HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Figures Only Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Related articles in Crop Science Similar articles in this journal Similar articles in ISI Web of Science Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via ISI Web of Science (8) Citing Articles via Google Scholar Google Scholar Articles by Chen, Y. Articles by Nelson, R. L. Search for Related Content PubMed Articles by Chen, Y. Articles by Nelson, R. L. Agricola Articles by Chen, Y. Articles by Nelson, R. L. Related Collections Soybean Plant Genetic Resources

PLANT GENETIC RESOURCES

Genetic Variation and Relationships among Cultivated, Wild, and Semiwild Soybean

^a Dep. of Crop Sciences, 1101 W. Peabody Dr., University of Illinois, Urbana, IL 61801
^b USDA-ARS, Soybean/Maize Germplasm, Pathology, and Genetics Research Unit, Dep. of Crop Sciences, 1101 W. Peabody Dr., University of Illinois, Urbana, IL 61801

^* Corresponding author (rlnelson{at}uiuc.edu).

	ABSTRACT

TOP NOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION REFERENCES

 
Some annual Glycine accessions are intermediate between the standard phenotypes of Glycine max (L.) Merr. and Glycine soja Sieb. & Zucc. and have been labeled semiwild. Few studies have examined both the genetic and phenotypic relationships among G. soja, G. max, and semiwild-types by combining morphological traits and DNA markers. The objectives of this research were to quantify genetic variation within G. soja, G. max, and semiwild accessions; to investigate the relationships among the G. soja, G. max, and semiwild accessions; and to examine the relationships among phenotypes on the basis of morphological traits and genotypes on the basis of DNA markers. Ninety-two semiwild, G. soja, and G. max accessions from the USDA Soybean Germplasm Collection were evaluated for 20 phenotypic traits and with 137 RAPD markers. Mahalanobis distances and a Jaccard genetic similarity matrix were calculated for phenotypic traits and DNA data, respectively. Nonhierarchical and hierarchical clustering as well as multidimensional scaling (MDS) were used to evaluate relationships among semiwild, G. soja, and G. max accessions. Principal component analysis was applied to identify the morphological traits that were most significant in separating the three groups. For the accessions examined, unique RAPD markers were found for each taxonomic type. Three clusters defined by either phenotypic or DNA data are highly consistent and strongly corresponded to G. soja, G. max, and semiwild classifications. On the basis of the analysis of RAPD data, G. soja accessions have the greatest genetic diversity and semiwild accessions the least. Glycine max and semiwild accessions are more closely related to each other than to G. soja accessions. These data will be useful in helping to define a core collection of annual Glycine.

Abbreviations: MDS, multidimensional scaling • MG, maturity group • PCA, principal component analysis • PCR, polymerase chain reaction • PI, Plant Introduction • RAPD, random amplified polymorphic DNA

	INTRODUCTION

TOP NOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION REFERENCES

 
THERE ARE TWO SPECIES usually recognized within the genus Glycine subgenus Soja, Glycine max and Glycine soja. On the basis of data from morphology (Palmer et al., 1987), cytogenetics (Hymowitz and Singh, 1987), phytoalexins (Keen et al., 1986), restriction endonuclease fragment analysis of mitochondrial DNA (Doyle, 1988), ribosomal RNA (Doyle and Beachy, 1985), chloroplast DNA (Shoemaker et al., 1986), and sequences from the ITS region of nuclear ribosomal DNA (Kollipara et al., 1997), G. soja is considered the ancestor of G. max. Besides G. max and G. soja, an intermediate form sometimes known as G. gracilis Skvortz. has been described. This form has numerous characteristics intermediate between G. max and G. soja and was first proposed as a new species by Skvortzow (1927). Fukuda (1933) proposed that G. gracilis is an intermediate evolutionary type between G. soja to G. max, but Hymowitz (1970) suggested that G. gracilis is a hybridization product of G. max and G. soja. The latter hypothesis was supported by Broich and Palmer (1981) on the basis of the results from their study of the frequency and distribution of 10 alleles among G. max, G. soja, and G. gracilis accessions. On the basis of numerical taxonomic analysis, Broich and Palmer (1980)(1981) recommended that the designations G. max and G. gracilis both be utilized. They reasoned that G. gracilis phenotypes can be distinguished from G. max and they represent an intermediate form of domesticated soybean. In addition, G. max and its semiwild relative should be regarded as taxonomically distinct from G. soja since both are domesticated.

