| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a USDA-ARS Sugarcane Field Station, 12990 U.S. Highway 441 N, Canal Point, FL 33438
b School of Plant, Environmental, and Soil Sciences, Louisiana State Univ. Agric. Center, Baton Rouge, LA 70803-2110, current address: Punjab Agricultural Univ., Ludhiana 141 004, India. Mention of trade names or commercial products is solely for the purpose of providing specific information and does not imply recommendation or endorsement by USDA, Louisiana State Univ. Agric. Center, or Punjab Agric. Univ. over others not mentioned
* Corresponding author (Barry.Glaz{at}ars.usda.gov).
 |
ABSTRACT |
|---|
 TOPNOTES ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: CP, Canal Point • DU, A. Duda and Sons, Inc. • CGL, crop cycle x genotype x location • EAA, Everglades Agricultural Area • EG, Eastgate Farms • GGE, genotype main effect (G) plus genotype x environment interaction effect (GE) • GL, genotype x location • KN, Knight Management, Inc. • LY, Lykes Brothers, Inc. • OK, Okeelanta Corporation 1 • OU, Okeelanta Corporation 2 • OS, Sugar Farms Cooperative North-Osceola Region • SF, Sugar Farms Cooperative North-SFI Region • TRS, theoretical recoverable sucrose • WW, Wedgworth Farms, Inc
|
INTRODUCTION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
The organic soils in the EAA need substantial micronutrient fertilization (Sanchez, 1990). More importantly, from a cultivar selection perspective, microbial oxidation of these organic soils makes excessive N available to sugarcane. Estimates suggest that 892 kg N ha–1 are mineralized annually under current growing conditions in the EAA (Glaz and Gilbert, 2006). Compared with other sugarcane-growing regions with annual growth cycles, sugarcane in the EAA has low values of theoretical recoverable sucrose (TRS). It is hypothesized that these low TRS values are caused by the sugarcane plant responding to the excessive N in these organic soils by producing more vegetative growth with less partitioning of photosynthates toward sucrose. Thus, a primary challenge for developing productive sugarcane cultivars in the EAA is to identify genotypes that produce high levels of cane tonnage and TRS in the presence of high N nutrition.
Commercial sugarcane yields in Florida improved substantially from 1968 through 2000 because of productive cultivars released by the CP program. Edmé et al. (2005) reported a mean annual improvement in TRS of 0.80 g kg–1 cane for this 33-yr period, which represented a 26% improvement in TRS and indicated that the CP program successfully identified cultivars with improved TRS yields in the high-N environment of the EAA. Edmé et al. (2005) also reported annual increases in commercial cane and sucrose yields of 0.31 and 0.10 Mg ha–1, respectively, for the same 33-yr period. These translated to 15.5 and 47.0% increases in cane yield and sucrose yield, respectively.
The commercial yield gains reported by Edmé et al. (2005) were for the entire Florida sugarcane industry. However, the gains were not similar across soil types (i.e., organic and sand soils). Gains attributable to CP cultivars on organic soils were similar to those of the entire industry, whereas there was no significant change in production on sand soils. Thus, a comprehensive review of the CP program to develop strategies that improve cultivar selection on sand soils without compromising successful cultivar selection on organic soils is needed.
Typically, an initial planting of about 100,000 seedlings from true seed is made in the CP program, followed by four clonally propagated genotype selection stages with a selection intensity of approximately 10% at each selection stage (Poehlman, 1986; Brown and Glaz, 2001). The seedlings and the first two clonally propagated selection stages are planted on organic soil at Canal Point. It is not until the third clonal selection stage, when sufficient planting material is available, that genotypes are tested in commercial fields at multiple sites (i.e., three organic-soil locations and one sand-soil location). In the final selection stage (fourth clonal stage), traditionally, performance testing occurred on eight locations with organic soils and one location with a sand soil. Recently, a second sand location was added in the final testing stage. However, the most recent complete set of retrospective data available for this study was with a sand soil at only one location and organic soils at the remaining eight locations.
