| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Monsanto, 1203A Airport Road, Ames, IA 50010
b Pioneer Hi-bred International, Johnston, IA 50131
c Univ. of Minnesota, Dep. of Agronomy and Plant Genetics and Plant Molecular Genetics Institute, 411 Borlaug Hall, 1991 Upper Buford Circle, St. Paul, MN 55108
* Corresponding author (phill005{at}tc.umn.edu).
 |
ABSTRACT |
|---|
 TOPNOTES ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONSUMMARYREFERENCES |
|---|
Abbreviations: AK, aspartate kinase • CP, crude protein • DM, dry matter • kDa, kilodalton • MET, methionine • NIRS, near infrared reflectance spectrophotometry • QPM, quality protein maize
|
INTRODUCTION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONSUMMARYREFERENCES |
|---|
In addition to productivity gains realized by supplementing poultry rations with MET, Schutte (1989) has shown that optimizing the limiting amino acid (MET) content increases the efficiency of dietary protein utilization. By supplementing MET and cysteine to 100% of normal requirements, up to 15% of total feed ration protein could be eliminated with no decrease in weight gain of young turkeys (Waibel et al., 1995). Increasing the N use efficiency of poultry or reducing the total amount of protein fed could potentially decrease N excretion, resulting in diminished N pollution (Schutte, 1989; van Weerden, 1989).
Conventional breeding efforts to improve the protein quality of maize have focused primarily on increasing lysine and tryptophan. The opaque-2 mutation in maize confers increased kernel lysine and tryptophan and has been used in breeding programs to improve the protein quality of maize. Generally, opaque-2 maize yields less and is more susceptible to disease and insect pests than normal maize (Brown, 1975). Maize breeders at CIMMYT (the International Maize and Wheat Improvement Center) have produced high-lysine modified opaque-2 cultivars, designated quality protein maize (QPM), with normal kernel characteristics (National Research Council, 1988; Bjarnason and Vasal, 1992). Pixley and Bjarnason (2002) evaluated 62 QPM cultivars in a diverse set of 13 locations and found that protein concentration, tryptophan content of grain, and tryptophan content of protein behave as relatively stable traits. Zuber and Helm (1975) conducted recurrent mass selection for whole-kernel lysine within three normal dent maize populations and were able to achieve 31 to 48% increases in lysine levels.
The maize inbred line BSSS53 has relatively high levels of kernel MET (Phillips et al., 1981). BSSS53 has elevated levels of a high-MET 10-kilodalton (kDa) zein protein (Phillips and McClure, 1985). The structural gene for the 10-kDa zein has been cloned and found to contain 23% MET (Kirihara et al., 1988). This gene, designated dzs10, is regulated posttranscriptionally by a trans-acting regulatory gene, dzr1 (Benner et al., 1989; Cruz-Alvarez et al., 1991; Chaudhuri and Messing, 1995). The dzr1+BSSS53 allele is recessive to the dzr1+Mo17 allele; however, the dzr1+Mo17 allele is parentally imprinted and silenced when transmitted through the male (Chaudhuri and Messing, 1994). Mo17 accumulates low levels of 10-kDa zein (Schickler et al., 1993; Chaudhuri and Messing, 1994) and is low in whole-kernel MET. Although the 10-kDa zein is high in MET, attempts to increase the whole kernel MET level of Mo17 through backcrossing with selection for high 10-kDa expression have failed to produce lines with increased MET (Krone, 1994).
Transgenic approaches have been attempted to increase the MET level of maize. The high-MET 10-kDa zein was fused with a 27-kDa zein promoter in an attempt to effect seed-specific overexpression of this protein (Kleese et al., 1991). The 10-kDa zein gene under the control of endogenous promoters has been used to increase the level of MET in 19 separate transformants (Anthony et al., 1997). In that study, MET levels were increased by up to 30%. Lai and Messing (2002) constructed a chimeric dzs10 gene that preserves the coding sequence of the gene but eliminates the putative cis-acting target sequences of dzr1. Transformed lines developed with this construct stably express the 10-kDa zein protein and MET levels are increased by 80% (Lai and Messing, 2002).
