| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Dep. of Agronomy, Iowa State Univ., Ames, IA 50011
* Corresponding author (wfehr{at}iastate.edu)
 |
ABSTRACT |
|---|
 TOPNOTES ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
80 mg seed-1 are needed to produce grain for export to Japan for the production of natto, a fermented food. The purpose of this study was to compare three population types for the recovery of lines that would have adequately small seed size for natto. Two small-seeded and one normal-size conventional cultivar or line were used to produce a small-seeded x small-seeded two-parent population, a small-seeded x normal-size two-parent population, and a small-seeded x (small-seeded x normal-size) three-parent population. Five sets of the three population types were developed with different parents in each set. The seed size of 100 random F2 plants was determined from each of the 15 populations and 10 plants of each parent. For each of the five sets, the progeny of the 100 F2 plants of each population type were compared as F2:3 lines with 10 entries of each of the three parents at two Iowa locations. The average percentage of lines with a seed size equal to or smaller than one of the parents in a cross was 90% for the small-seeded x small-seeded populations, 4% for the small-seeded x normal-size populations, and 20% for the three-parent populations. An average of 10% of the lines from the small-seeded x small-seeded populations had significantly smaller seed size than either of the parents, and no transgressive segregation for small seed size was observed in the other two population types. For the development of small-seeded cultivars, small-seeded x small-seeded and three-parent populations would provide lines with acceptable seed size. A small-seeded x normal-size population may provide suitable lines if seed size of the conventional cultivar is sufficiently small and adequate resources are available to select for the limited number of small-seeded segregates. |
INTRODUCTION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
80 mg seed-1 are desirable for producing grain for use by Japanese natto manufacturers. By comparison, conventional cultivars grown in the USA range in seed size from 120 to 180 mg seed-1 (Hartwig, 1973).
Small-seeded cultivars available in the northern United States currently yield 
20% less than conventional cultivars. It would be desirable in breeding for improved small-seeded cultivars to capitalize on the high yield potential of conventional cultivars. No research has been reported on the segregation for seed size that would be expected from single crosses between a small-seeded parent with 80 mg seed-1 and a conventional cultivar or line with normal size. The most closely related studies involved single crosses between accessions of Glycine soja Sieb. & Zucc. that have 20 mg seed-1 and conventional cultivars of G. max (Weber, 1950; Buhr, 1976; Carpenter and Fehr, 1986). In all of the studies, none of the lines derived from the populations had a seed size as small as the G. soja parent.
It may be possible to take advantage of the high yield potential of conventional cultivars to improve the yield of small-seeded cultivars by the development of three-parent populations. A small-seeded parent would be crossed to a normal-size parent with high yield potential, and the F1 plants from the single cross would be mated to a second small-seeded parent to increase the frequency of alleles for small seededness in the population. Segregates from the population would have an average of 25% of their alleles from the high-yielding normal-size parent and 75% from the small-seeded parents. The utility of three-parent populations for developing small-seeded cultivars is not known because no studies have been reported in which the segregation for seed size was evaluated for three-parent populations involving parents with 80 mg seed-1.
The objectives of this study were (i) to evaluate the segregation for seed size among three population types involving parents with 80 mg seed-1 and normal-size parents and (ii) to evaluate the feasibility of selection for small seed size among single plants.
|
MATERIALS AND METHODS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
|
The F1 seeds of populations 12 and 13 were planted in October 1996 at Isabela. Each F1 plant was harvested separately to form a subpopulation. Seeds from the F1 plants of the three populations were planted as subpopulations at the Agronomy Farm in 1997, and seeds of the three parents were planted in adjacent plots. The seeds were planted 10 cm apart in rows 69 cm wide. Either molecular marker analysis or pubescence color was used to confirm that subpopulations of populations 11 and 12 originated from hybrid plants, and the identity of the subpopulations was not maintained thereafter. The identity of the subpopulations of population 13 was maintained at harvest. The F2 and parent plants were harvested individually. Seed size of 100 random F2 plants from each population type and 10 random plants of each parent was measured by dividing the weight of all the seeds from a plant by the number of seeds. For evaluation of their F3 progeny, 72 seeds from each F2 plant were desired to plant in two replications at each of three locations. That amount of seed was not available from each of 100 plants. To determine if using plants with less than 72 seeds might bias the results, a phenotypic correlation between seed size and seed number was calculated for each population. The correlation coefficients of 0.04 for population 11, 0.00 for population 12, and 0.10 for population 13 were not significant (P > 0.05). Plants with as few as 66 seeds were used in Set 1. The number of lines tested per subpopulation of the three-parent population was dependent on the number of F2 plants with at least 66 seeds and ranged from 2 to 13.
