| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Plant Sciences Div., Univ. of Missouri, Columbia, MO 65211
b Agronomy Dep., Lilly Hall, Purdue Univ., West Lafayette, IN 47907
c Crop Science Dep., AW-101 Turner Hall, Univ. of Illinois, Urbana, IL 61801
* Corresponding author (scharfp{at}missouri.edu)
Received for publication March 8, 2005.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: CM, chlorophyll meter • EONR, economically optimal nitrogen rate
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
The Minolta SPAD (Konica Minolta, Hong Kong) CM measures transmission of red and near-infrared light through individual leaves. Its output is in arbitrary units and has been shown to be strongly related to leaf chlorophyll concentration (Markwell et al., 1995). Previous research has shown that CM readings can reliably indicate N stress in corn. Research in corn has focused mainly on differentiating locations that will respond to N fertilizer from locations that will not, on evaluating the meter as a tool to indicate whether and when fertigation is needed, and on relationships between meter readings and soil and plant N content. Findings in the first two areas are described briefly below.
Several studies have shown that CM readings of corn at stage V6 were able to separate responsive from nonresponsive sites with 65 to 80% accuracy (Piekielek and Fox, 1992; Jemison and Lytle, 1996; Sims et al., 1995; Zebarth et al., 2002). In these four studies, the success rate of the CM in separating responsive from nonresponsive sites was essentially identical to that of the pre-sidedress soil nitrate (NO3–) test. One of the studies (Sims et al., 1995) used only absolute CM readings. The other three (Piekielek and Fox, 1992; Jemison and Lytle, 1996; Zebarth et al., 2002) used both absolute and relative CM readings and found them about equally effective in separating responsive from nonresponsive sites.
Chlorophyll meter readings would be much more useful for making management decisions in corn if they could, in addition to predicting whether corn will respond to N fertilizer, predict how much fertilizer is needed, or the size of the expected yield response, in time to make management decisions. The prediction of how much N to apply appears to be more generally useful for making management decisions, but in cases where a rescue N application is being considered (e.g., weather prevented a planned N application, or resulted in loss of previously applied N fertilizer), an assessment of the size of the yield reduction that could be expected if nothing is done would be a useful decision aid.
An indirect answer to how much N is needed has been developed for irrigated corn, where irrigation water can be used as a N delivery system with repeated opportunities for application. In this system, the question of how much N to apply can be answered by repeatedly checking the N status of the corn with a CM and applying a fixed low rate of N whenever the meter reading falls below a critical value (Varvel et al., 1997; Shapiro, 1999). This system has relied on relative CM readings, that is, the reading from the area being diagnosed divided by the reading from a reference area that has received a high or nonlimiting N rate. There do not appear to be any reports contrasting the accuracy of absolute vs. relative CM readings in this type of management system.
In rainfed systems, the opportunity for making N applications is more constrained. The CM will only be useful in guiding N application rate if it can be the basis for a single quantitative rate recommendation. Results to date have been mixed. There do not appear to be any reports regarding the relationship between relative CM readings and the amount of N needed. Scharf (2001) reported that absolute CM readings at the V6 stage were related to EONR and produced N rate recommendations that were lower than N rates used by producers in the same fields, but that performed as well or better economically than producer rates. In contrast, Bullock and Anderson (1998) concluded that absolute CM readings were not useful for predicting N fertilizer need. However, average yield response to N was only 0.5 Mg ha–1 in their six experiments, indicating that little N stress was observed and little N was needed. Piekielek and Fox (1992) also concluded that absolute CM readings were not useful for making N rate recommendations, but they reported a correlation coefficient between CM reading and soil N-supplying capacity of 0.67, which was similar to the correlation coefficients between soil N-supplying capacity and the pre-plant or pre-sidedress soil NO3– tests. Deciding what level of accuracy is adequate to serve as a basis for N rate recommendations is difficult, but predictive relationships with R2 of 0.5 can provide N rate recommendations that are economically superior to current producer practices (Scharf, 2001; Scharf et al., 1993). Additional environments and growth stages need to be investigated before firm conclusions can be drawn regarding the usefulness of CM readings in making N fertilizer rate recommendations for corn.
Some success has been achieved with making predictions of EONR or yield response to N for other crops based on CM readings. In wheat (Triticum aestivum L.), the relationship between CM readings and EONR has been successfully calibrated (Murdock et al., 1994), with differential readings (difference between observed CM value and high-N reference CM value) performing better than absolute readings. In rice (Oryza sativa L.), a relationship to yield response to N based on absolute meter readings has been established (Turner and Jund, 1991).
