundation grants (96-0500-006-2 and 971-0505-029-2) to MGC. Back

2 Current address: Department of Agronomy, Iowa State University, Ames, Iowa 50011 USA. Back

3 Author for correspondence (FAX: +82-55-754-0086, email: mgchung{at}nongae.gsnu.ac.kr ). Back


   LITERATURE CITED
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 LITERATURE CITED
 
Aspinwall, N., and T. Christian. 1992 Clonal structure, genotypic diversity, and seed production in populations of Filipendula rubra (Rosaceae) from the northcentral United States. American Journal of Botany 79: 294–299.[CrossRef][ISI]

Broyles, S. B., and R. Wyatt. 1993 Allozyme diversity and genetic structure in southern Appalachian populations of poke milkweed, Asclepias exaltata. Systematic Botany 18: 18–30.

Brunsfield, S. J., D. E. Soltis, and P. S. Soltis. 1991 Patterns of genetic variation in Salix section Longifoliae (Salicaceae). American Journal of Botany 78: 853–869.

Cheliak, W. M., and J. P. Pitel. 1984 Technique for starch gel electrophoresis of enzyme from forest tree species. Petawawa National Forest Institute, Information Report P1-X-42. Canadian Forestry Service, Agriculture, Canada, 1–49. Chalk River, Ontario, Canada.

Chung, M. G. 1994a Genetic variation and population structure in Korean endemic species: III. Hosta minor (Liliaceae). Journal of Plant Research 107: 377–383.[CrossRef][ISI]

———. 1994b Genetic structure in Korean populations of Hosta capitata (Liliaceae). Journal of Plant Biology 37: 277–284.

———. 1994c Allozyme diversity and population genetic structure in Hosta jonesii (Liliaceae), a Korean endemic species. Korean Journal of Genetics 16: 147–156.

———. 1995a Genetic diversity in two island endemics, Hosta venusta and H. tsushimensis (Liliaceae). Journal of Japanese Botany 70: 322–327.

———. 1995b Low levels of genetic diversity within populations of Hosta clausa (Liliaceae). Plant Species Biology 9: 177–182.[CrossRef]

———, and H. G. Chung. 1994 Allozyme diversity and population genetic structure in Korean endemic plant species: II. Hosta yingeri (Liliaceae). Journal of Plant Biology 37: 141–149.

———, J. L. Hamrick, S. B. Jones, and G. S. Derda. 1991 Isozyme variation within and among populations of Hosta (Liliaceae) in Korea. Systematic Botany 16: 667–684.[CrossRef][ISI]

———, and S. S. Kang. 1994a Morphometric analysis of the Hemerocallis L. (Liliaceae) in Korea. Journal of Plant Research 107: 165–175.[CrossRef][ISI]

———, and ———. 1994b Hemerocallis hongdoensis (Liliaceae), a new species from Korea. Novon 4: 94–97.[CrossRef]

———, and ———. 1994c Genetic variation and population structure in Korean populations of Eurya japonica (Theaceae). American Journal of Botany 81: 1077–1082.[CrossRef][ISI]

———, and J. Noguchi. 1998 Geographic spatial autocorrelation of morphological characters of the Hemerocallis middendorffii complex (Liliaceae). Annales Botanici Fennici 35: 183–189.[ISI]

———, and K. B. Park. 1998 Spatial genetic structure in populations of Hosta capitata and H. minor (Liliaceae). Israel Journal of Plant Sciences 46: 181–187.

Clayton, J. W., and D. N. Tretiak. 1972 Amine-citrate buffers for pH control in starch gel electrophoresis. Journal of Fisheries Research Board of Canada 29: 1169–1172.

Cliff, A. D., and J. K. Ord. 1981 Spatial processes—methods and applications. Pion Limited, London, UK.

Cohen, S. 1986 The well-tinkered daylily. Garden September&October: 16–19.

Crawford, D. J. 1983 Phylogenetic and systematic inferences from electrophoretic studies. In S. D. Tanksley and T. J. Orton [eds.], Isozymes in plant genetics and breeding, part A, 257–287. Elsevier, New York, New York, USA.

———. 1989 Enzyme electrophoresis and plant systematics. In D. E. Soltis and P. S. Soltis [eds.], Isozymes in plant biology, 146–164. Dioscorides, Portland, Oregon, USA.

