Agronomy Journal Grow Your Career With ASA
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published in Agron J 97:1603-1611 (2005)
DOI: 10.2134/agronj2005.0129
© 2005 American Society of Agronomy
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Web of Science (3)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Martin, K. L.
Right arrow Articles by Raun, W. R.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Martin, K. L.
Right arrow Articles by Raun, W. R.
Agricola
Right arrow Articles by Martin, K. L.
Right arrow Articles by Raun, W. R.
Related Collections
Right arrow Maize
Right arrow Plant Nutrition
Right arrow Spatial Distribution

Corn

Plant-to-Plant Variability in Corn Production

K. L. Martina, P. J. Hodgenb, K. W. Freemana, Ricardo Melchioric, D. B. Arnalla, R. K. Teala, R. W. Mullend, K. Destaa, S. B. Phillipse, J. B. Soliea, M. L. Stonea, Octavio Cavigliac, Fernando Solarib, Agustin Bianchinif, D. D. Francisb, J. S. Schepersb, J. L. Hatfieldg and W. R. Rauna,*

a Oklahoma State Univ., Stillwater, OK
b USDA-ARS, Lincoln, NE
c INTA, Parana, Argentina
d The Ohio State Univ., Columbus, OH
e Virginia Polytechnic Inst., Blacksburg, VA
f AAPRESID, Rosario, Argentina
g Usda-Ars, Ames, Ia

* Corresponding author (wrr{at}mail.pss.okstate.edu)

Received for publication May 3, 2005.

   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Corn (Zea mays L.) grain yields are known to vary from plant to plant, but the extent of this variability across a range of environments has not been evaluated. This study was initiated to evaluate by-plant corn grain yield variability over a range of production environments and to establish the relationships among mean grain yield, standard deviation, coefficient of variation, and yield range. A total of forty-six 8- to 30-m corn transects were harvested by plant in Argentina, Mexico, Iowa, Nebraska, Ohio, Virginia, and Oklahoma from 2002 to 2004. By-plant corn grain yields were determined, and the average individual plant yields were calculated. Over all sites in all countries and states, plant-to-plant variation in corn grain yield averaged 2765 kg ha–1 (44.1 bu ac–1). At the sites with the highest average corn grain yield (11478 and 14383 kg ha–1, Parana Argentina, and Phillips, NE), average plant-to-plant variation in yield was 4211 kg ha–1 (67 bu ac–1) and 2926 kg ha–1 (47 bu ac–1), respectively. As average grain yields increased, so did the standard deviation of the yields obtained within each row. Furthermore, the yield range (maximum corn grain yield minus the minimum corn grain yield per row) was found to increase with increasing yield level. Regardless of yield level, plant-to-plant variability in corn grain yield can be expected and averaged more than 2765 kg ha–1 over sites and years. Averaging yield over distances >0.5 m removed the extreme by-plant variability, and thus, the scale for treating other factors affecting yield should be less than 0.5 m. Methods that homogenize corn plant stands and emergence may decrease plant-to-plant variation and could lead to increased grain yields.

Abbreviations: CV, coefficient of variation • NDVI, Normalized Difference Vegetative Index = ({rho}NIR – {rho}Red)/({rho}NIR + {rho}Red) • PSV, plant-spacing variability


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
VARIABILITY IN PLANT stands is well documented. Nielsen (2001) studied plant-spacing variability (PSV) in 354 commercial fields of corn throughout Indiana and Ohio. This work showed that the standard deviation of plant spacing was 7.5 cm or less in only 16% of the fields. Sixty percent of the sampled fields exhibited standard deviations of plant spacing between 10 and 12.5 cm. Plant-spacing standard deviations in 24% of the fields were 15 cm or greater (up to 30 cm). Their results showed that for every 2.54 cm increase in the standard deviation in plant-to-plant spacing, 156 kg ha–1 (2.5 bu ac–1) in grain yield was lost. The average standard deviation of plant spacing was 17.2 cm (6.8 in), resulting in an estimated 1066 kg ha–1 (17 bu ac–1) yield loss over 354 commercial fields. Lauer and Rankin (2004) and Liu et al. (2004) had differing results, noting that PSV did not significantly alter grain yields in Wisconsin and Ontario, Canada, respectively. Nafziger et al. (1991) noted that uneven emergence of corn can occur when soils are dry at the time of planting and could lead to decreased grain yields. It is generally accepted that when adjacent plants differ by more than two leaf stages, the younger plant may not develop to its fullest potential. A two leaf stage difference can result from delayed emergence ranging from 5 to 10 d, which can cause a 1% yield loss for each 1-d delay (Robert L. Nielsen, Purdue University, personal communication, 2004). Tollenaar and Wu (1999) found increased stress tolerance in corn when plant-to-plant variability was lower. In general, these statistics identify a twofold problem: first, the need to homogenize plant spacing and emergence and second, the need to recognize differences in yield potential that clearly exist by plant.

