| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Dep. of Crop Science, North Carolina State Univ., Raleigh, NC 27695-7631
b USDA-ARS and Dep. of Crop Science, North Carolina State Univ., Raleigh, NC 27695-7631
* Corresponding author (tommy_carter{at}ncsu.edu)
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
Abbreviations: CP, coefficient of parentage • NA, North American • PS, phenotypic similarity • N, northern NA • S, southern NA • NEC, northeastern China • NC, northern China, i.e., Huanghe, Huaihe, and Haihe valleys • SC, southern China • SEPEL, Southeastern Plant Environment Laboratories
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
Carter et al. (2000) suggested that Chinese cultivars may be an important reservoir of diversity with which to expand the genetic base of NA breeding. In that regard, China has released 651 cultivars from 1923 to 1995 (Cui et al., 1999). Through pedigree analysis, Cui et al. (2000a)(b) established that Chinese cultivars were derived from a far greater number of ancestors than were NA cultivars (339 vs. 80 ancestors) and that the two regions had few identifiable ancestors in common. Carter et al. (2000) also demonstrated that several Chinese cultivars performed well agronomically in North America.
In China, soybean breeders have recognized the potential importance of China and NA as mutual reservoirs of diversity. They have released 67 cultivars since 1974 with at least 25% of the pedigree derived from modern NA cultivars (Cui et al., 1999, 2000a). Many of these cultivars are high yielding, a finding that is consistent with the general notion that modern Chinese and NA cultivars may contrast in a beneficial way.
Despite important signals pointing to the desirability of exotic cultivars in Chinese and NA soybean breeding, the true utility of cross breeding these two cultivar pools, especially in terms of NA yield improvement, remains an open question. No NA cultivar has been developed from modern Chinese cultivars. No unique yield genes have been identified in these two breeding pools.
The uniqueness of Chinese and NA breeding pools have been elucidated to date primarily through pedigree analysis. The value of cross-breeding Chinese with NA cultivars will be greater than otherwise if the two groups of cultivars are distinct not only in terms of published pedigree, but in the underlying genetics as well. A limiting factor in the pedigree analyses of diversity for Chinese and NA cultivars is that, although the genetic bases of North America and China contrast in size and specific members, most members of these two bases trace their origins to China. Thus, it is possible that Chinese and NA cultivars may be derived from a common preexisting genetic base despite the lack of supporting pedigree evidence. In that regard, as many as 20 000 landraces were available in China by 1900 for use as parents in modern NA and Chinese breeding. There is no clear historical record which relates the selection of founding stock for Chinese and NA breeding. Relatively few of these landraces were used as breeding stock in North America or China, while many of these landraces are preserved in China, but the genetic structure of this large collection is not well understood (Chang et al., 1992, 1999).
One approach to elucidate underlying genetic differences between NA and Chinese breeding pools is to estimate genetic distinctness between them on the basis of measures other than pedigree. Molecular markers have been used to estimate genetic differences in germplasm accessions of soybean and other crops (Li et al., 2001; Thompson et al., 1997; Autrique et al., 1996; Johns et al., 1997). Preliminary analysis of simple sequence repeat (SSR) markers indicates that Chinese and NA soybean cultivars form fairly distinct breeding pools (Carter et al., 2000). Li et al. (2001) showed that ancestors of NA and Chinese breeding differ in terms of random amplified polymorphic DNA (RAPD) markers.
Phenotypic differences may also elucidate genetic differences. In the context of Chinese and NA breeding, we suggest that distinctions between the two pools, exemplified by pedigree and DNA marker analysis, should also be expressed phenotypically. Theoretically, phenotypic diversity should approximate genetic diversity. As the number of phenotypic traits increases in a comparison of breeding pools, the number of genes involved in the control of phenotypic traits should increase accordingly and, thereby, improve the utility of phenotypic diversity in predicting genotypic diversity. Employing this concept, van Beuningen and Busch (1997), Johns et al. (1997), and Autrique et al. (1996) used morphological, developmental, and physiological traits to create distance measures and examine genetic diversity in large collections of crop genotypes. Grafius et al. (1976) and Grafius (1978) were among the first to apply this concept to practical breeding by employing cultivar differences in morphological traits to select genetically diverse breeding pairs. In their work, the limited backcrossing of contrasting quantitative seed traits (presumably controlled by an array of genes on multiple linkage groups) from one cultivar into another tended to improve seed yield.