Our objective was to develop calibrations to predict corn N need and yield response based on CM readings over a wide range of environments and growth stages to improve N rate recommendations and inform management decisions.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Sixty-six N rate experiments were conducted in seven north-central states [Illinois (5), Kansas (3), Michigan (4), Minnesota (12), Missouri (4), Nebraska (28), and Wisconsin (10)] (Fig. 1 ) over a 4-yr period from 1995 to 1998. Weather conditions were generally favorable for corn production in these years. Experimental treatments were N rates ranging from zero to a rate that was expected to be nonlimiting, which varied from one experiment to another. All but seven of the experiments used at least five N rates, and all but one had at least four N rates. Fifty-four of the experiments had a top N rate of 200 kg ha–1 or higher. Of the 12 experiments with top N rates <200 kg ha–1, 8 experiments clearly had no economic response to N from the second highest to the highest N rate. Three experiments with possible economic yield responses from the second highest N rate to the highest N rate nonetheless had very high yields (ranging from 11.3–13.4 Mg ha–1) at the second-highest N rate. Thus it seems likely that in all experiments, the highest N rate was in fact nonyield limiting or close to it.
|
Broadcast ammonium nitrate was used as the N fertilizer source at most locations. Exceptions were the 5 Illinois experiments, in which injected urea-ammonium nitrate solution was the N source, and 10 of the Minnesota experiments, which used urea broadcast and incorporated the same day as the N source. Fertilizer was applied either between planting and emergence or as a sidedress application at growth stage V4 to V7.
Twenty-six of the 66 experimental locations either had a manure history or had received manure in the study year. Among locations receiving manure, an attempt was made to avoid sites where it was virtually certain that no yield response to N fertilizer would occur.
Previous crop was soybean [Glycine max (L.) Merr.] or corn in 58 experiments, while eight experiments had alfalfa (Medicago sativa L.) or a small grain as the previous crop. Most of the experiments conducted in Nebraska and Kansas were irrigated, with a few irrigated experiments in other states as well.
A broad range of soil properties were represented in these experiments. Soil texture ranged from sandy loam to clay loam, with silt loam and silty clay loam as the most common textural classes. Soil organic C ranged from 9 to 42 g kg–1.
Soil properties with potential to predict N mineralization were characterized for the surface soil of each site, including organic C, total N, water pH, salt pH (in 0.01 M CaCl2), phosphate-borate extractable NH4+ (Gianello and Bremner, 1986) steam distilled for either 4 or 8 min, borate-extractable NH4+ (same procedures but phosphate omitted), hot KCl-extractable NH4+ (Gianello and Bremner, 1986) steam distilled for either 4 or 8 min, NH4+ released by incubation with hot KCl for 20 h (Øien and Selmer-Olsen, 1980), NH4+ released with aerobic (Bremner, 1965) or anaerobic (Waring and Bremner, 1964) incubation, and bulk density.
Chlorophyll meter readings were taken using Minolta SPAD CMs. Twenty to 30 readings were taken per plot, and values from four plots (four replications) were combined to create the treatment average readings used in our analyses. At most experimental sites, meter readings were taken numerous times during the growing season, and the growth stage was recorded. Before tasseling, readings were taken on the uppermost fully expanded (i.e., collared) leaf. From tasseling onward, readings were taken on the ear leaf. All readings were taken midway between the stalk and the tip of the leaf (Peterson et al., 1993).
Relative CM value was calculated as:
 |
The reference CM value was specific to each experimental location and growth stage. This reference value was calculated by averaging all readings from a group of high N treatments. A group of high N treatments, instead of just the highest N rate, was used to maximize the number of observations used to calculate the reference value, and thereby to maximize its accuracy and robustness. This group always included the highest N rate treatment, plus other N rate treatments judged to be sufficient by the criteria that follow. If the N rate below the highest N rate: (i) produced an average yield no more than 0.3 Mg ha–1 below the yield with the highest N rate, and (ii) had average CM reading no more than two units lower than the reading from the highest N rate (at any stage during the season), then this treatment was also included in the group of high N treatments used to calculate the reference CM reading. If the second-highest N rate was included in this group, then the third-highest N rate was considered for inclusion based on the same two criteria. This procedure, using these criteria, was repeated until an N rate treatment was found whose average yield was more than 0.3 Mg ha–1 lower than the highest yield associated with a higher N rate, or that produced CM readings more than two units below the highest N rate. This N rate was not included in the high-yielding group, but all N rates above it were included.