Both whole-kernel and free MET levels can also be influenced by modification of the aspartate pathway, the biosynthetic pathway leading to MET, lysine, isoleucine, and threonine. Aspartate kinase (AK) is the first enzyme of the aspartate pathway leading to the biosynthesis of lysine, MET, threonine, and isoleucine (Azevedo et al., 1997). Ask1-LT19 is a mutant allele of Ask1 that increases whole-kernel threonine, MET, and lysine (Muehlbauer et al., 1994). Wang and Larkins (2001) found that the effect of the o2 mutation in Oh545 on the free amino acid levels of aspartate-derived amino acids was much greater than the effect of the same mutation in the W64A background. Free MET is 14.4 times greater in Oh545o2 than in W64Ao2, and one of the largest quantitative trait loci identified for free amino acid accumulation in a population derived from these two o2 lines corresponds to the lysine-sensitive AK 2 gene (Wang and Larkins, 2001; Wang et al., 2001).
A native landrace, IAPO-13, was found to have very high MET levels ranging from two to four times the MET level of normal dent maize (Zarkadas, 1997). This landrace should be an excellent source of increased MET that may be used as a donor in traditional breeding programs. The authors are unaware of any reported efforts to increase whole kernel MET through conventional plant breeding approaches. For this reason, there is no information currently available regarding strategies or expectations of conventional selection for increased kernel MET in maize.
The inbred BSSS53 has been identified as a potential source of increased MET and is 20 to 100% higher in MET than other corn-belt dent inbreds (Phillips et al., 1981; Krone, 1994). Krone (1994) used BSSS53 as a donor parent to generate BC2S2:3 versions of A632, B73, and Mo17 that were significantly higher in MET than their respective recurrent parents. One objective of the present research was to further backcross these high-MET BC2S2:3 lines while maintaining increased MET levels. High and low MET backcross-derived lines generated from this effort have been genotyped as part of a larger effort to identify and confirm genomic locations of quantitative trait loci controlling MET accumulation. High-MET materials resulting from this program also represent potentially useful intermediary donor sources for future MET breeding efforts since each of the inbred lines improved for MET in this study are important progenitors of current elite corn-belt inbred lines. The ultimate objectives of this effort were to increase the MET level of the three maize hybrids A632/Mo17, B73/A632, and B73/Mo17, and to evaluate the efficacy of per se selection for improving hybrid MET levels.
|
MATERIALS AND METHODS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONSUMMARYREFERENCES |
|---|
The nomenclature used to designate experimental lines in this study describes the genetic background, the specific BC2S2:3 source line, and the backcross generation. For example, A632MET-79 refers to a specific high-MET A632 BC2S2:3 line. This line was designated Treatment 79 in the BC2S2:3 inbred amino acid trial of Krone (1994). A632MET-79-3-53 is a BC3S2:3 line derived from A632MET-79, and was designated Treatment 53 in the A632 trial of the current study. Similarly, A632MET-79-4-42 is a BC4S1:2 line derived from A632MET-79, designated Treatment 42 in the current study.
Selection among BC3S2:3 and BC4S1:2 lines within each genetic background was made based on MET levels predicted with near infrared reflectance spectrophotometry (NIRS). After 2 yr of evaluation for whole-kernel MET content, selected high-MET lines were crossed to evaluate MET levels of resulting hybrids. Selected high-MET lines from each genetic background were crossed with selected high-MET lines from each of the other two genetic backgrounds. Sixteen experimental hybrids were made by intercrossing the seven BC2S2:3 lines. One hundred and one experimental hybrids were made by intercrossing the highest MET BC3S2:3 and BC4S1:2 lines. Seed of each experimental hybrid was produced by crossing eight plants of the male parent line with eight plants of the female parent line. B73 lines were used as females in crosses with A632 and Mo17 lines. A632 lines were used as females in crosses with Mo17 lines.