Recently, many modern Chinese cultivars were evaluated extensively in a series of field tests in North America for yield and other agronomic traits (Carter et al., 2000). However, we could find no comparison of genotypic diversity in NA and Chinese cultivars on the basis of a wide array of phenotypic traits. The objectives of this study were to (i) quantify diversity for 72 representative modern NA and Chinese cultivars by a phenotypic similarity (PS) index based on biochemical, morphological, and agronomic traits, (ii) compare Chinese and NA cultivars for PS through cluster analysis, and determine the relation between PS and previously reported coefficient of parentage (CP) estimates, and (iii) incorporate results into guidelines for management of the contrasting breeding programs as reservoirs of diversity.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
|
|
Traits Evaluated
The traits measured before flower initiation were as follows. (i) Leaf length and leaf width: the most recently fully developed leaves (2nd and 3rd trifoliate leaves) and the petioles were removed from each plant at 28 DAP. Leaflets were measured for length and width. (ii) Leaf Ratio: ratio of leaf length to width. Area per leaf: the area of a leaf determined using a leaf area meter for those leaflets employed in length measurements. (iii) Petiole length: petioles of leaves used in leaf area measurements were excised and measured. (iv) Dry weight per leaf: leaf samples were dried at 60°C for 24 h and weighed. (v) Specific leaf weight: the ratio of dry weight per leaf to area per leaf. (vi) Leaf nitrogen content: leaf samples from the leaf area measurement were dried and ground. The N content of samples was determined by the Kjeldahl method (Nelson and Sommers, 1973; Glowa, 1974). (vii) Chlorophyll a, b, and total content: three leaf disks 6.5 mm in diameter were punched from each leaflet of the 4th trifoliate. The chlorophyll content of the samples was determined by means of the technique described by Moran and Porath (1980) and Moran (1982). (viii) Chlorophyll ratio: ratio of chlorophyll a to b. (ix) Vegetative plant height and node number: height and number of nodes recorded at 28 d after planting. (x) Vegetative internode length: the ratio of plant height to number of nodes at 28 d after planting.
The traits measured after maturity were as follows. (i) Stem diameter: main stem diameter measured between the first and second nodes using a caliper. (ii) Internode length: the ratio of plant height to number of nodes at harvest. Number of seeds per pod: ratio of number of seeds per plant to number of pods per plant. (iii) Hundred-seed weight: weight/100 seed. (iv) Seed protein content and seed oil content: seed protein and oil contents were determined by near infra-red (NIR) analysis at USDA Northern Regional Research Center in Peoria, IL. All seeds were yellow, which is a prerequisite for the NIR analysis. (v) Palmitic acid (16:0), stearic acid (18:0), oleic acid (18:1), linoleic acid (18:2), and linolenic acid (18:3) contents in seed oil were determined by gas chromatography as described by Carver et al. (1984).
Days to flowering and days to maturity from planting, flower, and pubescence color were also recorded. Maturity Group was determined from a field study (Carter et al., 2000). Traits were measured for all six chambers with the following exceptions. Protein and oil content, and fatty acid composition of oil were determined for four chambers. Chlorophyll, plant height, and leaf N were determined for five chambers.
Statistical Analysis
Analysis of variance and least squares genotypic means were computed for all traits by the GLM procedure of SAS (SAS Institute, 1985b). Normality of genotypic means was tested by the UNIVARIATE procedure of SAS (SAS Institute, 1985a). The Shapiro-Wilk statistic, W, indicated that genotypic means approximated normal distributions except for the trait, leaf ratio, which was dropped from further analysis.
Principal component analysis was performed on the standardized least squares genotypic means to calculate eigenvalues, eigenvectors, and principal component scores calculated by the PRINCOMP procedure of SAS (SAS Institute, 1985b). Pairwise genetic distances among the 72 genotypes were computed as follows:
 |
Where, Dij = genetic distance between ith and jth genotypes; yik and yjk are the kth principal component scores for Genotypes i and j; 
k is the kth largest eigenvalue; q is the smallest eigenvalue that is >1.0 (Goodman, 1972; Gizlice et al., 1993).
Phenotypic similarity was derived from genetic distance by the formula:
 |
Where, PSij = phenotypic similarity between ith and jth genotypes; Dij = genetic distance between ith and jth genotypes; and Dmax is the largest genetic distance.