Eight small-seeded and four normal-size cultivars or lines were used to form the populations for Sets 2 to 5 (Table 1). Two-parent populations were made during March 1998 at Isabela. The F1 seeds from the populations were planted in May 1998 at the Agronomy Farm near Ames. Molecular marker analysis or pubescence color was used to confirm that the plants originated from hybrid seed, and subpopulations of the two-parent populations were not maintained thereafter. The F1 plants of the small-seeded x normal-size populations were mated to a small-seeded parent to form three-parent populations. The F1 seeds of the three-parent populations were planted in October 1998 at Isabela. Molecular marker analysis or pubescence color was used to confirm that the plants originated from hybrid seed. Each F1 plant was harvested individually, and subpopulations of the three-parent populations were maintained. The F2 seeds and seeds of the parents of all the populations of Sets 2 to 5 were planted in January 1999 in Isabela. The soil type was a Coto clay (Very-fine, koalinitic, isohyperthermic, Typic Haplorthox). The seeds were planted 15 cm apart in rows 102 cm wide, and the plants were grown under artificial lights to extend the day length for increased seed production. The F2 and parent plants were harvested individually. Seed size of 10 random parent plants and 100 random F2 plants from each of the populations was measured by dividing the weight of all the seeds by the number of seeds. The 100 F2 plants of the three-parent populations included four progeny from each of 25 F1 plants.
The F2:3 lines and parents of Set 1 were grown in 1998, and those of Sets 2 to 5 were grown in 1999 at the Agronomy Farm and the Burkey Farm, which are part of the Agricultural Engineering and Agronomy Research Center near Ames. The soil type at both locations is a Nicollet loam (Fine-loamy, mixed, superactive, mesic Aquic Hapludoll). Set 1 also was grown in Puerto Rico, but the data were not included in the analysis because an insufficient number of seeds were obtained from many of the plots. The 330 lines of each set included 100 F2:3 lines from each of the populations and 10 entries of each of the three parents. Each set was planted as a separate experiment. The plots were single rows 76 cm long with 102 cm between rows and a 107-cm alley between the ends of plots. A maximum of 12 seeds and a minimum of 11 seeds were planted in each plot at the locations. Each plot was harvested in bulk with a self-propelled combine. Seed size was measured by weighing a 400-seed sample from each plot.
The data from all populations were analyzed using the general linear model (GLM) procedure of the SAS software (SAS Institute, 1992). Genotypes, locations, and replications were considered random effects. Parent entries were excluded from the analyses of variance.
For each cross, broad-sense heritability estimates and their standard errors were computed on a plot and entry-mean basis from variance component estimates for the F2:3 lines combined across the two locations (Hallauer and Miranda, 1995). Estimates of heritability were calculated in standard units by the phenotypic correlation of the seed size of the F2 plants with the mean of their F2:3 lines over locations (Frey and Horner, 1957). Heritability estimates also were calculated by regressing the mean of the F2:3 lines across locations on the seed size of the F2 plants. Phenotypic correlation coefficients were calculated using the correlation (CORR) procedure, and regressions were calculated using the regression (REG) procedure of the SAS software (SAS Institute, 1992).
|
RESULTS AND DISCUSSION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
There were significant differences in seed size among the F2:3 lines within each of the 15 populations. Because of environmental influence on seed size, segregation within the populations was evaluated by comparing the seed size of F2 plants and F2:3 lines with that of their parents, instead of using a fixed seed size of 80 mg sd-1. The small-seeded x small-seeded populations had the highest percentage of lines of the three population types with seed size equal to or smaller than the larger of the small-seeded parents in the set (Table 1). There was a range of 81 to 100% of the lines in the five populations equal to or smaller in size than the larger of the small-seeded parents in the set, with an average for the five populations of 90%. Transgressive segregation for seed size less than the smaller parent ranged from 4 to 21% and averaged 10% for the five populations. The percentage of F2:3 lines larger than the larger parent ranged from 0 to 19% and averaged 10%. The transgressive segregation for seed size observed in the crosses between small-seeded parents agreed with the transgressive segregation reported by Weber and Moorthy (1952) for crosses between cultivars with normal seed size.