One hypothesis that needs to be tested as part of the comprehensive review of the CP program is that the final selection stage does not effectively identify genotypes that yield well on sand soils because of the disproportionate representation of sand locations (eight organic to two sand) in this stage. Alternatively, improved cultivar selection for sand soils may require more changes to earlier stages of the selection program rather than, or in addition to, changes in the final selection stage. As part of the comprehensive review of the CP program, several studies of the early selection stages are being undertaken to answer these questions. A major concern in conducting early selection stages with large numbers of genotypes in small, unreplicated plots on sand soils in Florida is that sugarcane growth is highly variable on these soils. The present study was undertaken because replacing at least one organic-soil location with a sand-soil location in the final genotype-testing stage was considered a rational step due to the low representation of sand soils in this stage. This, however, requires an assessment of the relative contributions and importance of sand and organic test sites through a scientific approach. This approach should seek replacement of an organic-soil location with a sand-soil location that would not seriously compromise genotype selection on organic soils.
The last formal assessment of CP testing locations was reported by Glaz et al. (1985). Since that assessment, Yan (2001) has developed GGEbiplot software to help graphically visualize performance patterns of genotypes and to efficiently assess representativeness and discriminating ability of test locations. The biplot technique has been used to assess genotype x environment interactions in a variety of crop species (Ma et al., 2004; Casanoves et al., 2005; Kang et al., 2005; Blanche and Myers, 2006; Fan et al., 2007).
To improve the efficiency of the CP cultivar selection program in Florida, it is imperative to conduct a comprehensive analysis of historic data from the final testing stage and identify possible organic-soil locations as replacement candidates for a sand location. Thus, the objectives of the present study were to assess the contributions of a location with a sand soil to the final stage of multienvironment testing of sugarcane genotypes in Florida, and to identify locations with organic soils that, if replaced with a sand soil location, would be least likely to compromise the ability of the CP program to identify productive sugarcane cultivars for organic soils in Florida.
|
MATERIALS AND METHODS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
.
|
|
|
|
The 14 experimental genotypes were CP 98-1107 (G07), CP 98-1118 (G18), CP 98-1139 (G39), CP 98-1325 (G325), CP 98-1335 (G335), CP 98-1417 (G17), CP 98-1457 (G57), CP 98-1481 (G81), CP 98-1497 (G97), CP 98-1513 (G13), CP 98-1569 (G69), CP 98-1725 (G725), CP 98–2047 (G47), and cultivar CP 98-1029 (G29) (Edmé et al., 2006). Reference cultivar CP 72-2086 (G86) (Miller et al., 1984) was planted at all nine locations, reference cultivar CP 70-1133 (G33) (Rice et al., 1978) was included at the seven locations planted in 2001, and reference cultivar CP 89-2143 (G43) (Glaz et al., 2000) was included in the two locations planted in 2002. CP 70-1133 was widely planted on sand soils in Florida for about 20 yr and on organic soils for about 5 yr; its use declined after it became susceptible to brown rust (caused by Puccinia melanocephala Syd. and P. Syd.).
Parental genotypes in the CP program are recycled. Any genotype that advances to the fourth clonal stage and flowers is used as a parent for at least 1 yr. Eight of the 17 genotypes used in this study had at least one parent that did not advance beyond the fourth clonal stage. Sugarcane genotypes selected in Louisiana and Texas are also used as parents in the CP program. Four of the 17 genotypes in this study had one parent selected in Louisiana and one genotype had a parent that was selected in Texas. Seven of the genotypes in this study had at least one parent that was a commercial cultivar in Florida. Cultivars CL 61-620 (Holder and Todd, 1981), CP 63-588 (Rice et al., 1969), CP 70-1133, CP 72-1210 (Miller et al., 1981), and CP 84-1198 (Glaz et al., 1994), all historically widely planted on both organic and sand soils in Florida, were parents or grandparents of seven of the 16 genotypes used in this study. Parents of each of the 17 genotypes were reported by Glaz et al. (2005).