Ward's minimum variance method and eleven other clustering techniques were applied to PS scores, employing the CLUSTER procedure of SAS (SAS Institute, 1985b). A separate additional nonhierarchical clustering approach was also used, where multidimensional scaling (MDS) was applied to produce Euclidean coordinates (71 dimensions) based on the 72 by 72 PS matrix (SAS Institute, 1992; Gizlice et al., 1996). The nonhierarchical cluster analysis FASTCLUS procedure was then applied to MDS-derived 71-dimensional Euclidean coordinates (SAS Institute, 1985b). For each FASTCLUS analysis, it was necessary to specify the number of clusters desired prior to conducting the analysis. To find an optimum analysis, we performed 24 separate FASTCLUS analyses, specifying 2 to 25 clusters. The MEANS procedure of SAS was used to calculate mean PS within and among clusters for each analysis (SAS Institute, 1985a).
To determine the importance of a discriminator, such as cluster, maturity group, geographic origin, and yield level in explaining PS, we computed the percentage of variation accounted for by the discriminator in the PS matrix. An arithmetic shortcut described by Gizlice et al. (1996) was adopted to compute the coefficient of determination (R2) by means of PROC GLM (SAS Institute, 1985b). Correlation analysis was applied between CP and the phenotypic similarity measures derived from the phytotron data by the CORR procedure in SAS (SAS Institute, 1985b).
The PS matrix for the 72 cultivars was subjected to the multidimensional scaling (MDS) procedure of SAS with options LEVEL = ABSOLUTE, DIMENSION = 2, and SIMILAR = 1 (SAS Institute, 1992). Two-dimensional graphs were constructed from MDS-derived coordinates (Table 1) by the PLOT procedure (SAS Institute, 1985a). The graphs depicted approximate genetic distances among entries conveniently in the same units as the original PS values (Gizlice et al., 1996).
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
The 25 NA public cultivars were selected as representative of NA breeding. Seventeen were chosen to represent the nine major clusters of NA soybean cultivars identified by Gizlice et al. (1996), on the basis of pedigree analysis, and were released from 1978 to 1988. The remaining eight NA cultivars, released from 1989 to 1997, were selected as representative of more recent cultivars in both northern and southern USA. The only exception was the cultivar Johnston which was not a member of a cluster and released in 1983, but was included because of its high yield potential. Maturity ranged from 00 to VIII.
Use of Phytotron to Minimize Maturity Effects
In soybean, similarity measures based on phenotypic data are difficult to employ because of the extreme cultivar range in photoperiod sensitivity and attendant maturity. In the field, for example, genotypic differences in photoperiod sensitivity may cause one cultivar to flower just as another matures. The large pleiotropic effect of maturity on plant height, lodging, and seed yield tends to overshadow other phenotypic differences among genotypes. Thus, these traits are not readily used in the calculation of genetic distance. In addition to the clear pleiotropic effects of maturity on some plant traits, maturity also introduces a more subtle but equally important bias effect in the cultivar comparisons with respect to seed characters. A cultivar maturing more than 1 mo later than another in the same field in North America, for example, will likely experience a cooler temperature regime during pod development and, as a result, produce seed with altered size and composition (Martin et al., 1986). Thus, seed traits obtained from field studies are not well suited to studies of genetic distance when genetic differences in maturity are present. While there is not a clear solution to the problem of maturity and pleiotropism in distance estimation, Gizlice et al. (1993) demonstrated that the bias effect of maturity on seed traits is minimized when (i) phenotypic data are derived from cultivars grown in temperature-controlled growth chambers instead of field plots, and (ii) traits inherently related to maturity group, such as plant height at maturity, are deleted from analysis. Thus, growth chambers should provide an effective aid in the development of distance measures based on phenotypic traits.
The phytotron protocol employed in this study effectively minimized the maturity group effect on cultivar comparisons. Although the actual maturity groups for the 72 entries ranged from 000 to VIII (an approximate range of 9 wk in maturity in the field), the observed range in flowering and maturity dates was only 8 and 24 d, respectively. The range in flowering dates is similar to that reported for a previous phytotron-based study (Gizlice et al., 1993). Analysis of variance revealed significant (P < 0.01) phenotypic differences for all 25 traits studied. However, none was highly correlated with maturity group (all but one r value between ±0.20), indicating that constant temperature regime of the phytotron was effective in removing bias effects of maturity on seed trait comparisons. Thus, the data employed in this study appeared appropriate for assessment of phenotypic diversity in Chinese and NA cultivars.