Experimental Trials
All of the experimental trials reported were grown on Waukegan silt loam soil (fine-silty over sandy or sandy-skeletal, mixed, superactive, mesic Typic Hapludolls) at the Minnesota Agricultural Experiment Station at St. Paul, MN. Separate inbred line trials containing A632-, B73-, or Mo17-derived lines were conducted in 1996, 1997, and 1998. Hybrid trials were conducted in 1998. In each growing season, urea was applied to provide 112 kg ha-1 N. All inbred and hybrid trials were planted as single-row plots, 5.49 m long, spaced 91 cm apart. Thirty-six kernels were planted in each plot. All trials were overplanted and thinned at the V4 stage. Inbred trials were thinned to 26 plants per plot (
52 000 plants ha-1) and hybrid trials were thinned to 28 plants per plot (56 000 plants ha-1). Within each plot, six to eight plants were self-pollinated. Equal volumes of seed from each ear were bulked to form a representative sample. For MET concentration analyses, 23.5 g of seed from each line was ground to pass a 1-mm screen, dried at 65°C for 24 h (60 g kg-1 moisture), and tumbled for 45 min to homogenize ground samples.
During the 1996 summer growing season, three separate inbred trials were conducted. In the A632 and B73 genetic backgrounds, 68 experimental lines were grown together with the two high-MET BC2S2:3 source parents, the recurrent parent, and BSSS53. In each of these trials, the 72 entries were planted in an 8-by-9 rectangular lattice with three replications. In the Mo17 genetic background, 127 experimental lines were grown together with the three high-MET BC2S2:3 source parents, Mo17, and BSSS53. The 132 entries in this trial were planted in an 11-by-12 rectangular lattice with three replications. The three replications of the Mo17 trial were delay-planted at 4-d intervals to facilitate pollinations.
High-MET selections from the 1996 trials were reevaluated during the summer of 1997. Within each genetic background, a number of selected lines were compared with the BC2S2:3 source parents, the recurrent parent, and BSSS53. There were 24 entries in each of the A632 and B73 experiments, and 30 entries in the Mo17 experiment. Lines advanced from 1996 and 1997 were evaluated in inbred trials during the summer of 1998. Inbred trials in 1997 and 1998 were planted as randomized complete block design experiments with three replications.
Each hybrid trial (B73/A632, B73/Mo17, A632/Mo17) consisted of 42 entries, planted as a 6-by-7 rectangular lattice with three replications. The B73/A632 trial included 35 hybrids derived from BC3S2 and BC4S1 lines, four hybrids derived from crosses among BC2S2:4 lines, B73/A632, B73/BSSS53, and A632/BSSS53. The B73/Mo17 trial included 33 hybrids derived from BC3S2 and BC4S1 lines, six hybrids derived from crosses among BC2S2:4 lines, B73/Mo17, B73/BSSS53, and Mo17/BSSS53. The B73/Mo17 trial included 33 hybrids derived from BC3S2 and BC4S1 lines, six hybrids derived from crosses among BC2S2:4 lines, A632/Mo17, A632/BSSS53, and Mo17/BSSS53. No specific mating design was employed to determine which crosses were made. The highest MET BC3S2 and BC4S1 lines within each genetic background were crossed with the highest MET BC3S2 and BC4S1 lines from each of the other two genetic backgrounds. The focus of this strategy mimics conventional breeding approaches in that the aim was to identify the highest MET hybrids within the group of hybrids evaluated. In general, BC3S2 lines were crossed with BC3S2 lines and BC4S1 lines were crossed with BC4S1 lines. This structure was maintained to compare the best BC2, BC3, and BC4–derived hybrids across backcross generations. However, in some cases, crosses between lines of different filial generation were made to test hybrid combinations of lines with the most promising per se data.
Estimation of Methionine and Crude Protein Levels
Near infrared reflectance spectrophotometry was used to estimate crude protein (CP) and MET levels of all samples from experimental inbred and hybrid trials. A Foss North America model 6500 Near Infrared Reflectance Spectrophotometer was used to estimate CP and MET levels. The NIRS equation for estimation of MET levels was developed in 1996 and expanded in subsequent years (Olsen, 1999). The final equation, developed from samples collected during the 3-yr period, was used to repredict MET levels of all samples. MET levels of all experimental units were expressed on a dry matter (DM) basis (g MET kg –1 DM) and as a percentage of CP (g MET 100 g-1 CP).