Except for the trial at OU, all trials were planted in fields that had not been cropped to sugarcane for at least 9 mo. The trial at OU was planted within 2 mo of the final ratoon harvest of a sugarcane crop, thus it was planted in a successive sugarcane cycle.
Experimental units were three rows wide, an outside border row, one inside row, and a third inside row that bordered the inside row of the adjacent plot. The two inside rows of each plot were used for yield determination. Plot rows were 10.7 m long and the distance between rows was 1.5 m. Total number of stalks predicted to be sufficiently large for harvest from the two inside rows of each plot was recorded each year from June through September. One sample consisting of 10 stalks was cut from the middle row of each plot from October through March. Stalk weights were determined from these samples. Cane yield in megagrams per hectare was calculated by multiplying stalk weight by number of stalks. Brix and pol were determined for each 10-stalk sample and from these values, TRS was determined as grams of sucrose per kilogram of cane according to Legendre (1992). Sucrose yield in megagrams per hectare was the product of cane yield by TRS divided by 1000. Full explanations of trial planting and sampling procedures are reported in Glaz et al. (2005).
Data Analysis
The trial at each location was conducted using a randomized complete-block design with six replications. Genotypes and crop cycles were considered as fixed effects and locations and replications were considered as random effects. Analyses of variance were computed using PROC MIXED (SAS Institute, 2003). Data were analyzed for each crop cycle separately and analyses were also conducted with the combined data of the plant, first-ratoon, and second-ratoon crops. Results were declared significant at P = 0.05 and highly significant at P = 0.01.
The GGEbiplot software (Yan, 2001) was used to visualize the genotype x environment two-way data; GGE denotes genotype main effect (G) plus genotype x environment interaction effect (GE). Information on this software is available at http://www.ggebiplot.com (verified 14 Mar. 2008). Data from all locations were combined to construct the biplots even though seven locations were planted in 2001 and two locations were planted in 2002. The rationale for combining all locations was that an objective of the study was to identify locations with organic soils that, if replaced by a location with a sand soil, would be least likely to compromise the ability of the CP program to identify productive sugarcane cultivars for organic soils in Florida. In this genotype selection program, the locations that were planted in 2002 are routinely planted 1 yr after the seven locations that were planted in 2001; thus comparing all nine locations together was the most logical approach. This approach is similar to what Yan et al. (2007) classified as the Type 3 mega-environment because we used multiyear and multilocation testing to identify sugarcane cultivars for both major soil types in Florida based on mean performance and stability.
Two GGE biplots were generated for each yield trait (cane yield, TRS, and sucrose yield). One issue when comparing locations in a genotype selection program with the intent to optimize locations when resources are limited is to determine if some locations provide similar information on genotype differentiation. A second issue is to identify locations that better discriminate among genotypes than other locations. The biplot software provides output that can help visualize both of these location characteristics. In the "Relationship among testers" output, the cosine of the angle between two vectors approximates the correlation between the two locations represented by their vectors (Fig. 1
). The length of a vector represents a location's ability to discriminate among genotypes (Yan and Kang, 2003)—the longer the vector, the greater the discriminating ability. For all biplots (Fig. 1–6
), data were not transformed ("Transform = 0") or scaled ("Scaling = 0") before biplot analysis, and data were environment-centered, meaning that biplots displayed genotype and genotype x location (GL) effects, but did not display the main environmental (L) effect ("Environment = 2"). These biplot options were explained in more detail by Yan et al. (2007).