Some technologies in precision agriculture have been driven commercially. The most notable has been combine yield monitors. Depending on combine speed, header width, and the smoothing effect as grain moves through the combine, each sensed element represents more than 48 m2 (width of swath times the distance traveled in 2 to 4 s). However, Lengnick (1997), Solie et al. (1999), and Raun et al. (1998) found significant soil variability at distances less than 30 m apart, and in many cases, less than 1 m. Furthermore, large differences in measured yield have been reported on a small scale (<0.4 m2) for winter wheat (Triticum aestivum L.) (Raun et al., 2002) and by plant in corn (Raun et al., 2005). For corn, the expressed spatial variability was greatest at the V6 growth stage (Ritchie et al., 1996). This peak in the within-row variability was thought to occur at the same growth stage where treating the variability would have the greatest impact (Raun et al., 2005). Maddonni and Otegui (2004) reported that the greatest difference in estimated shoot biomass between plant types occurred between V7 and V13 (Ritchie et al., 1996) and remained constant from V13 onward. Vega and Sadras (2003) found a strong inequality in reproductive output within high populations of corn, indicating an apparent breakage of reproductive allometry.

Varvel et al. (1997) noted that when sufficiency indices (determined with a SPAD meter) were lower than 90% at V8, maximum yields were not achieved with in-season N fertilizer applications because early-season available N was below that needed for optimum growth and yield potentials had already been reduced. Therefore, if added N was needed, making the decision to apply should likely take place at or before V8.

Fundamental field element size is the area where maximum relatedness exists between adjacent elements. Treatment at scales larger than the fundamental field element size compromises the effectiveness since independent variation of nutrient levels exists within a single treatment level. Treatment at scales less than the fundamental field element size is pointless, as nutrient levels within this scale are related. When N decisions are at 1-m2 resolution, the variability present can be detected (e.g., Normalized Difference Vegetative Index, or NDVI) and treated accordingly with foliar N (Solie et al., 1996; Stone et al., 1996), thus increasing N use efficiency. Taylor et al. (1999) reported that smaller plot sizes employed in variety trials reduced the variability encountered in estimating the mean yield. This was consistent with the resolution where detectable differences in soil test parameters exist that should be treated independently.

Porter et al. (1998) observed that temporal yield variability was approximately three times greater for soybean [Glycine max (L.) (Merr.)] and four times greater for corn than spatial variability among plots. They also reported that producers should not change management practices (as a function of yield maps) unless the differences were shown to be consistent over years. Mallarino et al. (1999) employed grid sampling and factor analysis to investigate the relationship between several site variables (soil tests, plant population, weed control, etc.) and corn grain yields in five producer fields. They reported that some of the variables collected were correlated with grain yields, but that the relationships changed between fields. When collecting corn grain yield data from twenty-four 4.6- by 3.0-m subplots within a larger farmer field, Schmidt et al. (2002) showed that yields ranged from 4.7 to 9.5 Mg ha–1. It is important to note that this large range in yield was from plots that did not receive any fertilizer N. They also noted that variable N application is needed to achieve maximum grain yield and improved N management over different locations in the same field. Sadler et al. (1998) noted that Coastal Plain soils required study at finer resolutions than the >100-m grids commonly used in precision farming.

The objectives of this study were (i) to evaluate by-plant corn grain yield variability over a range of production environments; (ii) to determine the relationships among mean grain yield, standard deviation of yield, coefficient of variation (CV) of yield, and yield range; and (iii) to evaluate the relationship between NDVI and corn grain yields.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 REFERENCES
 
In accordance with the countries and states where data were collected for this paper, the following production statistics are provided. In 2003, world corn grain production averaged 4.5 Mg ha–1, coming from 142 million ha. Average corn grain yields in the USA, Argentina, and Mexico were 8.9, 6.4, and 2.5 Mg ha–1 from 28, 2.3, and 7.8 million ha, respectively (http://faostat.fao.org; verified 19 Aug. 2005). In Iowa, Nebraska, Ohio, Virginia, and Oklahoma, average corn grain yield for 2003 was 9.8, 9.2, 8.7, 7.2, and 7.8 Mg ha–1 from 4.8, 3.1, 1.2, 0.13, and 0.08 million ha, respectively (www.usda.gov/nass/nasshome.htm; verified 19 Aug. 2005).