Hermann (1962) removed G. gracilis from the species rank and incorporated it into G. max on the basis of classical taxonomy. Several studies support the elimination of G. gracilis as a separate species. Singh and Hymowitz (1989) demonstrated that G. max, G. soja, and G. gracilis all hybridized readily, and F₁ seeds produced viable, vigorous, and fertile plants with normal meiotic pairing. Wang (1976) suggested all three classifications in subgenus Soja could be a single species since they were not reproductively isolated, but on the basis of cultivated status, he recommended that G. soja be kept as a species and G. gracilis reclassified as G. max.

Dae et al. (1995) applied isozyme and RAPD techniques to evaluate genetic variation within the subgenus Soja and concluded on the basis of morphological appearances that the intermediate forms of G. max were also intermediate between G. max and G. soja on the basis of genotypic measurements. Fei and Chen (1996) analyzed genetic diversity of the Glycine genus with RAPD markers using 21 accessions from 10 species of the Glycine subgenus and the three species of the Soja subgenus (G. max, G. gracilis, and G. soja) with eight primers. In this analysis, they found that the three species within the Soja subgenus were clustered as one group with G. gracilis classified a subgroup within G. max. This research supports the idea that there should only be one species, and earlier Smartt (1984) had proposed that G. max, G. soja, and G. gracilis should all be classified as subspecies. Although there are many arguments about the designations of species in subgenus Soja, it is well accepted that G. soja is the ancestor of cultivated soybean and most taxonomists have kept G. max and G. soja as separate species.

Several traits have distinct differences between G. max and G. soja accessions. In general, G. soja has much smaller seeds (<3.0 g 100 seeds^–1) than G. max (generally >9.0 g 100 seeds^–1). Glycine soja also has viney and twining stems, severe shattering before plant maturity, and impermeable seed coats, which are all rare in G. max. Glycine soja also has much lower oil and oleic acid concentration, and higher linolenic acid concentration. There are many accessions in annual Glycine collections that are intermediate between the typical G. soja and G. max types. Chang et al. (1999) reported that among the 17 613 accessions of the Chinese G. max collection 1.5% of the accessions have 100-seed weights of less than 6.0 g, and 33% of the accessions were between 6.1 and 12.0 g 100 seeds^–1. Dong et al. (1999) examined 6172 G. soja accessions in the Chinese wild soybean collection and found that 8.5% of the accessions have 100-seed weights of more than 5.0 g. The appropriate classification of these intermediate types is not well defined. The objectives of this research were to quantify genetic variation within the G. soja, G. max, and semiwild Glycine accessions; to investigate the relationships among the G. soja, G. max, and semiwild accessions; and to examine the relationships among phenotypes on the basis of morphological traits and genotypes on the basis of DNA markers.

	MATERIALS AND METHODS

TOP NOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION REFERENCES

 
Thirty semiwild, 31 G. max, and 31 G. soja accessions, previously classified on the basis of morphological traits when the accessions were initially evaluated, were selected from the USDA Soybean Germplasm Collection for this study. Accessions within each group were selected to have similar origins and maturity dates (Table 1). All G. max accessions are primitive types that predate scientific plant breeding. The lines were evaluated at Urbana, IL, in 1999 and 2000. The G. max and semiwild accessions were grown in three replications in one-row plots 2.5 m long and 0.75 m apart, and the G. soja lines were grown in hill plots 0.75 m apart in an aphid-proof cage with one replication in 1999. In 2000, the experiment was repeated with three replications for all entries.