———, R. Ornduff, and M. C. Vasey. 1985 Allozyme variation within and between Lasthenia minor and its derivative species, L. maritima (Asteraceae). American Journal of Botany 72: 1177–1184.[CrossRef][ISI]

Dolan, R. W. 1994 Patterns in isozyme variation in relation to population size, isolation, and phytogeographic history in royal catchfly (Silene regia; Caryophyllaceae). American Journal of Botany 81: 965–972.[CrossRef][ISI]

Epperson, B. K. 1993 Recent advances in correlation analysis of spatial patterns of genetic variation. Evolutionary Biology 27: 95–155.

———, Z. Huang, and T.-Q. Li. 1999 Measures of spatial structure in samples of genotypes for multiallelic loci. Genetical Research, Cambridge 73: 251–261.

Ford, B. A., D. A. R. Mcqueen, R. F. C. Naczi, and A. A. Reznicek. 1998 Allozyme variation and genetic relationships among species in the Carex willdenowii complex (Cyperaceae). American Journal of Botany 85: 546–552.[Abstract]

Gottlieb, L. D. 1973 Genetic differentiation, sympatric speciation, and the origin of a diploid species of Stephanomeria. American Journal of Botany 60: 543–553.

———. 1981 Electrophoretic evidence and plant populations. Progress in Phytochemistry 7: 1–46.

Hamrick, J. L., and M. J. W. Godt. 1989 Allozyme diversity in plants species. In A. H. D. Brown, M. T. Clegg, A. L. Kahler, and B. S. Weir [eds.], Plant population genetics, breeding and genetic resources, 43–63. Sinauer, Sunderland, Massachusetts, USA.

———, and J. D. Nason. 1996 Consequences of dispersal in plants. In O. E. Rhodes, R. K. Chesser, and M. H. Smith [eds.], Population dynamics in ecological space and time, 203–236. University of Chicago Press, Chicago, Illinois, USA.

———, M. J. W. Godt, D. A. Murawski, and M. D. Loveless. 1991 Correlations between species traits and allozyme diversity: implications for conservation biology. In D. A. Falk, and K. E. Holsinger [eds.], Genetics and conservation of rare plants, 75–86. Oxford University Press, New York, New York, USA.

———, ———, and S. L. Sherman-Broyles. 1992 Factors influencing levels of genetic diversity in woody plant species. New Forests 6: 95–124.[CrossRef]

Haufler, C. H. 1985 Enzyme variability and modes of evolution in Bommeria (Pteridaceae). Systematic Botany 10: 92–104.

Holsinger, K. E., and L. D. Gottlieb. 1991 Conservation of rare and endangered plants: Principles and prospects. In D. A. Falk and K. E. Holsinger [eds.], Genetic and conservation of rare plants, 195–208. Oxford University Press, New York, New York, USA.

Hotta, M. 1986 Hemerocallis aurantiaca group in northern Kyushu, Japan. Acta Phytotaxonomica et Geobotanica 37: 17–21 (in Japanese).

———, M. Ito, and I. Okata. 1984 Anthesis of the genus Hemerocallis and its variation. Acta Phytotaxonomica et Geobotanica 35: 84–93 (in Japanese).

———, ———, and ———. 1985 Differentiation and species relationships of island populations of Hemerocallis around Kyushu, Japan. In H. Hara [ed.], Origin and evolution of diversity in plants and plant communities, 18–31. Academia Scientific Book Inc., Tokyo, Japan.

Kang, S. S. 1997 Morphometric analysis and population genetic structure of Hemerocallis (Liliaceae) in Korea. Master's thesis, Gyeongsang National University, Chinju, Korea.

———, and M. G. Chung. 1994 Hemerocallis hakuunensis (Liliaceae) in Korea. Sida 16: 23–31.

———, and ———. 1997a Genetic variation and population structure in Korean endemic species: IV. Hemerocallis hakuunensis (Liliaceae). Journal of Plant Research 110: 209–218.[CrossRef][ISI]

———, and ———. 1997b Hemerocallis taeanensis (Liliaceae), a new species from Korea. Systematic Botany 23: 427–431.