Corn grain was harvested at physiological maturity, typically using plot combines or harvesting by hand. Reported grain yields are corrected to a moisture content of 150 g kg–1. Responsiveness to N was evaluated using a multi-step procedure. If analysis of variance with 
= 0.10 did not indicate yield differences between treatments at an experimental location, then EONR was assigned a value of zero. If analysis of variance indicated treatment differences, then four regression models (linear, quadratic, linear plateau, and quadratic plateau) were fitted to the yield data. The best model for describing the data at each location was selected using highest R2 and lowest standard error as criteria. This model was then used to calculate EONR for the location using a price ratio of 3.4 kg corn grain kg–1 fertilizer N. This price ratio was representative of the time period during which the experiments were conducted, but has since widened. Using the current price ratio would result in slightly lower EONR values at locations that were modeled as quadratic or quadratic-plateau responses.
Although EONR is a property of the experimental location, CM readings were taken from plots receiving a range of N rates at many locations. Thus it was necessary to calculate EONR values for each N-rate treatment, based on the EONR value determined for the location. For plots that had received N before taking CM readings, EONR was defined as the location EONR minus the N rate that had been applied. Similarly, the potential for yield response to additional N was calculated for each N rate treatment as non-N-limited yield for the site minus the average yield for that N rate.
Stepwise regression was used to model EONR as a function of independent site variables, including all soil measurements noted above and CM readings. This procedure was performed using the MAXR option of SAS software employing the default settings.
Linear and quadratic regression were used to relate CM readings over a range of growth stages to yield response variables. Both absolute and relative CM readings were used as independent variables in these models. Economically optimal N rate and yield response to N applied at or before stage V7 were used as dependent variables. Values used in the regression models were treatment means for a site. An F test with = 0.10 was used to determine whether two regression lines described the data better than a single line, and therefore were different. SAS software was used to perform all statistical calculations and procedures.
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Regardless of the growth stage at which CM readings were acquired, relative CM reading was the first variable entered into the stepwise regression model to describe EONR, with R2 ranging from 0.53 to 0.66, dependent on stage. The second term entering the model varied (depending on the stage of the CM readings used, and whether variables with missing values were included), but the increase in R2 due to the introduction of this second term was on average only 0.03. None of the soil measurements contributed to models for predicting EONR to a degree that would have any practical significance. This analysis demonstrated that relative CM reading was by far the strongest predictor of EONR in our data set, and most likely to be useful for making N-rate recommendations for corn production.
Soil NO3– measurements are one of the most-used diagnostic tools for N-fertilizer need, thus we gave special attention to them. Linear regression analysis of EONR as a function of soil NH4+ (preplant or pre-sidedress, to 30-, 60-, 90-, or 120-cm depth) for our data set produced coefficients of determination ranging from 0.04 to 0.16. In some cases, a few sites with EONR = 0 and high soil NO3– values produced a poor fit for linear regression as evaluated by regression residuals. Use of a linear-plateau model to relate EONR to soil NO3– improved fit in these cases and increased coefficient of determination to as high as 0.23. However, this is still much lower than the coefficients of determination seen for CM readings. Thus the focus of the remainder of our analysis and discussion will be on CM readings.
Relative CM reading was significantly related to EONR at all growth stages studied (Fig. 2 ). This relationship changed as the season progressed, making it necessary to divide the data into early- (Fig. 2A), mid- (Fig. 2B), and late-season (Fig. 2C) CM readings. Simple linear functions of relative CM reading were highly significant (P < 0.0001) predictors of EONR for all three time periods, and described from 0.528 to 0.656 of the total variability in EONR. Introduction of a quadratic term resulted in a slight improvement in the model's ability to describe the data. The quadratic term was barely significant in describing the V5 to V9 data (P = 0.07), but was highly significant in describing the V10 to R1 data (P = 0.011) and the R2 to R5 data (P < 0.0001). The accompanying increases in R2 with the introduction of the quadratic term were quite small: 0.008 for the V5 to V9 data, 0.006 for the V10 to R1 data, and 0.028 for the R2 to R5 data. This is reflected in the relatively small differences between the linear and quadratic functions, which deviate from each other mainly at low relative CM values (Fig. 2A, 2B, and 2C).