NA vs. Chinese Cultivar Differences in Phenotypic Traits
Chinese and NA cultivar groups differed in phenotypic means for 13 of 25 traits (P < 0.05) (Table 2). On average, Chinese cultivars exhibited longer leaves, larger seed, higher protein and lower oil content in the seed, higher oleic and lower linoleic acid content in the seed oil, and lower leaf chlorophyll and lower nitrogen contents in comparison with NA cultivars (P < 0.05). Chinese cultivars were also more diverse than NA cultivars for 24 of the 25 traits on the bais of range of—and variance among—genotypic means (Table 2). The NA cultivars had a broader genotypic range than Chinese cultivars only for specific leaf weight.
|
Phenotypic Similarity Estimates
The large genotypic variation in phenotypic traits prompted the use of multivariate analysis to identify major genotypic patterns. Principal component analysis (PCA) was used to transform cultivar least squares means (Table 3) for phenotypic traits into principal component scores. The first seven principal components (those with eigenvalues greater than one) summarized 79% of the variation in cultivar means (data not shown). The first two principal components separated Chinese and NA cultivars well, indicating a clear phenotypic distinction between the two cultivar pools in terms of leaf morphology, chlorophyll content, and seed composition.
|
Cluster Analysis
Many clustering programs and algorithms are available for study of genetic diversity in crops. It is seldom clear a priori which will be best for any particular data set. To minimize the effect of clustering approach on results, we employed a number of methods. Cluster analysis was performed, initially, by the CLUSTER procedure of SAS (SAS Institute, 1985b). The PS matrix was subjected to Ward's minimum variance method and 11 other methods available in the CLUSTER procedure. Ward's minimum variance method explained the greatest proportion of the variation in the PS matrix and was the only method from the CLUSTER procedure retained for further use. Because a nonhierarchical clustering approach had been shown to be effective in the cluster analysis of pedigree data, we also applied a nonhierarchical analysis to the present data set and compared results with those from Ward's minimum variance method (Cui et al., 2000b). In this separate alternative approach to clustering, the PS matrix was converted into Euclidean coordinates (71 dimensions) via the MDS procedure (SAS Institute, 1992). The MDS-derived coordinates were used subsequently as source data in a series of FASTCLUS analyses, where the analysis was optimized sequentially for 2 to 25 clusters. The FASTCLUS analyses employing three and seven clusters accounted for only 15 and 39% of the variation in the original PS matrix, whereas Ward's minimum variance method accounted for 42 and 59% for the same number of clusters. Because the MDS plus FASTCLUS approach was less effective than Ward's minimum variance method, only the output from Ward's minimum variance method was retained for analysis and interpretation.
Ward's minimum variance method produced a dendrogram of cultivar relationships (Fig. 2) . The cubic clustering criterion and the pseudo F statistic had peaks at three clusters, indicating the existence of at least three major clusters (SAS Institute, 1985b; Milligan and Cooper, 1985). Examination of the dendrogram in relation to the origin and pedigree of cultivars revealed that each of the three major clusters could be decomposed further in a meaningful way to produce a total of seven clusters (Table 1 and Fig. 1). The decomposition of the three initial clusters into seven produced an increase in R2 from 0.42 to 0.59.
|
|
Cluster E was composed of 16 NA and 11 Chinese cultivars and exhibited intermediate values for most traits. Almost all Chinese cultivars in this cluster were from northern China and the NA cultivars in this cluster were derived largely from landraces from northern China. Although they appeared similar to each other phenotypically, no pedigree relation was found between these Chinese and NA cultivars. Cluster F contained five NA cultivars which exhibited thin leaves, large stem diameter, high oil, linoleic acid, and chlorophyll contents, and low seed protein content. Cluster G had only one member, the NA cultivar Braxton, which was unique phenotypically and represented the genetic extreme for several traits in this study (Table 4). The pedigree of Braxton was also unique in that 19% of its ancestry traced to the Japanese cultivar Tokyo. No other NA entry in the study had as high a percentage of Japanese germplasm in the pedigree.