Crude protein levels were estimated for all lines with an NIRS commercial corn grain equation (Infrasoft International, 1998). The commercial corn grain equation was monitored by measuring micro-Kjeldahl CP levels of 27 experimental lines from 1996 and 1997. Samples were analyzed in duplicate and a conversion factor of 5.70 was used to determine CP. The 5.70 conversion factor was used instead of the commonly used 6.25 conversion factor to correct for nonprotein N (Zarkadas, 1997). Micro-Kjeldahl determined CP levels were regressed on NIRS-determined CP levels to check the calibration of the commercial corn grain equation.
Statistical Analysis
Combined analyses of inbred trials were conducted across years for lines grown in all three years within each genetic background; there were 14 lines for the A632 genetic background, 17 lines for the B73 background, and 19 lines for the Mo17 background. Each year was analyzed as a randomized complete block experiment within each genotype. Randomized complete block experiments were analyzed with PROC GLM procedure of SAS (SAS Institute, 1994). In the combined analysis, years and replications within years were considered as random effects, and genotypes were considered fixed. The genotype x year interaction mean square was used to test significance of differences among genotypes. Homogeneity of error variances across years was tested with the Bartlett test (Steel and Torrie, 1980, p. 471–472). Although heterogeneity of error was detected in some cases, the results of the combined analysis are reported. Hybrid experiments were analyzed by PROC LATTICE in SAS (SAS Institute, 1994). Adjusted treatment means are reported for experiments for which incomplete blocks had relative efficiency of 
105%. The average value of the variance of means for genotypes occurring together in the same incomplete block and the variance of means for genotypes occurring in different blocks was used for testing differences among treatment means. Mean separation for all trials was performed with the LSD test criterion. For MET expressed on either a DM or percentage protein basis, a one-sided test was conducted to determine if experimental genotypes were significantly higher in MET than the control inbreds or hybrids. For protein level, a two-sided test was conducted to determine if experimental genotypes were significantly different from the control inbreds or hybrids.
Hybrid MET levels were regressed on their midparent MET values as a preliminary evaluation of the efficacy of backcross selection without testcrossing. Heritability estimates were not inferred as the parental lines were highly selected for MET content.
|
RESULTS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONSUMMARYREFERENCES |
|---|
Selection of High-Methionine Backcross-Derived Inbred Lines
Within the A632 genetic background, individual BC3S2:3 and BC4S1:2 lines were identified with significantly higher MET levels than A632 on a DM basis (Table 1). One BC2S2:4 (A632MET-80) line and one BC3S2:3 (A632MET-80-3-1) line exhibited significantly increased levels of MET as a percentage of total protein. None of the BC4S1:2 lines were significantly higher in MET as a percentage of total protein when compared with A632. A632MET-80-3-1 was significantly lower in total protein than A632 while A632MET-79-3-53 was higher in total protein than A632.
|
In the Mo17 genetic background, BC2S2:4, BC3S2:3, and BC4S1:2 lines with significantly higher levels of MET were identified (Table 1). Selected lines exhibited significant increases in MET on a DM basis and as a percentage of total protein. Selected BC3S2:3 lines were from 38 to 55% higher in MET than Mo17 on a DM basis. Selected BC4S1:2 lines were from 22 to 31% higher in MET than Mo17. High-MET lines in the Mo17 genetic background were also consistently higher than Mo17 in total protein. The MET levels of the highest BC2, BC3, and BC4 lines within each genetic background are compared with their respective recurrent parents and the donor line, BSSS53, in Fig. 1. A summary of the ANOVA for each inbred trial is provided in Table 2.
|
|
|
Several backcross-derived versions of A632/Mo17 were significantly higher in MET than A632/Mo17 (Table 3). Six A632 BC3/Mo17 BC3 experimental hybrids were significantly higher than A632/Mo17 in MET on both a DM and a percentage protein basis. Of these hybrids, only two had significantly higher grain protein concentration than A632/Mo17. A summary of the MET levels of the highest BC2/BC2, BC3/BC3 and BC4/BC4 hybrids compared with the three check hybrids is provided in Fig. 2. A summary of the ANOVA for each hybrid trial is provided in Table 4.
|
|
|
|
DISCUSSION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONSUMMARYREFERENCES |
|---|
The percentage of the increase in MET due to increased total protein can be estimated by predicting the expected increase in MET, assuming no overall change in the amino acid profile of the kernel protein. Under this assumption, a simple increase in kernel protein would result in a proportionate increase of all amino acids. With this approach, 49% of the increase observed in B73MET-127 and its derivatives was accounted for by increased protein levels. On average, the percentage of the MET increase in high-MET Mo17 derivatives accounted for by increased kernel protein was 22%. These results indicate that both increased protein levels and increased MET as a percentage of total protein contributed to higher MET in these lines.