|
Timing of N application, however, complicates the use of the relationships in Fig. 2 for making predictions of how much N to apply. Our N treatments were applied sometime between planting and stage V7, and our EONR values are probably appropriate for N applications from V5 to V9 (Fig. 2A), but less appropriate for N applications from V10 to R1, while N applications after R2 are probably not very useful. There is evidence that EONR may change little if at all during the planting to V7 period (e.g., Scharf and Lory, 2002), and also that large yield response to N can be obtained throughout the vegetative period and at least into the early reproductive period (Scharf et al., 2002; Binder et al., 2000; Russelle et al., 1983). However, it is certain that at some point during the season, the potential for the crop to respond to N will decline and then disappear. For corn experiencing substantial N stress, a significant decline in yield potential can begin to occur as early as stage V6 (Binder et al., 2000) but more commonly does not occur until sometime between V12 and silking (Scharf et al., 2002; Binder et al., 2000; Russelle et al., 1983). There is little if any research that tracks this decline past silking, or that addresses whether N-use efficiency and EONR might change before the decline in yield response. It seems likely that the relationship shown for growth stages R2 to R5 would be useful mostly in diagnosing what N rates would have been optimal, rather than how much N to apply following the CM measurements.
Sawyer et al. (J.E. Sawyer, D.W. Barker, and J.P. Lundvall, http://extension.agron.iastate.edu/soilfertility/info/chlorophyl04.pdf, accessed October 2005, verified 20 Feb. 2006) have performed experiments and analyses similar to ours, with similar conclusions. In their Fig. 2, a relative CM value of 0.8 at stage R1 is associated with an optimal N rate of 150 kg N ha–1, while in our Fig. 2b it is associated with an optimal N rate of 138 kg N ha–1. The precision of their relationship in this part of the curve is probably lower than ours due to having only four data points with EONR > 160 kg N ha–1. In addition, their coefficient of determination for this relationship (0.69) is similar to ours (0.66).
A fundamental question about these CM calibrations is whether they are influenced by previous N fertilizer applications. In fields that have not yet been fertilized, yield response to N is very common. Guiding fertilizer rate decisions in this situation is one important use of the calibration. There are also situations in which it may be useful to diagnose the need for additional N in fields that have already received N fertilizer. One example is fields that have received fertilizer N but then have been subjected to weather-related N loss conditions. Another example is fields that have received a modest N application at planting, with the intention of later diagnosing whether more N is needed and responding appropriately. It is important to know whether the same calibrations can be used in these contrasting situations. Thus we have separated our data into treatments that received N fertilizer at planting and treatments that never received N fertilizer, in order to examine whether relationships between CM readings, EONR, and yield response are different for these two groups.
Our data suggest that, in general, when predicting EONR from relative CM readings, the same calibration can be applied regardless of previous N-fertilizer applications. Calibrations on the left side of Fig. 2 include treatments that have and treatments that have not received N fertilizer at planting. On the right side of Fig. 2, the same data are presented but are separated into these two respective groups. The best-fitting simple regression line is in all cases nearly identical (Fig. 2D, 2E, and 2F) and not statistically different ( = 0.10) for the two groups. In these graphs, EONR for treatments that had received N at planting was calculated as the site EONR minus the N rate applied at planting. Thus, it is an estimate of the additional N that would have been required to bring the treatment to full economic potential.
One possible exception to this conclusion comes at the earliest growth stages that we considered. If data for growth stages V5 and V6 are evaluated separately, there is some evidence that the relationship between relative CM reading and EONR is different for treatments that have received N fertilizer at planting and treatments that have not (Fig. 3
). The probabilities that the two regression lines are not different are 0.08, 0.19, and 0.03 for the V5, V6, and combined V5-V6 data, respectively. At these stages, corn that has received some N fertilizer at planting may often be nearly as dark as, and have nearly the same CM reading as, corn that has received a high N-fertilizer rate, even if additional N is needed to optimize yield. For the five location-treatments with relative CM 
0.95 at growth stage V5 or V6 when N had been applied at planting, EONR was 75 or greater, but at higher relative CM values, there was no relationship to EONR. For growth stage V7 and all later stages, there was no evidence of any difference between the regression lines with and without previous N. We suggest that caution should be used in making sidedress decisions based on CM data at stages V5 and V6 if some N fertilizer has already been applied to the crop. Because V5 and V6 are important stages for application of N with tractor-mounted equipment, the practical implications of this exception are relatively important.
|
Coefficients of determination for relative CM reading as a predictor of EONR or yield response to N were lower for the V5 to V9 stage than for later stages. This may be due to changes in the amount of soil-derived N available to the crop as the season progresses, such that later meter readings are a better measure of season-long soil-derived N supply, which in turn is an important determinant of EONR. Factors that may cause temporal changes in soil-derived N available to the crop include more complete root exploration, in-season N loss due to leaching or denitrification, and weather-related changes in soil N mineralization rates. Variations from site to site in temporal patterns of relative CM reading support this concept. For example, in some experiments where large yield response to N and high EONR were observed, relative CM readings started low and dropped slowly, while in other experiments they stayed fairly high for a while, then dropped precipitously.