By-plant harvested corn grain yields from 13 different sites in Argentina, Mexico, Iowa, Nebraska, Ohio, Virginia, and Oklahoma were evaluated to determine relationships among by-plant and averaged yield and ranges, standard deviations, and coefficients of variation of yields. At each location, corn rows (transects) ranging from 8 to 30 m in length were selected for by-plant harvesting. At most of the sites, individual plants were marked at or before the V8 growth stage to ensure detection of barren and/or lost plants at harvest (60–85 d later depending on the maturity). At the same time that plants were tagged, a tape measure was extended the length of the row, and cumulative distances were recorded for each plant.

At most sites, based on the row spacing used at each location, the area occupied by each plant was calculated. This was done by assuming that each plant occupied half the distance to and from its nearest neighbor (Eq. [1]):

[1]
where Ai is the area occupied by the ith plant; di–1, di, and di+1 are the distances to the i – 1, i, and i + 1 plants; and R is the row spacing.

Each ear was harvested individually from each plant and the weight recorded. When more than one ear per plant was present, total weight was recorded on a by-plant basis. At those sites where actual distances between plants were not recorded, an average distance occupied per plant was determined based on row spacing and total transect or row distance and number of plants harvested per row. Once removed from the stalk, ears were dried at 66°C for 48 h and weighed before and after shelling. The dry grain weight taken from the shelled corn was the final weight used for yield determination. The locations sampled, number of transects at each site, planting date, harvest date, row spacing, plant population, hybrid, and maturity are reported in Table 1. Location, soil series, texture, and transect length are included in Table 2. At some of the sites, NDVI data were collected at the V8 growth stage, directly over the corn row using a shaft encoding GreenSeeker sensor placed 76 cm above the corn canopy and that was capable of recording NDVI readings with computed distance every 0.5 cm. Sensor NDVI readings were then averaged for consecutive four-plant clusters. Corn ears from these four-plant clusters were harvested and shelled to evaluate the relationship between NDVI readings at V8 and final grain yield.


View this table:
[in this window]
[in a new window]
  		Table 1. Location, transect, year, planting date, harvest date, row spacing, plant population, and corn hybrids where corn grain yields were evaluated by plant from 2002 to 2004 from transects ranging from 8 to 30 m of row, USA, Argentina, and Mexico.

 

View this table:
[in this window]
[in a new window]
  		Table 2. Soil series and texture from location transects ranging from 8 to 30 m of row, USA, Argentina, and Mexico.

 
Statistical analysis included regression of average grain yield per transect on the standard deviation, CV, and yield range of by-plant grain yield over all locations using SAS (SAS Inst., 2002). The areas where by-plant harvest data were collected were representative of a range of corn production environments around the world. The previous crop and tillage practice employed at each site are reported in Table 2.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 REFERENCES
 
Average transect corn grain yield plotted against standard deviation, CV, and yield range over all locations is illustrated in Fig. 1, 2, and 3 , respectively. Prediction equations reported and plotted on Fig. 1–3 do not exceed the limits of the collected data. The standard deviation of by-plant corn grain yield increased with increasing yield level up to 13000 kg ha–1 (Fig. 1). This is consistent with several sources reporting that the standard deviation of yields increases with increasing yield level (Taylor et al., 1999; Dobermann et al., 2003). The CV of by-plant yields was negatively correlated with mean grain yield across the range of experiments studied (Fig. 2). The test for the slope (negative) being different from zero was highly significant (P > |t|, 0.0007). Even though CVs were lower for the higher-yielding sites, the actual plant-to-plant variation in grain yield (kg ha–1) was greater when compared with sites with lower average yields. The yield range (maximum yield observed in each transect minus the minimum yield observed) increased with average corn grain yields (Fig. 3).



View larger version (21K):
[in this window]
[in a new window]
  		Fig. 1. Average corn grain yield plotted against the standard deviation from by-plant yield over 46 transects in Argentina, Mexico, Iowa, Nebraska, Ohio, Virginia, and Oklahoma.