Seed composition measurements included protein and oil concentration, and concentration of the following fatty acids: palmitic, stearic, oleic, linoleic, and linolenic. Nitrogen content of whole seed was determined with a LECO FP-428 Nitrogen Determinator (LECO Corp., St. Joseph, MI). The 6.25 conversion factor was used to calculate protein concentration on a dry weight basis. Oil concentration (dry weight basis) of whole seed was determined with a 5 MHz nuclear magnetic resonance spectrometer (Newport Oxford Instruments, Newport Pagnell, England). Fatty acid methyl esters were prepared from chloroform/hexane/methanol (8:5:2, v/v/v) extracts of crushed seed by transmethylation with sodium methoxide. Fatty acid composition was determined with a Hewlett-Packard 5890-II (Palo Alto, CA) gas chromatograph equipped with dual flame ionization detectors, and a 0.53-mm by 30-m AT-Silar capillary column (Alltech Associates, Deerfield, IL). Authentic fatty acids were used for calibration.

Genomic DNA was isolated from the first trifoliate leaves of five greenhouse grown seedlings for each accession. Harvested leaves were placed in 15-mL screw-cap tubes and frozen at –80°C before lyophilizing the tissue. Four glass beads were added to each tube and shaken on a shaker for 3 min. DNA was extracted by the CTAB (hexadecyltrimethyl ammonium bromide) method of Kisha et al. (1997). The DNA concentration of all extracted samples was calculated from spectrophotometer readings at wavelengths of 260/280 and adjusted to a concentration of 10 ng µL^–1. Forty-four decanucleotide primers from Operon Technologies Inc. (Alameda, CA) were chosen for this study (Table 2). These included 35 primers of a core set identified by Thompson and Nelson (1998) and nine randomly selected primers. The amplification protocol of Kresovich et al. (1994) was used with minor modifications. Amplified products were separated by 1% (w/v) agarose gels in 1x Tris-acetate buffer for 2.5 h at 125 V with constant power, stained with ethidium bromide, and visualized under UV light.

RAPD fragments were scored as either present (1) or absent (0). Jaccard's coefficient was used to measure the distance between each pair of genotypes with the following formula: S_ij = a/(a + b + c), where a is the number of common bands; b is the number of bands present in first accession and absent in the second; and c is the number of bands absent in first accession and present in the second. D_ij = 1 – S_ij was calculated as a measure of dissimilarity.

A hierarchical cluster analysis was performed on the 92 by 92 genetic dissimilarity matrix using the WARD option of PROC CLUSTER of PC SAS (SAS Institute, 1999). Mean distances within and between clusters were calculated using a SAS Interactive Matrix Language (SAS/IML, SAS Institute, 1999) program provided by D.Z. Skinner (personal communication, 2001). Values of the cubic clustering criterion (CCC), pseudo F statistic (PSF), and Hotelling's pseudo T² statistic were also considered for defining optimum cluster numbers (SAS Institute, 1999). A nonhierarchical cluster analysis procedure, VARCLUS option of PROC CLUSTER in PC SAS (SAS Institute, 1999), was also applied to the original fragment data to divide the accessions into nonoverlapping clusters. The data were also subjected to principal component analysis.

The matrix of genetic distances generated from Jaccard's genetic dissimilarity coefficient was subjected to multidimensional scaling (MDS) (Shepard, 1974) by the MDS procedure in PC SAS (SAS Institute, 1999). The ABSOLUTE option was used to maintain the scale of 0 and 1 for making interpretation and graphing easier. The criteria are similar to that described by Thompson et al. (1998) and Gizlice et al. (1996). To evaluate the effectiveness of 2 to 22 dimensions, the goodness of fit criterion (R²) between the original data and the predicted values that were derived from the MDS coordinates was used. The best MDS analysis was considered to be the fewest dimensions that resulted an R² > 0.95 with the original genetic distance matrix. The matrix of the Mahalanobis distances from twelve phenotypic traits was also subjected to multidimensional scaling.