All the F2:3 lines from the small-seeded x normal-size populations had larger seed size than the small-seeded parent in the cross, except for one line in Population 51 (Table 1). The percentage of lines in the five populations with a seed size equal to the larger of the small-seeded parents in the set ranged from 0 to 13%, with an average of 4%. In previous evaluations of segregation from small-seeded x normal-size crosses, the small-seeded parent was an accession of G. soja with a seed size of 20 mg seed-1, and the normal-size parent was a G. max cultivar (Weber, 1950; Buhr, 1976; Carpenter and Fehr, 1986). None of the lines recovered in these three studies were equal in size to the G. soja parent in a cross. Although the small-seeded parents used in our study were about three times larger than those of the G. soja parents used in the previous research, only 1 of 500 F2:3 lines was equal to the small-seeded parent in a cross. The results indicated that the recovery of lines with seed size equal to the small-seeded parent of a small-seeded x normal-size cross would be difficult due to the allelic differences in the parents for the multiple genes controlling seed size.
For the three-parent populations, the number of lines equal to or smaller than the larger small-seeded parent in the set ranged from 7 to 53%, with an average for the five populations of 20% (Table 1). The variation in the percentage of acceptable lines among the five three-parent populations indicated that it would be difficult to reliably determine the population size needed to recover a desired number of small-seeded lines that could be evaluated for yield and other traits in a cultivar development program.
The effectiveness of single-plant selection for small seed size was evaluated by computing heritability estimates based on regression and phenotypic correlation of the seed size of the F2 plants and the mean size of their F2:3 lines. The estimates from the regression analysis ranged from 0.15 to 0.70 among the populations (Table 2). The heritability estimates obtained by phenotypic correlation ranged from 0.30 to 0.87. Carpenter and Fehr (1986) reported heritability estimates of 0.72 and 0.81 based on the phenotypic correlation between the seed size of F2 plants and their F2-derived lines from two crosses of small-seeded G. soja parents with normal-size G. max parents grown in two replications at two locations. LeRoy et al. (1991) obtained heritability estimates on a single-plant basis averaging 0.35 that were computed from variance component estimates from F2-derived lines of three G. max x G. soja crosses grown in two replications at four locations. For nine crosses involving normal-size and large-seeded parents, Bravo et al. (1980) reported an average heritability of 0.27 on a single-plant basis computed from variance component estimates obtained from F2-derived lines grown in two replications at two locations.
|
+ 2 standard deviations (SD) of the larger small-seeded parent in the population. No error in acceptance of the F2 plant occurred when the line derived from that plant had a mean seed size in the replicated test equal to or smaller than the + 2 standard errors of the mean (SEM) of the larger small-seeded parent. An acceptance error occurred when the line derived from an acceptable F2 plant had a mean seed size in the replicated test that exceeded the + 2 SEM of the larger small-seeded parent. An F2 plant was rejected if its mean seed size exceeded the + 2 SD of the larger small-seeded parent in the set. No error in rejection of the F2 plant occured when the line derived from that plant had a mean seed size in the replicated test that exceeded the + 2 SEM of the larger small-seeded parent. A rejection error occurred when the line derived from a rejected F2 plant had a mean seed size in the replicated test equal to or smaller than + 2 SEM of the larger small-seeded parent. For the small-seeded x normal-size populations, the percentage of F2 plants that were incorrectly accepted ranged from 0 to 9% and none of the plants were incorrectly rejected (Table 3). The low frequency of misclassification was due to the absence of segregates with seed size equal to or smaller than the small-seeded parent of the population. Selection for seed size among F2 plants or the evaluation of random lines from small-seeded x normal-size populations comparable with those used in this study would not be warranted because such a low frequency of segregates would have seed as small as the small-seeded parent.