Selection for increased MET levels was conducted on a DM basis and no attempt was made to account for increased protein levels. Evaluation of the high-MET inbreds and hybrids will be necessary to determine if the increased protein content of these materials will adversely affect yield or other agronomic characteristics. Depending on the influence of high kernel protein on other traits of importance, it may be necessary to place greater emphasis on breeding for increased MET as a percentage of total protein. The fact that BSSS53 has higher MET as a percentage of total protein than A632, B73, and Mo17 and is not significantly higher in total protein than these inbreds suggests that the high-MET level of BSSS53 results from an altered amino acid profile. Since the donor inbred, BSSS53, has an altered amino acid profile, it was possible to develop genotypes from this source having increased MET and unchanged protein levels.
Among the A632 high-MET selections, A632MET-80 and A632MET-80-3-1 showed significant increases in MET as a percentage of total protein. A632MET-80-3-1 had significantly lower total protein than A632. Interestingly, A632MET-80-3-1 produced high-MET hybrids in both the B73/A632 trial and the A632/Mo17 trial even though this line was not higher in MET on a DM basis compared with A632. This line has a relatively low protein level and an amino acid profile very similar to BSSS53. It may be that this line expresses relatively high-MET proteins, and that, in combination with high protein levels from lines like Mo17MET-158-3-113 or Mo17MET-158-3-84, it produces hybrids with increased MET levels.
No backcross-derived lines were selected which had MET levels as high as BSSS53, suggesting either that some beneficial alleles have been lost during backcrossing or that epistatic effects of genes within these genetic backgrounds have reduced the magnitude of MET accumulation by introgressed BSSS53 alleles.
In each of the three genetic backgrounds, no BC4S1:2 lines were recovered with MET levels as high as the best BC3S2:3 genotypes. This could be caused by a loss of favorable alleles following the fourth backcross or it may be that some of the MET loci in the BC4S1:2 lines have not been fixed homozygous for the favorable alleles. If heterozygosity of MET loci exists within the high-MET BC4S1:2 lines, it should be possible to select segregants derived from these lines with further increases in MET. If, for example, dzr1 or one of the AK genes were heterozygous in a BC4S1 line, it should be possible to select homozygous progenies of this BC4S1 line that would have higher MET than the BC4S1 line.
Among the B73/Mo17 experimental hybrids with increased MET, a consistent increase in CP as a percentage of dry weight was observed. On average, 46.2% of the MET increase in these lines can be attributed to increased total protein as a percentage of dry weight. Consistent with results from the inbred trials, the BC4/BC4 hybrids were lower in MET than the BC3/BC3 hybrids. The BC4/BC4 hybrids also tended to have lower protein as a percentage of dry weight than the BC3/BC3 hybrids. Although MET levels of B73/Mo17 have been increased, the correlated increase in CP as a percentage of dry weight may have unanticipated effects both on the agronomic performance of these hybrids and on the feeding value of the grain.