	RESULTS AND DISCUSSION

TOP NOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION REFERENCES

 
Variation Based on Phenotypic Data
On the basis of phenotypic data, G. max has the greatest diversity and G. soja has the least. All of the evaluated qualitative traits are uniform for G. soja except for pubescence form. Purple flowers, tawny pubescence color, normal pubescence density, seed coat bloom, and black pod, seed coat, and hilum color are common for all G. soja entries. Twelve quantitative traits (lodging, shattering, leaflet shape, stem ratio, seed weight, and seed concentrations of protein, oil, and five fatty acids) were subjected to analysis of variance (Table 3). Statistically significant differences were found between years for all traits except stem ratio, protein, and oleic acid concentration; however, the differences between the years were quite small for most traits. The means of the three taxonomic classes for all 12 traits were nearly all significantly different, but most of the traits have overlapping ranges across the three classes (Table 3). Seed weight, oil concentration, oleic, and linolenic acid concentration have highly significant differences among the three classes and little or no overlap in ranges of values (Table 3). Glycine soja has a viney stem that is never erect, severe shattering, a small stem ratio (<4.5), low seed weight (<2.5 g 100 seeds^–1), low oil concentration (<130 mg g^–1), low oleic acid concentration (<140 mg g^–1), and high linolenic acid concentration (>140 mg g^–1) (Table 3). Glycine max is variable for lodging and shattering, has a high stem ratio (>4.5), larger seed weight (>9.0 g 100 seeds^–1), high oil concentration (>185 mg g^–1), high oleic acid concentration (>190 mg g^–1), and low linolenic concentration (<95 mg g^–1). The semiwild accessions are intermediate between G. soja and G. max for most traits (Table 3).

Five dimensions in multidimensional scaling adequately captured the information in the original Mahalanobis distance matrix (R² = 0.97). Data from both years resulted in nearly identical MDS plots and results were consistent with the Ward's clustering method. The first dimensions accounted for 59% of the total variation and the two-dimensional plot showed that G. max and G. soja were in two distinct groups with the semiwild accessions generally distributed between these two groups (Fig. 1). The majority of semiwild accessions were clearly separated from the two species, but G75, G06, G70, and G16, the semiwild accessions in the predominantly G. max cluster by Ward's method, are positioned in the G. max group (Fig. 1). In the fifth dimension of the MDS, it was possible to separate G06 from the G. max group. Two other lines (G10 and G45) were on the boundary between the semiwild and G. max groups, and it is difficult to define them as either G. max or semiwild on the basis of this analysis, but in the fourth dimension G10 could be distinguished from the G. max accessions. The G. soja accessions were more tightly clustered than the other groups indicating less variability for these phenotypic traits (Fig. 1).

View larger version (15K): [in this window] [in a new window]   Fig. 1. Two dimensional scatter plot of 92 accessions of G. max, G. soja, and semiwild accessions obtained from multidimensional scaling analysis of genetic distance estimates based on the Mahalanobis distance matrix of 12 phenotypic traits measured in 2000.

The VARCLUS analysis resulted in four clusters with both 1999 and 2000 data. Cluster 1 and cluster 2 included 29 semiwild accessions. Although half of the accessions in cluster 1 weigh less than 4.0 g 100 seeds^–1, the mean 100-seed weight of cluster 1 (6.3 g) is nearly the same as cluster 2 (5.8 g). The range of seed weights in cluster 2 is from 4.4 to 8.9 g 100 seeds^–1. Cluster 3 contained all of the 31 G. max lines and one semiwild accession (G06) with a 12.9 g 100-seed weight mean. Cluster 4 was composed of all 31 G. soja lines and had the smallest 100-seed weight (mean = 1.4 g) compared with other three clusters. All four analytical procedures (Ward's, MDS, PCA, and VARCLUS) identified G06 as not being part of the semiwild group. G75, G70, and G16 were identified as such by all but the VARCLUS procedure. From the phenotypic data, we can conclude that the G. max and G. soja groups are clearly distinct from each other. Those classified as semiwild form an intermediate but not always unambiguous grouping.