|
|
|
|
|
|
|
RESULTS AND DISCUSSION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
). Therefore, all characters were analyzed separately in the plant-crop, first-ratoon, and second-ratoon crop cycles. The GL interaction was highly significant for each of the three traits analyzed in each crop cycle. The variation attributed to the GL interaction was smaller than the variation among genotypes for TRS in the plant-crop and first-ratoon crop cycles (Table 3). In the second-ratoon crop, variation for TRS attributable to the GL interaction was greater than variation among genotypes. Variation in either cane or sucrose yield attributed to the GL interaction was greater in all three crop cycles than was variation among genotypes (Tables 2 and 4). In the plant-crop cycle, variation among genotypes (43.06%) accounted for more of the total TRS variability associated with genotypes and locations than locations (27.41%) (Table 3). Otherwise, for TRS, cane yield, and sucrose yield, variation among locations was greater than that among genotypes or GL. Except for TRS in the plant-crop cycle, variation among locations for the three traits explained between 55 and 85% of the total variation (Tables 2–4). That variation among locations was generally the most important source of variation justifies the selection of the biplots based on the site regression model (Yan et al., 2000; Yan and Kang, 2003). The amount of variability attributed to the GL interaction in the combined analysis of each character was similar to that attributed to the CGL interaction, suggesting that the GL interaction is expected to be repeatable across crop cycles. Therefore, the biplots will be presented using overall crop analyses, but important information from individual crop analyses that differs from combined crop analyses will be noted.
The first two principal components (PC1 and PC2) obtained by singular-value decomposition of environment-centered or within-environment standardized genotype x environment data (Yan et al., 2007) explained from 59.1 to 82.3% of the total variability caused by G + GL for cane yield, TRS, and sucrose yield (Fig. 1–6). These percentages indicated that the GGE biplots of PC1 vs. PC2 adequately displayed the GGE patterns for the three traits.
Yan and Rajcan (2002) defined a mega-environment as a group of locations that consistently shared the most productive set of genotypes across years. Yan et al. (2007) explained how to identify mega-environments within a data set using the "which-won-where" view of the GGE biplot. The which-won-where biplot of each trait indicated that all nine locations in this study were in the same mega-environment except for DU and EG, each of which was not in this mega-environment for two of three traits. In the biplot for cane yield, both DU and EG were not in the mega-environment that included the other seven locations; in the TRS biplot, only DU was not in the same mega-environment as the other eight locations; and in the biplot for sucrose yield, only EG was not in the same mega-environment as the other eight locations (data not shown).
The major mega-environment for each trait included LY, the sand-soil location. We chose to include DU and EG in this study to determine preliminarily if one of these locations would be a logical one to replace with a sand-soil location. The highest yielding genotype in the major mega-environment was cultivar CP 98-1029 which was the only genotype from this group that was released, and it was released for both organic and sand soils (Edmé et al., 2006).
Cane Yield
Locations OS and KN were highly correlated (small angle between their vectors) and nearly identical in their ability to discriminate among genotypes for cane yield, as revealed by the length of their vectors (Fig. 1). Also, genotypes had nearly identical relative cane yields at these two locations. Relative genotype performance was nearly identical at OK and OU. Cane yields among genotypes were similar at WW compared with OS and KN, and were also similar at WW compared with LY. Among these four locations, genotype discrimination was lowest at WW. Discrimination of genotypes was highest at EG and was also relatively high at OU and LY. Locations SF and DU were valuable locations for testing cane yield based on the large angles formed by their vectors with those of other locations.
The ideal location for testing cane yield among sugarcane genotypes in Florida was OU; it was nearly on the average environment axis and its discriminating ability was relatively high (Fig. 2). Locations LY, the only location with a sand soil, and SF were also valuable for testing genotypes because they were both discriminating and highly representative. Although OK was highly representative, its vector was somewhat shorter than vectors of others, indicating that it was not highly discriminating among genotypes. Eastgate (EG) was identified as a valuable location because it was highly discriminating, but it was the least representative. This was not surprising because EG often has the highest yields among locations in the CP testing program. Because it is located closest to Lake Okeechobee, EG benefits more than other locations from the moderating temperature effects of the lake during winters. Also, the organic soil at EG, a Torry muck, has substantially more mineral content than the soils at the other locations with organic soils (Table 1), and the organic soil at EG has a depth to bedrock of about 3 m compared with less than 1 m for the other organic soils. In a trial of recently released cultivars, Gilbert et al. (2006) also reported differences in sugarcane cultivar performance at a Torry muck site compared with locations with organic soils composed primarily of decomposed sawgrass.