A632/Mo17 experimental hybrids with increased MET generally had smaller increases in protein as a percentage of dry weight. Only one BC3/BC3 hybrid was significantly higher in protein as a percentage of dry weight than A632/Mo17, and A632MET-80-3-1/Mo17MET-158-3-84 was nonsignificantly lower in protein as a percentage of dry weight. In the A632/Mo17 trial, the BC3/BC3 hybrids tended to be higher in MET than the BC2/BC2 or BC4/BC4 hybrids. Only one of the BC2/BC2 hybrids was significantly higher than A632/Mo17, and 58% of the increase in MET in this hybrid was attributable to increased protein as a percentage of dry weight. By comparison, only 22% of the MET increase in the BC3/BC3 and BC4/BC4 hybrids resulted from increased protein as a percentage of dry weight. It is difficult to explain the large difference in MET levels between the BC2/BC2 hybrids and the best BC3/BC3 hybrids derived from these parental BC2 inbreds. Each of the Mo17BC3S2:3 high-MET lines derived from Mo17MET-158 had higher MET levels than Mo17MET-158. It is possible that this could be because of fixation of a favorable allele from BSSS53 in the Mo17BC3S2:3 lines that was heterozygous in the BC2S2:3. Alternatively, the additional backcross could have recovered an allele from Mo17 contributing to increased MET levels. Although Mo17 is fairly low in MET, results of a QTL mapping effort with F2:3 lines from the cross of BSSS53/Mo17 indicate that there is a QTL on chromosome 1 for which Mo17 contributes the positive allele (unpublished data, 2003).
The significant increase in MET levels of experimental hybrids indicated that selection of high-MET inbreds is useful for producing high-MET maize hybrids. Zuber and Helm (1975) reported that normal dent maize inbred lines with increased whole-kernel lysine levels did not produce hybrids with increased whole-kernel lysine levels. They speculated that this might have been caused by the lower germ–endosperm ratio of the hybrid kernel since the embryo has a considerably higher lysine level than the endosperm. Comparison of the MET levels of the embryo and endosperm fractions of BSSS53 and Mo17 revealed that the majority of the MET increase in BSSS53 is in the endosperm (Olsen, 1999). It is possible that the increased MET level of the selected high-MET inbred lines is occurring in the endosperm and that the lower embryo-endosperm ratio of the hybrids does not prohibit high-MET levels from being attained.
Although selection on inbred MET levels enabled the selection of lines that produce high-MET hybrids, the relationship between inbred MET levels and hybrid MET levels was not strong enough to enable accurate prediction of hybrid MET levels. It should be noted, however, that the parameter estimates associated with the regression of hybrid MET on midparent MET in this study may not accurately reflect the results that might be expected if an unselected group of inbred parents were crossed. By including more low-MET parents (as would be the case for unselected lines), the variance of the independent variable would increase, raising the coefficient of determination and improving the perceived predictive power. Still, these results suggest that final discrimination of the most useful high-MET inbreds will need to be determined in testcross trials.
|
SUMMARY |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONSUMMARYREFERENCES |
|---|
Approximately half of the increase in MET observed in the B73 genetic background could be attributed to increased grain protein concentration, and half of the increase was apparently because of an altered amino acid profile. In the Mo17 genetic background, only 22% of the MET increase resulted from increased protein concentration. In both genetic backgrounds, increased kernel protein levels contributed significantly to the final MET increases.
Significant increases in whole-kernel MET were observed in each of the B73/A632, B73/Mo17, and A632/Mo17 hybrid trials. Whole-kernel MET was increased by 0.45 g kg-1 over B73/A632, 0.70 g kg-1 over B73/Mo17, and 0.65 g kg-1 over A632/Mo17. Higher protein levels generally accompanied increased MET levels in the experimental hybrids. The identification of high-MET hybrids from crosses of selected high-MET inbred lines indicates that selection for increased whole-kernel MET can be implemented to improve the nutritional quality of maize hybrids.
|
ACKNOWLEDGMENTS |
|---|
|
NOTES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONSUMMARYREFERENCES |
|---|
Received for publication January 8, 2002.
|
REFERENCES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONSUMMARYREFERENCES |
|---|
Related articles in Crop Science:
This article has been cited by other articles:
|
|
 |
|
  R. L. Phillips, J. Suresh, M. Olsen, and T. Krone Registration of High-Methionine Versions of Maize Inbreds A632, B73, and Mo17 Journal of Plant Registrations, September 1, 2008; 2(3): 243 - 245. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Feil, S. B. Moser, S. Jampatong, and P. Stamp Mineral Composition of the Grains of Tropical Maize Varieties as Affected by Pre-Anthesis Drought and Rate of Nitrogen Fertilization Crop Sci., January 31, 2005; 45(2): 516 - 523. [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 | |||