 


View larger version (18K):
[in this window]
[in a new window]
  		Fig. 2. Average corn grain yield plotted against the coefficient of variation from by-plant yields over 46 transects in Argentina, Mexico, Iowa, Nebraska, Ohio, Virginia, and Oklahoma.

 


View larger version (20K):
[in this window]
[in a new window]
  		Fig. 3. Average corn grain yield plotted against the by-plant yield range (maximum minus minimum yield) in 46 transects ranging from 10.5 to 30 m in length in Argentina, Mexico, Iowa, Nebraska, Ohio, Virginia, and Oklahoma.

 
Average grain yield across all regions ranged from 4268 to 14383 kg ha–1, with an average of 8495 kg ha–1 (135 bu ac–1), close to the U.S. average and above that for Argentina and Mexico (Table 3). The average differences in measured yield plant to plant ranged from 1724 to 4367 kg ha–1 (excluding barren plants) and averaged 2765 kg ha–1 (44.1 bu ac–1) (Table 3). At those sites where the average yields were the highest (Phillips, NE, and Argentina), the standard deviations about the yield mean were 2926 kg ha–1 (47 bu ac–1) and 4211 kg ha–1 (67 bu ac–1), respectively. Although a trend for decreased CVs at the higher yield levels was observed (Fig. 2), the average plant-to-plant yield differences that would be encountered at both these high-yielding sites exceeded the average over all locations where yields were much lower.


View this table:
[in this window]
[in a new window]
  		Table 3. Minimum, maximum, mean, standard deviation, maximum/minimum, and coefficient of variation (CV) for by-plant corn grain yields from 46 transects in Argentina, Mexico, Iowa, Nebraska, Ohio, Virginia, and Oklahoma, 2002–2004.

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
REFERENCES
 
The sites reported in this paper were planted and treated using normal practices in each region. No steps were taken to minimize cultural, nutrient, or environmental factors that would keep the corn hybrids from achieving their genetic yield potential. To achieve the theoretical maximum genetic yield potential, stands must be optimum, plant spacing must be exact, seed must be planted at ideal and uniform depth, germination should be 100%, all nutrients must be nonlimiting, soil types and physical properties must be uniform and ideal for the cultivar, and moisture, temperature, and all other environmental factors must be ideal during the entire growing season. All plants must emerge within 1 d. All plants must set at least one ear of corn, and the ear must completely fill. Under these conditions, the range of yield and standard deviation of yield should theoretically approach zero.

Causes for the Large Differences in By-Plant Corn Grain Yields
There are many variables that likely contributed to the extensive variability in by-plant corn grain yield seen at all sites included in this study. Delayed and uneven emergence can be caused by variable depth of planting, wheel compaction, location of the seed within the furrow, surface crusting, random soil clods, soil texture differences, variable distance between seeds, seed germination, variable soil compaction around the seed, insect damage, moisture availability, variable surface residue, variable seed furrow closure, and/or many other factors that influence nonuniformity of plants. In light of the many factors known to influence plant stands, within-row variability in corn grain yield should be expected, and that was present in the trials evaluated here. In all trials, common hybrids were employed for each respective region. Each row was inspected early in the season (excluding the Phillips, NE location) for volunteer corn, and these plants were removed. The presence of volunteer corn plants was scarce, especially in Virginia, Argentina, Mexico, Ohio, and some sites in Oklahoma where corn was not the previous crop (Table 2).

The range of average yields (2700–16100 kg ha–1) included in this study was representative of a wide array of production environments (Fig. 1–3). Some of these sites were irrigated while others relied on natural precipitation. One of the sources of plant-to-plant variability could be competition for soil moisture, especially in dryland fields. However, this would be an unlikely source of plant-to-plant variability at the higher yield levels where moisture was not limiting (>13000 kg ha–1), unless soil texture differences were expected at the by-plant level. Similarly, the extensive plant to-plant differences in grain yield within 8 to 30 m of row were unlikely due to plant-to-plant differences in early-season N availability.

Maddonni and Otegui (2004) noted that increased interplant competition in corn hybrids enhanced the appearance of plants with different competitive abilities. Thus at higher populations, plant-to-plant variation can also be expected. They noted that the onset of interplant competition started very early during the life cycle and that differences in estimated plant biomass between stand densities were detected as early as V6. Furthermore, they reported that plant population and row-spacing treatments alone did not modify the onset of the hierarchical growth among plants. The same causes for delayed and uneven emergence discussed earlier in this section would likely be expressed later in the life cycle with variable plant growth.