Comparison of PS-Based and CP-Based Clusters
Because clear PS-based clusters were identified in this study, we sought to validate their integrity through a comparison with CP-based clusters described previously by Gizlice et al. (1996) (North America) and Cui et al. (2000b) (China) (Table 1). Comparisons were made in the following way. All cultivars in this study were assigned to a PS-based cluster. Many were also assigned to CP-based clusters in previous studies. Discrepancies between the former and latter assignments provided the basis for comparison of PS and CP results. A total of 20 Chinese and 16 NA cultivars were available to compare CP and PS cluster assignments (Table 1). Results showed a trend that members from a CP-based cluster were usually assigned to a single PS-based cluster. Thus, a positive relation existed between CP- and PS-based clusters. For example, the eight pairs of NA cultivars representing eight CP-based clusters were assigned to only two PS-based clusters. For six of the eight pairs, members of a CP based cluster were assigned to the same PS-based cluster. Groups of Chinese cultivars representing eight CP-based clusters were assigned to four PS-based clusters. For 12 of the 20 Chinese cultivars, all members of a CP-based cluster were also assigned to a single PS-based cluster. Where PS-based cluster assignments differed from CP-based assignments, the cultivars were often near each other on the graph of PS derived from MDS analysis (Fig. 1). This agreement tended to validate the PS-based clustering approach.
Relationship between PS and CP
China
Because PS- and CP-based clusters tended to show some agreement in characterization of cultivars within a country, we also examined the overall correlation of CP and PS estimates within a country. The mean CP relation for the 47 Chinese soybean cultivars was 0.03 and ranged from 0 to 0.63, with distribution of CP values strongly skewed toward values below 0.25 (Cui et al., 2000b). The mean PS relation for this group was higher (0.58) and ranged from 0.24 to 0.88. A weak but significant positive correlation (r = 0.14, P < 0.01) was found between the CP and corresponding PS values. The weakness of this linear association may have resulted partly from the preponderance of low CP values. Based on empirical results from Manjarrez-Sandoval et al. (1997), CP and PS would be expected to have little relation for CP values below 0.25 (the majority of CP values here), while the relation between genetic PS and CP should increase for higher CP values. When the cultivar pairs with low CP (<0.25) were dropped from analysis, the correlation between CP and corresponding PS rose from 0.14 to 0.43 (P < 0.01).
The correlation of PS with CP would also be expected to rise as the number of measured traits increases, because PS would then more completely represent the true underlying genetic pool effects. Had we employed a larger number of phenotypic traits in this study, the relation may have improved beyond the observed 0.43. Van Beuningen and Busch (1997) employed 35 traits and found a significant correlation of 0.68 between CP and PS.
North America
The mean and range in CP for 25 NA soybean cultivars was 0.145 (0.00–0.62) and the corresponding mean and range for PS was 0.69 (0.06–0.92). The correlation between CP and PS was low but significant and similar to that found for Chinese cultivars (r = 0.13, P = 0.02). The relation of PS and CP for the 25 NA soybean cultivars illustrated the same trend observed for Chinese cultivars. However, deletion of CP values below 0.25 did not improve the correlation between CP and PS, because few pairs had CP relationships above 0.25.
Five Chinese cultivars contained NA parentage in the pedigree (Table 1) and were a potential basis for comparing the impact of innate diversity (derived from founder effects) vs. breeder selection pressure on phenotype and PS estimates. If these five ‘bridge cultivars’ appeared more similar to Chinese than NA cultivars, phenotypically, then the result would suggest that contrasting breeder selection pressure on phenotype in China and North America may be important in the interpretation of our results. By contrast, if bridge cultivars appeared intermediate, phenotypically, between Chinese and U.S. cultivars, then this result would suggest that innate founder effects might be more important than breeder selection on PS estimates. Unfortunately, the parents (the appropriate controls for examining breeder selection effects) of the five bridge cultivars were not evaluated in this study; and, none of the other cultivars were closely related to the bridge cultivars. Thus, interpretation of the PS relationships between the bridge cultivars and Chinese or NA cultivars was unclear. However, it is interesting to note that none of the five bridge cultivars fell into clusters dominated by NA cultivars and that each was phenotypically more similar to Chinese than NA cultivars. Thus, there is a suggestion that breeder selection effects on PS may have been important.