Based on 3-yr cane yields, the sand location, LY, provided valuable information to the CP cultivar selection program (Fig. 1 and 2). Were a decision made to replace an organic-soil location with a new sand-soil location, OS, KN, or WW would be a logical choice based on cane yields. These three locations provided similar information on genotypes and were similarly (poorly) representative of the Florida sugarcane industry. Locations OS and KN provided nearly identical information, and although information provided by WW differed moderately, WW had poor discriminating ability. In individual crop-year analyses of cane yield, results were similar except that in the second-ratoon crop, KN was the least representative but most discriminating location (data not shown).
Theoretical Recoverable Sucrose
Six of the nine locations provided similar information for relative TRS production among genotypes (Fig. 3). Locations OK and KN nearly superimposed on each other, which indicated that the TRS-based ranking of genotypes was nearly identical at these two locations. Along with OK and KN, TRS ranking among genotypes was similar at SF, LY, OS, and OU. Among these locations, discrimination among genotypes was highest at OU, and OU also discriminated well among genotypes for cane yield (Fig. 1). Because OU was at one of the extreme sides of this group of six similar locations, and it was the most discriminating, it emerges as a valuable location for testing TRS among sugarcane genotypes. For purposes of testing TRS, it would be logical to replace any organic soil location, from among KN, OK, SF, and OS, with a location with sand soil.
Among the three locations that were not in the group of six similar locations, EG was the most extreme to one side, and also highly discriminating among genotypes, making it another valuable location for testing TRS. Duda (DU) had a short vector, indicating that it did not discriminate well among genotypes, but it was different from all other locations, making it valuable. Wedgworth (WW) had a relatively short vector, but, although moderately similar to EG, based on its separation from the group of six similar locations, WW should be considered valuable from the standpoint of TRS for the CP cultivar selection program.
The three locations that differed from the others in relative genotype performance for TRS represent locations with high cane production. At DU, the genotypes are often exposed to high fertility. Sugarcane and vegetables are rotated more at DU than at other locations, and higher residual P and K are available at DU because these nutrients are added to vegetable crops at high rates. As discussed previously, cane yields were high at EG due to the moderating temperature effects of Lake Okeechobee and the higher mineral content of its organic soils. The organic soils at WW are generally deeper than at other locations, and although not as close to Lake Okeechobee as EG, winters are generally not as cold at WW as at several of the other locations with organic soils.
It was surprising that relative genotype TRS performance at LY was similar to other locations. Lykes (LY) was the only location in this study with a sand soil and generally, TRS yields are substantially higher on sand soils than on organic soils. It is assumed that the lower TRS values on organic soils are related to the extremely high N availability on those soils. High rates of N encourage high cane tonnage on organic soils, but relatively low TRS yields compared with most other regions where sugarcane is grown. By contrast, there is not excessive N available at LY where N nutrition is provided primarily through N fertilization. Thus, TRS values are usually higher and cane yields are lower at LY compared with the locations with organic soils. These results suggested that to identify sugarcane cultivars with high 3-yr mean TRS values for sand soils in Florida, it might not be necessary to test genotypes on sand soils. However, this was a case where using 3-yr means would be misleading. The biplots for individual year TRS yields identified LY to be different from most other locations in both the first- and second-ratoon crops (data not shown).
As it was for cane yield, OU was the location considered to be ideal for testing sugarcane genotypes for TRS in Florida (Fig. 4). Although not exactly on the average environment axis, OU was more discriminating than OS, SF, OK, KN, and LY, all of which were also close to the center of the innermost circle. However, all six of these locations are not essential for identifying sugarcane genotypes that will have high TRS yields in Florida. Locations KN, OK, SF, and OS are all logical choices for replacement with a sand location. Although all four are nearly ideal locations, they rank and discriminate among the genotypes similarly for TRS (Fig. 3 and 4).
The sand soil location, LY, discriminated poorly among genotypes for TRS. It is generally expected in Florida that residual variation on sand soils is greater than on organic soils for similar experiments. The poor discrimination among genotypes at LY suggests that the CP program may compromise its overall ability to discriminate among genotypes for TRS as more organic-soil locations are replaced with sand-soil locations.