Expression of Variability
Seed suppliers do not normally publish genetic yield potential data (Fig. 1–3). However, the National Corn Growers Association Corn Yield Contest results (www.ncga.com/02profits/CYC/winners/winners.html; verified 19 Aug. 2005) can serve as a surrogate for these data. To achieve maximum yields, contest participants attempt to manage all factors under their control to minimize reduction in corn yield from the cultivars' genetic yield potential. The highest first-place yields for all classes from 2002, 2003, and 2004 ranged from 19000 to 22000 kg ha–1.

Nonlinear equations were fitted to the data using Table Curve 2D (Systat Software, 2000). The equations with the highest R2, which conformed to the upper and lower boundary conditions, were selected. In all cases, a partial cubic polynomial model met these requirements, and that was fit using a zero intercept. Examination of Fig. 1 through 3 yielded the following additional observations on the relationship of average corn yield to CV, range, and standard deviation. The range and standard deviation curves peaked near 13000 kg ha–1. The corn by-plant yield CV was nearly constant for yields under 10000 kg ha–1 and declined moderately to 15000 kg ha–1. Average field scale corn yields in all areas reported in this paper were generally much lower than 15000 kg ha–1. The by-plant corn yield variability was large within the yield ranges achieved by producers, as indicated by the state and country averages cited previously. Because the overall plant-to-plant variation in yield was found to be 2765 kg ha–1 (44.1 bu ac–1) (Table 3), it will likely be important to recognize and treat these differences. If it is feasible to recognize 2765 kg ha–1 plant-to-plant yield differences when average yields are 4300 kg ha–1, it should be feasible to detect them when average yields are 14000 kg ha–1.

Errors Associated with By-Plant Yield
Using the average plant yield of 120 g (dry shelled weight) over all trials included in this study and randomly applying all the errors included in the estimate (scale precision of 0.05 g, by-plant tape measure precision of 1.0 cm, and row spacing error of 1 cm), the yield estimate was 645 g m–2 or 6453 kg ha–1, which would be off by 5.8% when compared with the true value with no errors (6073 kg ha–1) using a 26-cm distance between plants and a 76-cm row spacing. Similarly, a 5.2% error was found when estimating yield from larger plots (harvesting two rows, 13.6 m in length) using a field plot combine [plot weight of 10.5 ± 0.2 kg (on-board scale precision) and a row distance of 13.6 ± 0.2 m, row spacing of 76 ± 1 cm]. This suggests that estimates of by-plant yields are no more problematic than small plot work using all the respective errors.

Work by Taylor et al. (1999) showed that standard deviations about yield means increased as mean yields increased in 220 fertilizer, weed management, and tillage trials, and that was similarly encountered in the trials reported here. Also, Taylor et al. (1999) reported a decrease in yield CV when mean yields increased. However, unlike the work of Taylor et al. (1999), which focused on plot data, we report on the standard deviations associated with by-plant differences in measured grain yield.

The average maximum/minimum range observed was 11x (46 transects ranging from 8 to 39 m of row) (Table 3). This came from experiments with an average yield of 8495 kg ha–1, well above the world average of 4500 kg ha–1 reported for corn grain yield in 2003. Furthermore, the data collected at specific sites within each location (Argentina, Mexico, Iowa, Nebraska, Ohio, Virginia, and Oklahoma) had yields equal to or exceeding each specific region's average. Dobermann et al. (2003) reported corn grain yields from 4- by 4-m grids determined from yield monitor data from 1996 to 2001. In this work, the maximum/minimum yields observed in the entire field exceeded 20x.

Corn Grain Yields Averaged over Larger Scales
If it were not possible to recognize each plant individually using sensors as has been published, it would be considered important to evaluate the error in predicting by-plant yields when yields were averaged over different scales, either by more than one plant or by a fixed distance. To investigate this, by-plant grain yields were averaged over two, three, and four plants in 15 m of row from the Ames, IA location (Fig. 4) . The largest differences in corn yield were discernable at the by-plant level. However, 4.9, 4.3, and 2.6x differences were detected within 15 m of row when averaged over two, three, and four plants, respectively (Fig. 4). At this site, the average differences in yield (standard deviation) were 3607, 2678, 2358, and 1996 kg ha–1 when yields were averaged over one, two, three, and four plants. When averaged over every four plants, this scale resulted in average differences between four-plant clusters that exceeded 1900 kg ha–1 (30 bu ac–1). In addition, when evaluating every four-plant cluster in 15 m of row, the range in yields was 5014 to 13210 kg ha–1 (difference of 8196 kg ha–1 or 131 bu ac–1).