Genetic Relationships Based RAPD Profiles
Forty-four primers generated 137 polymorphic fragments out of a total of 231 fragments (Table 2). The percentage of polymorphism (59%) is higher than reported by Thompson and Nelson (1998) for only G. max (30%) but only slightly higher than in the Li and Nelson (2001) data for both of G. max and G. soja (56%). Fragments OPO01₇₀₀, OPO01₈₅₀, OPX05₄₅₀, OPH02₂₁₀₀, OPO05₂₁₅₀, OPG11₂₅₀₀, and OPG11₁₉₀₀ were not found in one or more of the three classes, and a unique fragment was found within each class (Table 4). OPG11₂₅₀₀ was only found within the semiwild accessions and occurred in a majority of those accessions. OPG11₁₉₀₀ and OPO05₂₁₅₀ were present in 60% or more of the semiwild accessions, were totally absent from the G. max lines, and existed in low frequencies within G. soja. Neither the theory that semiwild-types are evolutionary intermediates between G. max and G. soja nor that they are products of more recent hybridizations provides a good explanation for unique fragments in this class of accessions, especially not for a fragment that occurred in more than 60% of the semiwild lines. Although the accessions in this study represent only a small portion of the available germplasm, this unique RAPD fragment indicates that the semiwild accessions are in some way genetically distinct from the standard types of the two annual species. OPH02₂₁₀₀ and OPX05₄₅₀ occurred in over 40% of the G. soja lines, but OPO01₈₅₀ was the only fragment that was found only in G. soja lines. The low frequency of OPO01₈₅₀ in these G. soja accessions may be one explanation for why it did not occur in either of the other groups that are derived from G. soja. OPO01₇₀₀ was found only in the G. max lines. Li and Nelson (2001) also reported this as a unique band in G. max. It is possible that changes in this region of the genome are partially responsible for the evolution of G. max. Extensive research would be required to confirm that OPG11₂₅₀₀, OPO01₇₀₀, and OPO01₈₅₀ are unique markers for semiwild, G. max, and G. soja, respectively, but they do demonstrate the genetic separation of these closely related groups. Removing these taxon-specific fragments from the analysis did not change the cluster groupings. The pattern of divergence among the three classes was primarily attributable to differences in fragment frequencies.

View larger version (18K): [in this window] [in a new window]   Fig. 2. Two dimensional scatter plot of 92 accessions of G. max, G. soja, and semiwild accessions obtained from multidimensional scaling analysis of genetic distance estimates based on Jaccard's genetic dissimilarity matrix of 231 RAPD fragments generated by 44 primers.

The semiwild group has the smallest within-cluster genetic distance (0.107), whereas the G. soja group has the largest genetic distance (0.219). These results agree with Maughan et al. (1995) and Li and Nelson (2001) showing the greatest genetic diversity in G. soja. The genetic distance between the semiwild cluster and the G. max cluster (0.199) was the least distance among the clusters indicating that the semiwild accessions have a closer relationship to G. max than to G. soja. Broich and Palmer (1980) also showed the semiwild and the G. max to be more closely related than either was to G. soja. If the semiwild-types are evolutionary intermediates between G. soja to G. max, theoretically semiwild-types should have a greater genetic variation than G. max and less than G. soja. If the semiwild accessions are hybridization products presumably only a small proportion of the plants from G. soja and G. max would have hybridized, which would cause the semiwild-type to have a narrower genetic base than either of the parental gene pools. If these assumptions are true, the data from this research support the theory that semiwild accessions are hybridization products.

The origin information for many of the G. soja lines is more precise than for the other accessions in this study. S50, S60, S72, and S73 were collected from the same field in Shenyang, China (41.48°N) (Bernard et al. 1989). The genetic distances among these four accessions range from 0.096 to 0.186. S22 and S87 were collected in the same pasture near Gongzhuling, Jilin, China (43.32°N) (Bernard et al., 1989). They were phenotypically similar but the RAPD profiles were quite different. The genetic distance between these two lines was 0.228, which is higher than the within G. soja cluster mean genetic distance (0.219). These data demonstrate that G. soja lines collected from the same area can be similar but in some cases are genetically quite distinct. It may be necessary to collect multiple samples within G. soja populations to sample completely the genetic diversity.