|
For the three-parent populations, the sum of the percentages of F2 plants incorrectly accepted and rejected ranged from 40 to 67%, which was at least twice as high as for the other two population types. The percentage of rejection error was greater than acceptance error in population 13 of Set 1, while the reverse was observed for the three-parent populations of Sets 2 to 5 (Table 3). The F2 plants and F2:3 lines of Set 1 were evaluated in different environments than the other sets, which was considered a major factor in the differences observed in acceptance and rejection errors.
The sum of the acceptance and rejection errors in the three-parent populations was so high that it brought into question the merit of investing the time and funds necessary to evaluate and select among F2 plants for seed size. The alternative would be to evaluate as lines the progeny of random F2 plants from a population. To compare the two alternatives, the percentage of lines that would have acceptable seed size was determined for each of the five three-parent populations. If selection was practiced among F2 plants, the percentage of acceptable lines based on the numbers in Table 3 would be determined by dividing the number of F2 plants accepted by the number of F2 plants correctly accepted and multiplying by 100. The percentage of acceptable lines with selection would be 93% for population 13, 24% for population 23, 12% for population 33, 14% for population 43, and 11% for population 53. If random lines were used from the populations, the percentage of acceptable lines would be equal to the sum of the percentage of F2 plants selected without error and the percentage of plants incorrectly rejected. The percentage of acceptable lines would be 53% for population 13, 21% for population 23, 7% for population 33, 10% for population 43, and 8% for population 53. In all of the populations, the percentage of acceptable lines was greater when selection was practiced among the F2 plants, although the difference was small in the three-parent populations of Sets 2 to 5. The F2 plants for those sets were grown in the Puerto Rico nursery, while the F2 plants of Set 1 were grown in Ames. The breeder would have to decide if the greater time and expense associated with selection among F2 plants would be justified in comparison with the evaluation of random lines from the three-parent populations.
Broad-sense heritability estimates for seed size on a plot basis ranged from 0.41 to 0.83 and averaged 0.69 for the 15 populations (Table 2). The heritability estimates on an entry-mean basis ranged from 0.74 to 0.94 and averaged 0.88. Weber and Moorthy (1952) reported an average broad-sense heritability estimate for seed size on a plot basis of 0.54 for three crosses between normal-size cultivars. Bravo et al. (1980) obtained average broad-sense heritability estimates for seed size of six crosses involving normal-size and large-seeded parents of 0.41 on a plot basis and 0.71 on an entry-mean basis. The average broad-sense heritability estimates for seed size obtained by LeRoy et al. (1991) from evaluation of F2-derived lines from three G. max x G. soja crosses were 0.52 on a plot basis and 0.89 on an entry-mean basis.
For the development of small-seeded cultivars, the most suitable population types in our study were the small-seeded x small-seeded two-parent crosses and the small-seeded x (small-seeded x normal size) three-parent crosses. The small-seeded x normal-size two-parent cross may be effective if the small-seeded parent had smaller size than that required of the segregates, the normal-size parent had smaller seed than that of common cultivars grown in the northern United States, or both. The seed size of potential small-seeded and normal-size parents also should be considered for three-parent populations as a means of increasing the frequency of segregates with acceptable size.
|
NOTES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
Received for publication October 1, 2000.
|
REFERENCES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  E. R. Cober and H. D. Voldeng Mass Selection for Small Seed Size in Natto Soybean Populations and the Resulting Effect on Seed Yield Crop Sci., July 1, 2008; 48(4): 1337 - 1340. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. A. Hoeck, W. R. Fehr, R. C. Shoemaker, G. A. Welke, S. L. Johnson, and S. R. Cianzio Molecular Marker Analysis of Seed Size in Soybean Crop Sci., January 1, 2003; 43(1): 68 - 74. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. J. Alt, W. R. Fehr, and G. A. Welke Selection for Large Seed and High Protein in Two- and Three-Parent Soybean Populations Crop Sci., November 1, 2002; 42(6): 1876 - 1881. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 | |||