The slope of the relationship between relative CM reading and EONR also changed as the season progressed, becoming shallower later in the season (Fig. 2). This was due to relative CM values usually declining over time in a given treatment, which was in turn due mainly to increasing meter values in well-fertilized reference plots, rather than to declining values in stressed plots (Fig. 4 ). Sunderman et al. (1997) observed a very similar trajectory for CM readings from well-fertilized plots over the course of a growing season.
|
|
|
|
Although relative CM readings were significantly related to EONR, there was a wide range of possible EONR values associated with any relative meter reading at all growth stages studied (Fig. 2). Some of the variability in the relationship between CM readings and yield response variables is related to the time-scale issues discussed above. Shifts in soil N availability that occur after a diagnosis of EONR is made represent an unavoidable and usually insoluble source of error.
Variability in EONR at a given relative CM value may also be associated with variability in N-uptake efficiency and, once N is taken up, in the physiological efficiency with which it is converted to grain. Nitrogen-uptake efficiency may vary for a variety of reasons, one of which is that dry conditions tend to reduce uptake (Eghball and Maranville, 1993). Physiological efficiency may depend on environmental conditions (lower when environmental conditions limit yield but not N uptake) and genetics, and tends to decrease slightly as yield increases (Cassman et al., 2002).
Additional variability in the relationship between CM readings and yield response variables is related to spatial scale issues. By using EONR and yield response to N as site variables, and correspondingly using site-average CM readings, we are ignoring spatial variability within individual experiments. Although these experiments were fairly small (<0.5 ha) and areas were selected for uniformity, spatial variability within experiments is probably responsible for some of the data scatter in our calibrations. It is well established that soil NO3– content (Cahn et al., 1994), soil N mineralization, soil water relations, soil compaction, and other factors that affect corn growth and yield can vary widely over short distances.
Spatial variability also creates an issue when applying our calibrations to make management decisions. Mamo et al. (2003) have shown that a substantial amount of spatial variability in EONR can exist within a field. Even if a field is subsampled extensively to accurately determine the field-average relative CM reading, and even if the N-rate recommendation based on this reading is the true field-average EONR, substantial areas of the field may still be over- and underfertilized. Techniques for basing N-rate recommendations on corn color measurements that are more easily collected in a spatially intensive manner may prove to be more practical than CM readings for routine whole-field management. Both aerial imagery (Blackmer et al., 1996; Scharf and Lory, 2002) and ground-based radiometers (Bausch and Duke, 1996) appear to have promise for making color measurements in a spatially-intensive manner. Bausch and Duke (1996) have shown that relative CM values and relative radiometer values are strongly correlated.
In some cases, when the main management decision to be made is not how much N to apply, but whether or not to apply N, a prediction of expected yield response to N may be more useful than a prediction of EONR. Fields that have received their primary N-fertilizer application but that have experienced weather conditions conducive to N loss are one example of this type of situation. Our data suggest that relative CM readings can be used to make reasonably good predictions of yield response to N (Fig. 6). The same points regarding the timing of N application, made above in regard to EONR, are true for yield response to N as well.
Predictive value of absolute CM readings was in all cases statistically significant (P < 0.01) (Fig. 7), but of lower quality than predictions based on relative CM readings. We found a much larger difference between absolute and relative CM readings in terms of predictive value than previously reported (Piekielek and Fox, 1992; Jemison and Lytle, 1996; Zebarth et al., 2002). A possible explanation for this observation is the more extensive range of genetic, geographical, climatic, and cultural conditions represented in our data set relative to these earlier studies. Although the earlier studies were quite extensive, with a minimum of 42 site years, the geographical range of each study was limited to one state or province and at most three weather years. Our data, collected over seven states and 5 yr, probably represent a broader range of weather and cultural conditions, as well as genetic material, all of which may lead to greater variability in CM readings for N-sufficient corn across locations. Differences in CM readings due to genetics have been reported by several authors (Costa et al., 2001; Peng et al., 1993), and similar effects of weather and cultural practices could be reasonably postulated.