In this study, the three clusters defined by phenotypic data and DNA profiles are highly consistent and strongly correspond to the original G. soja, G. max, and semiwild classifications. The current methods for defining G. max, G. soja, and semiwild are effective, and both phenotypic and DNA marker data can be used to classify Glycine accessions. Some accessions (G12, G13, G63, M92, G06, G16, G75, G37, and G70), mostly semiwild, were not consistently classified (Table 5). Four semiwild accessions (G06, G16, G70, and G75) were grouped with G. max on the basis of both phenotypic and genotypic data; three semiwild (G12, G13, and G63) and one G. max (M92) lines were clustered with G. soja; and one semiwild line (G37) was grouped with G. max on the basis of genotypic data. Although the three semiwild lines (G12, G13, and G63) share some attributes with G. soja, these accessions do not have the plant type that would justify changing their species classification. The five semiwild lines that were grouped with G. max are on that ambiguous boundary between semiwild and cultivated and could be included in either group. The results from this study showed that no single trait can be used to distinguish the semiwild soybean. Several characteristics seem to be unique for G. soja, and phenotypically seed weight and stem characteristics can be definitive. On the basis of classical taxonomy and cytogenetics, most authors have supported removing G. gracilis from the species rank and incorporating it into G. max (Hermann, 1962; Wang, 1976; Singh and Hymowitz, 1989).

	ACKNOWLEDGMENTS

 
The authors thank Dr. Richard F. Wilson for providing the chemical composition data used in this research.

	NOTES

TOP NOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION REFERENCES

 
Mention of a trademark, proprietary product, or vendor does not constitute a guarantee or warranty of the product by the USDA or the University of Illinois and does not imply its approval to the exclusion of other products or vendors that may also be suitable.

Received for publication December 16, 2002.

	REFERENCES

TOP NOTES ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION REFERENCES

 