View larger version (36K):
[in this window]
[in a new window]
  		Fig. 4. Average corn grain yields plotted by plant, every two plants, every three plants, and every four plants, using measured distances between plants at Ames, IA in 2004.

 
For the data in Fig. 4, grain yields were determined based on actual measured distances between plants and the area calculated from Eq. [1], which resulted in 10.6x differences in corn grain yield over 15 m of row. However, it is important to note that when by-plant grain yields were computed based on a fixed area (average distance between plants, 23.5 cm, over the entire row), a 12.5x difference in yield differences was observed at the Ames, IA site (Fig. 5) . At all sites, large differences in corn grain yield were observed over short distances whether or not grain yields were computed based on actual measured distances or an average (fixed) distance between plants.



View larger version (40K):
[in this window]
[in a new window]
  		Fig. 5. Average corn grain yields computed using fixed distances of 23.5, 47.0, 70.6, and 94.1 cm at Ames, IA in 2004.

 
At the Ames, IA site, midseason prediction of corn grain yields (same four-plant clusters) was quite good using NDVI collected at the V8 growth stage (Fig. 6) . The average yield difference of 1996 kg ha–1 between each four-plant cluster was greater than the precision at which final grain yields could be predicted midseason using NDVI (precision of ± 1565 kg ha–1).



View larger version (17K):
[in this window]
[in a new window]
  		Fig. 6. Normalized Difference Vegetative Index (NDVI) versus corn grain yield determined for every four plants using linear regression and associated 95% confidence intervals, east row at Ames, IA in 2004.

 
Errors in Corn Grain Yields from Larger Scales
Yet another alternative approach concerning this data was evaluating the errors associated with yield determined over fixed distances in a row (EFAW experiment station near Stillwater, OK; Shelton, NE; and Ames, IA). The distances used to calculate the yields were determined by dividing the total length of the row by the number of plants in the row. The yields were then averaged over multiples of that distance until the distance approached 1 m. Then, the absolute value of the errors in estimating the by-plant corn yield from the average value of the yields along fixed distances were calculated. Table 4 shows the mean, minimum, and maximum error associated with the fixed-distance yield calculation. At each location, the errors in the by-plant yield prediction decreased as yields were averaged over greater distances, with errors approaching a constant as distances approached 1 m (Fig. 7) . The distance where the true yield mean of the row could be estimated was between 0.5 to 0.6 m. All sites behaved similarly, with similar normalized errors in the by-plant yield estimates. Even though there are errors associated with predicting yield at early growth stages (V8, Fig. 6), these errors in yield prediction were dwarfed in comparison to the by-plant yield differences reported.


View this table:
[in this window]
[in a new window]
  		Table 4. Absolute value of the errors in estimating by-plant corn yield by averaging yield over a fixed distance along the row.

 


View larger version (16K):
[in this window]
[in a new window]
  		Fig. 7. Effect of averaging plant yields over a specified distance along the row on the absolute error incurred when using the average corn yield for estimating the by-plant yield. Yields were normalized by the average yield along the entire row.

 

   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 



This article has been cited by other articles:


Home page
 Agron. J.Home page
P. J. Hodgen, R. B. Ferguson, J. F. Shanahan, and J. S. Schepers
Uptake of Point Source Depleted 15N Fertilizer by Neighboring Corn Plants
Agron. J., January 8, 2009; 101(1): 99 - 105.
[Abstract] [Full Text] [PDF]

Home page
 Agron. J.Home page
K. W. Freeman, K. Girma, D. B. Arnall, R. W. Mullen, K. L. Martin, R. K. Teal, and W. R. Raun
By-Plant Prediction of Corn Forage Biomass and Nitrogen Uptake at Various Growth Stages Using Remote Sensing and Plant Height
Agron. J., March 12, 2007; 99(2): 530 - 536.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Web of Science (3)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Martin, K. L.
Right arrow Articles by Raun, W. R.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Martin, K. L.
Right arrow Articles by Raun, W. R.
Agricola
Right arrow Articles by Martin, K. L.
Right arrow Articles by Raun, W. R.
Related Collections
Right arrow Maize
Right arrow Plant Nutrition
Right arrow Spatial Distribution

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