———, J. Noguchi, K. B. Park, and M. G. Chung. 1998 Allozyme diversity in Japanese populations of Hemerocallis thunbergii, H. middendorffii, and H. exaltata (Liliaceae). Nordic Journal of Botany 18: 581–587.[ISI]

Kawano, S. 1961 On the natural hybrid population of Hemerocallis. Canadian Journal of Botany 39: 667–681.

———, and J. Noguchi. 1973 Biosystematic studies on the genus Hemerocallis (Liliaceae). I. Introgressive hybridization between H. citrina v. vespertima and H. fulva sensu lato. Journal of the College of Liberal Arts, Toyama University, Japan, 6: 111–137.

Lee, H.-W., M. G. Chung, Y. Suh, and C.-W. Park. 2000 Allozyme variation and genetic relationships among species in Cimicifuga (Ranunculaceae) from Korea. International Journal of Plant Sciences 112: 139–144.

Levin, D. A., and H. W. Kerster. 1974 Gene flow in seed plants. Evolutionary Biology 7: 139–200.

Lewis, P. O., and D. J. Crawford. 1995 Pleistocene refugium endemics exhibit greater allozymic diversity than widespread congeners in the genus Polygonella (Polygonaceae). American Journal of Botany 82: 141–149.

Li, C. C., and D. G. Horvitz. 1953 Some methods of estimating the inbreeding coefficient. American Journal Human Genetics 5: 107–117.

Maki, M., and M. Masuda. 1994 Spatial genetic structure within two populations of a self-incompatible perennial, Chinographis japonica var. japonica (Liliaceae). Journal of Plant Research 107: 283–287.[CrossRef][ISI]

Matsuoka, M., and M. Hotta. 1966 Classification of Hemerocallis in Japan and it's vicinity. Acta Phytotaxonomica et Geobotanica 22: 25–43 (in Japanese).

Mayer, M. S., P. S. Soltis, and D. E. Soltis. 1994 The evolution of the Streptanthus glandulosus complex (Cruciferae): genetic divergence and gene flow in serpentine endemics. American Journal of Botany 81: 1300–1308.[CrossRef][ISI]

McClintock, K. A., and M. J. Waterway. 1994 Genetic differentiation between Carex lasiocarpa and C. pellita (Cyperaceae) in North America. American Journal of Botany 81: 224–231.[CrossRef][ISI]

Mitton, J. B., Y. B. Linhart, K. B. Sturgeon, and J. L. Hamrick. 1979 Allozyme polymorphism detected in mature needle tissue of ponderosa pine. Journal of Heredity 70: 86–89.[Abstract/Free Full Text]

Nakao, S., and K. Yamashita. 1956 Variation in some plant populations. In Syundan idengaku (Population genetics), Tokyo, Japan (in Japanese).

Nei, M. 1972 Genetic distance between populations. American Naturalist 106: 283–292.[CrossRef][ISI]

———. 1973 Analysis of gene diversity in subdivided populations. Proceedings of the National Academy of Sciences, USA 70: 3321–3323.[Abstract/Free Full Text]

———. 1977 F-statistics and analysis of gene diversity in subdivided populations. Annals Human Genetics 41: 225–233.[CrossRef]

Noguchi, J. 1986 Geographical and ecological differentiation in the Hemerocallis dumortieri complex with special reference to its karyology. Journal of Science of the Hiroshima University Series B Division 2 (Botany) 20: 29–193.

Park, K. B., S. S. Kang, and M. G. Chung. 1998 Spatial genetic structure in a population of Hemerocallis taeanensis (Liliaceae). Journal of Japanese Botany 73: 179–184.

Parker, K. C., and J. L. Hamrick. 1992 Genetic diversity and clonal structure in a columnar cactus, Lophocereus schottii. American Journal of Botany 79: 86–89.

Qiu, Y.-L., and C. R. Parks. 1994 Disparity of allozyme variation levels in three Magnolia (Magnoliaceae) species from the southeastern United States. American Journal of Botany 81: 1300–1308.

Rohlf, F. J. 1988 Numerical taxonomy and multivariate analysis system. Exeter Publishers, Setauket, New York, USA.

Sakai, A. K., and N. L. Oden. 1983 Spatial pattern of sex expression in silver maple (Acer saccharium L.): Morista's index and spatial autocorrelation. American Naturalists 122: 489–508.