• Bernard, R.L., G.A. Juvik, and R.L. Nelson. 1989. USDA Soybean Germplasm Collection Inventory. Vol. 2. International Agricultural Publications. INTSOY Series Number 31. INTSOY, University of Illinois at Urbana-Champaign.
• Broich, S.L., and R.G. Palmer. 1980. A cluster analysis of wild and domesticated soybean phenotypes. Euphytica 29:23–32.
• Broich, S.L., and R.G. Palmer. 1981. Evolutionary studies of the soybean: The frequency and distribution of alleles among collections of Glycine max and G. soja of various origin. Euphytica 30:55–64.
• Chang, R.Z., L.J. Qiu, J.Y. Sun, Y. Chen, X.H. Li, and Z.Y. Xu. 1999. Collection and conservation of soybean germplasm in China. p. 172–176. In Proc. World Soybean Research Conference VI. Chicago, IL. 4–7 August 1999. National Soybean Research Lab., Urbana, IL.
• Dae, H.P., K.M. Shim, Y.S. Lee, W.S. Ahn, J.H. Kang, and N.S. Kim. 1995. Evaluation of genetic diversity among the Glycine species using isozymes and RAPD. Korean J. Genet. 17:157–168.
• Dong, Y.S., H. Sun, B. Zhuang, L. Zhao, and M. He. 1999. The genetic diversity in annual wild soybean. p. 147–155. In Proc. World Soybean Research Conference VI. Chicago, IL. 4–7 August 1999. National Soybean Research Lab., Urbana, IL.
• Doyle, J.J. 1988. 5S ribosomal gene variation in soybean (Glycine) and its progenitor. Theor. Appl. Genet. 75:621–624.
• Doyle, J.J., and R.N. Beachy. 1985. Ribosomal gene variation in soybean (Glycine) and its wild relatives. Theor. Appl. Genet. 70:369–376.
• Fei, D.W., and S.Y. Chen. 1996. Reestablish the relationships of species of Glycine Genus by RAPD. J. Hered. 23(6):460–468.
• Fukuda, Y. 1933. Cytogenetical studies on wild and cultivated Manchurian soybean (Glycine L.). Jpn. J. Bot. 6:489–506.
• Gizlice, Z., T.E. Carter, Jr., and J.W. Burton. 1996. Genetic diversity patterns in North American public soybean cultivars based on coefficient of parentage. Crop Sci. 36:753–765.[Abstract/Free Full Text]
• Hermann, F.J. 1962. A revision of genus Glycine and its immediate allies. USDA Tech. Bull. 1268. U.S. Gov. Print. Office, Washington, DC.
• Hymowitz, T. 1970. On the domestication of the soybean. Econ. Bot. 23:408–421.
• Hymowitz, T., and R.J. Singh. 1987. Taxonomy and speciation. p. 23–48. In J.R. Wilcox (ed.) Soybeans: Improvement, production, and uses. Agron. Monogr. 16. 2nd ed. ASA, Madison, WI.
• Keen, N.L., R.L. Lyne, and T. Hymowitz. 1986. Phytoalexin production as a chemosystematic parameter with Glycine ssp. Biochem. Syst. Ecol. 14:481–486.
• Kisha, T., C.H. Sneller, and B.W. Diers. 1997. Relationship between genetic distance among parents and genetic variance in populations of soybean. Crop Sci. 37:1317–1325.[Abstract/Free Full Text]
• Kollipara, K.P., R.J. Singh, and T. Hymowitz. 1997. Phylogenetic and genomic relationships in the genus Glycine Willd. Based on sequences from the ITS region of nuclear rDNA. Genome 40:57–68.[Medline]
• Kresovich, S., W.F. Lamboy, R. Li, A.K. Szewc-McFadden, and S.M. Bliek. 1994. Application of molecular methods and statistical analysis for discrimination of accessions and clones of vetiver grass. Crop Sci. 34:805–809.[Abstract/Free Full Text]
• Li, Z., and R.L. Nelson. 2001. RAPD marker diversity among soybean and wild soybean accessions from four Chinese provinces. Crop Sci. 41:1337–1347.[Abstract/Free Full Text]
• Maughan, P.J., M.A. Saghai Maroof, and G.R. Buss. 1995. Microsatellite and amplified sequence length polymorphisms in cultivated and wild soybean. Genome 38:715–723.[Medline]
• Palmer, R.G., K.E. Newhouse, R.A. Graybosch, and X. Delannay. 1987. Chromosome structure of wild soybean (Glycine soja Sieb. & Zucc.) accessions from China and the Soviet Union. J. Hered. 78:243–247.[Abstract/Free Full Text]
• SAS Institute. 1999. SAS/STAT user's guide, Version 8.0, First ed., SAS Inst., Inc., Cary, NC.
• Shepard, R.N. 1974. Representation of structure in similarity data: Problems and prospects. Psycometrika 39:373–421.
• Shoemaker, R.C., P.M. Hatfield, R.G. Palmer, and A.A. Atherly. 1986. Chloroplast DNA variation and evolution in the genus Glycine subgenus Soja. J. Hered. 77:26–30.[Abstract/Free Full Text]
• Singh, R.J., and T. Hymowitz. 1989. The genomic relationships between G. soja Sieb. and Zucc. G. max (L.) Merr. and ‘G. gracilis’ Skvortz. Plant Breed. 103:171–173.
• Skvortzow, B.V. 1927. The soybean-wild and cultivated in Eastern Asia. Proc. Manchurian Res. Soc. Pub. Ser. A. Nat. History Sec. 22:1–8.
• Smartt, J. 1984. Gene pools in grain legumes. Econ. Bot. 38:24–35.
• Thompson, J.A., and R.L. Nelson. 1998. Core set of primers to evaluate genetic diversity in soybean. Crop Sci. 38:1356–1362.[Abstract/Free Full Text]
• Thompson, J.A., R.L. Nelson, and L.O. Vodkin. 1998. Identification of diverse soybean germplasm using RAPD markers. Crop Sci. 38:1348–1355.[Abstract/Free Full Text]
• Wang, C.L. 1976. Review on the classification of soybean (in Chinese). Acta Phytotaxon. Sin. 14:22–30.

Related articles in Crop Science:

THIS ISSUE IN CROP SCIENCE

Crop Science 2004 44: 1-4. [Full Text]  

This Article Abstract Figures Only Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Related articles in Crop Science Similar articles in this journal Similar articles in ISI Web of Science Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via ISI Web of Science (8) Citing Articles via Google Scholar Google Scholar Articles by Chen, Y. Articles by Nelson, R. L. Search for Related Content PubMed Articles by Chen, Y. Articles by Nelson, R. L. Agricola Articles by Chen, Y. Articles by Nelson, R. L. Related Collections Soybean Plant Genetic Resources