In our experiments, when N fertilizer had been applied at planting, absolute CM readings were only weakly related to EONR and to yield response to N through silking (Fig. 7). After silking, absolute readings were more strongly related to EONR and yield response to N, even when N had been applied at planting. Thus, absolute readings at this time may serve as a useful feedback mechanism for producers, as suggested by Piekielek et al. (1995). Figure 4 from Piekielek et al. (1995) is very similar to our Fig. 7C, and with a similar coefficient of determination.
Our findings support the need to use a high-N reference area to establish the CM reading associated with N-sufficient corn, given the genetics and environment of each field. In some cases, this will be a small area of the field to which a high rate of N has been applied, while the rest of the field has not yet been fertilized, or has had a moderate N rate applied. In fields where N fertilizer has been applied, but wet conditions have led to loss of N, it is likely that some areas of the field will have experienced less N loss and may be used as reference areas.
While reasonably good predictions of EONR and yield response to N can be made based on absolute CM readings from V10 onward where no N fertilizer has been previously applied, this is expected to be a rare situation. However, in cases where no N has been applied due to wet weather and no high-N reference area is available, absolute CM readings may be the best tool available in making a decision such as whether to apply N aerially.
Coefficients of determination for predictive models at different growth stages are summarized in Fig. 8 . Chlorophyll meter readings were, in general, related to yield response to N more strongly than they were related to EONR. Environmentally-driven variability in N-use efficiency would increase variability in EONR without directly affecting the potential for yield response to N. Figure 8 suggests that predictions of EONR and yield response to N will be more reliable:
|
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  S. M. Samborski, N. Tremblay, and E. Fallon Strategies to Make Use of Plant Sensors-Based Diagnostic Information for Nitrogen Recommendations Agron. J., July 7, 2009; 101(4): 800 - 816. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. C. Scharf and J. A. Lory Calibrating Reflectance Measurements to Predict Optimal Sidedress Nitrogen Rate for Corn Agron. J., May 8, 2009; 101(3): 615 - 625. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. A. Ruiz Diaz, J. A. Hawkins, J. E. Sawyer, and J. P. Lundvall Evaluation of In-Season Nitrogen Management Strategies for Corn Production Agron. J., November 7, 2008; 100(6): 1711 - 1719. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. A. Coulter and E. D. Nafziger Continuous Corn Response to Residue Management and Nitrogen Fertilization Agron. J., November 7, 2008; 100(6): 1774 - 1780. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. Ziadi, M. Brassard, G. Belanger, A. Claessens, N. Tremblay, A. N. Cambouris, M. C. Nolin, and L.-E. Parent Chlorophyll Measurements and Nitrogen Nutrition Index for the Evaluation of Corn Nitrogen Status Agron. J., August 11, 2008; 100(5): 1264 - 1273. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. L. Ping, R. B. Ferguson, and A. Dobermann Site-Specific Nitrogen and Plant Density Management in Irrigated Maize Agron. J., June 23, 2008; 100(4): 1193 - 1204. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Zhang, A. M. Blackmer, J. W. Ellsworth, and K. J. Koehler Sensitivity of Chlorophyll Meters for Diagnosing Nitrogen Deficiencies of Corn in Production Agriculture Agron. J., May 7, 2008; 100(3): 543 - 550. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Solari, J. Shanahan, R. Ferguson, J. Schepers, and A. Gitelson Active Sensor Reflectance Measurements of Corn Nitrogen Status and Yield Potential Agron. J., May 7, 2008; 100(3): 571 - 579. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Zhang, A. M. Blackmer, J. W. Ellsworth, P. M. Kyveryga, and T. M. Blackmer Luxury Production of Leaf Chlorophyll and Mid-Season Recovery from Nitrogen Deficiencies in Corn Agron. J., May 7, 2008; 100(3): 658 - 664. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. W. Simpson, A. N. Sharpley, R. W. Howarth, H. W. Paerl, and K. R. Mankin The New Gold Rush: Fueling Ethanol Production while Protecting Water Quality J. Environ. Qual., March 1, 2008; 37(2): 318 - 324. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. A. Hawkins, J. E. Sawyer, D. W. Barker, and J. P. Lundvall Using Relative Chlorophyll Meter Values to Determine Nitrogen Application Rates for Corn Agron. J., June 5, 2007; 99(4): 1034 - 1040. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Crop Science | 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 | |||