Gilbert et al. (2006) noted that TRS relationships among genotypes were dependent on when during the harvest season the TRS information was collected. The present experiment was not designed to deal directly with this issue. However, all plant-cane TRS samples, except one, were collected in January and February, and all second-ratoon TRS samples, except one, were collected in October. In both cases, the one exception was EG, and this may partially explain the differences in relative TRS performance among genotypes at EG compared with other locations.
Sucrose Yield
Eastgate was a valuable location for testing sucrose yield of new sugarcane genotypes; its long vector indicated that discrimination among genotypes was greater at EG than at any other location (Fig. 5). An ideal location would be one that was both discriminating among genotypes and representative of other locations. There was no clear ideal location for testing sucrose yield (Fig. 6). Eastgate was highly discriminating, but its angle with the nearest location vector was about 90°, indicating that it is nearly independent of the other locations, making it a unique location (Fig. 5). Locations that were closest to the inner circle, SF, DU, OU, OK, and WW (Fig. 6), had short vectors, meaning that they did not discriminate well among genotypes (Fig. 5).
The vector representing SF nearly overlapped with WW on one side and with DU on the other side. In each case, SF would be the location that provided the most useful information to the genotype-testing program because its vector was the longest of the three. Similarly, OK and OU were highly correlated and OU was more discriminating than OK. Although not highly discriminating, OS fell in a unique region, making it least representative of the group.
Five locations (Group 1), WW, SF, DU, OK, and OU, were well correlated. The angle separating the mean vector of these locations and EG was about 90° in one direction, and the angle separating the mean vector of OS, LY, and KN (Group 2) from the mean vector of Group 1 was close to 90° in the other direction. Discrimination among genotypes in Group 1 was lowest at WW and OK, and in Group 2 it was low at all three locations (OS, LY, and KN). Thus, based on results for sucrose yield, one logical choice would be to replace either OS or KN with a location with a sand soil, and a second choice would be to replace either WW or OK. In analyses of sucrose yield by individual year, results were similar except that in the second-ratoon crop, KN was the least representative but most discriminating location (data not shown).
Each year, the experimental genotypes at OU and EG were planted the year after the other seven locations. Although it was not a pre-identified purpose, results of this study indicated that OU and EG provided valuable information for the CP sugarcane selection program. Eastgate was generally not representative of other locations; for sucrose yield, it was essentially independent of other locations. Like OU, EG was often highly discriminating among genotypes. Although both locations have similar soils and climates, OU was generally more discriminating of genotypes than OK. The difference between these two locations was that OU was planted in a successive rotation. Glaz and Ulloa (1995) reported that the mean sugar yield in successively planted sugarcane in Florida was about 19% less than in fallow-planted sugarcane. They determined that the later planting dates of successive planting accounted for about 30% of this yield loss. Meyer and Van Antwerpen (2001) reported that possible reasons for yield losses in continuously cropped sugarcane in South Africa were decreased soil pH, loss of soil organic matter, and changes in soil biota. Perhaps the stress of successive planting helped discriminate among genotypes at OU.
As it did for TRS, the sand-soil location, LY, discriminated poorly among genotypes for sucrose. This adds further to the concern that the CP program should monitor ability to discriminate traits among genotypes if progressively more locations with sand soils replace locations with organic soils.
|
CONCLUSIONS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
NOTES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Received for publication June 8, 2007.
|
REFERENCES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  J. C. Comstock, B. Glaz, S. J. Edme, R. W. Davidson, R. A. Gilbert, N. C. Glynn, J. D. Miller, and P. Y.P. Tai Registration of 'CP 00-1446' Sugarcane Journal of Plant Registrations, January 1, 2009; 3(1): 28 - 34. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. Glaz, S. J. Edme, R. W. Davidson, R. A. Gilbert, J. C. Comstock, N. C. Glynn, J. D. Miller, and P. Y.P. Tai Registration of 'CP 00-2180' Sugarcane Journal of Plant Registrations, January 1, 2009; 3(1): 35 - 41. [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 | |||