Factors Affecting Genotypic Variation in PS
Maturity group, geographic origin, and seed yielding ability, as well as PS-based clustering of cultivars can all be important factors, which explain or classify phenotypic diversity in soybean breeding programs. We examined the association between these factors (Table 1) and PS by subjecting portions of the 72 by 72 PS matrix to single and multiple-factor regression analyses with factors, such as maturity group, treated as independent class variables (Table 5). For all factors except clustering results and maturity groups, Chinese and NA cultivars were examined in separate regression analyses. For Chinese cultivars, only PS-based clusters were associated with cultivar variation in the PS matrix. For NA cultivars, maturity group was associated with PS (R2 = 0.37) and mirrored a relation between CP and maturity group detected previously by Gizlice et al. (1996) (Table 5). This relationship between PS and maturity group was related to a strong NA breeder. However, multiple regression analysis indicated that maturity effects were correlated almost completely with cluster effects. In summary, the seven PS-based clusters explained 59% of the variation in PS, indicating that cluster analysis was clearly the best discriminator of phenotypic diversity in this study. Inclusion of multiple factors in the regression analysis did not improve the R2 appreciably over that obtained by cluster analysis alone.
|
Implications to Soybean Breeding
The purpose of this study was to examine the hypothesis that NA and Chinese cultivars shared a common, though unrecorded, underlying genetic base. Experiments revealed that phenotypic diversity was much greater in Chinese than in NA cultivars (Fig. 1, Fig. 2, and Table 2). The results suggest that, although most NA and Chinese cultivars trace their pedigree to Chinese landraces, NA cultivars may be derived from a subset of the Chinese cultivar genetic base. In suppporting research, Li et al. (2001) compared Chinese and NA soybean ancestors using RAPD markers and found that some ancestors of NA breeding programs were similar genetically to ancestors used in Chinese breeding. It is also possible that decades of breeder selection for adaptation to contrasting environmental conditions in China and North America may have accentuated the genotypic and, thus, the phenotypic diversity between Chinese and NA cultivars. The findings here support the concept that Chinese and NA cultivars are phenotypically distinct and, thus, potentially good genetic sources for broadening the agronomic, morphological, and biochemical diversity of the contrasting breeding programs.
Neither the trait-based PS analysis reported here nor previously reported CP analysis were perfect measures of diversity, because they both depended on assumptions that may or may not have been met completely. However, they are sufficiently useful such that breeders may consider the use of PS and CP as well as agronomic performance as guides in the selection of exotic parents for local breeding (Cox et al., 1985; van Beuningen and Busch, 1997). We speculate that specific hybrid combinations between the two genotypic pools may be chosen best by avoiding matings within phenotypic-based PS clusters and by selecting parents with high yield. Agronomic studies have shown that 10 of the 47 Chinese cultivars studied here are relatively high yielding in the North America (Table 1). Only two of the 10 were associated with phenotypic clusters dominated by NA cultivars. If high protein content is a priority for a breeder, it was noted that cultivars from central and south China tended to have high protein content in the seed, especially those in the PS-based Cluster D.
China has successfully used elite NA cultivars as breeding stock, demonstrating the effectiveness of this approach in broadening the genetic base for future breeding. Contrary to the conventional expectations regarding the difficulty in using exotic materials in applied breeding, most new Chinese cultivars derived from U.S. cultivars yielded well. For example, ‘Si Dou 11’ was selected from the cross of ‘Si Dou 2 Hao’ x ‘Williams’ and released in 1987 in Jiangsu province. It has ranked high in Chinese yield trials to the present. ‘Ji Dou 7 Hao’ was developed from the cross Williams x ‘Cheng Dou 1 Hao’ and released in 1992 in Hebei province. It established a yield record of 4749 kg ha-1 in provincial yield trials (Cui et al., 1998).
It is interesting to note that among the NA cultivars studied here, both maturity group and PS-based cluster were useful in explaining phenotypic diversity. To maximize morphological diversity (and presumably agronomic diversity) in the progeny from NA x NA cultivars examined here, one might want to choose parents that differ in maturity, as well as PS and CP based cluster designation.
|
ACKNOWLEDGMENTS |
|---|
Received for publication December 21, 2000.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Vadose Zone Journal | |||
| Journal of Natural Resources and Life Sciences Education |
Soil Science Society of America Journal | ||||
| Journal of Plant Registrations | Journal of Environmental Quality |
The Plant Genome | |||
