Illinois
Fertilizer Conference Proceedings
January 23-25, 1995
| Home | 1995 Index | Search |
| Home | 1995 Index | Search |
D. Bullock, G. Bollero and D. Anderson1
Nitrogen is necessary for corn production and even moderate deficiencies will substantially reduce yield and profit, but excessive N can pollute both surface and ground water. Thus, farmers are walking a thin line. Too much N and they threaten the environment and add an unnecessary cost; too little N and they reduce their profit by reducing yield.
Virtually everyone is willing to decrease N fertilizer use if that reduction will improve N efficiency, increase profit and protect the environment. But current technology does not allow for rapid, accurate, on-farm decision making regarding N fertility. Thus, we can not answer the questions of if, when, where, and by how much N fertilizer rates can be changed without running the very real risk of affecting U.S. farm profit, efficiency, and competitiveness in the world market. For any given field the consequences of making the wrong decision are not trivial. Failure to reduce N fertilizer rates in many situations will result in farmers using more N than is necessary and in a continued decline in water quality. On the other hand, if we blindly accept large across-the-board reductions in N fertilizer rates we will, without a doubt, hurt many farmers and the U.S. position in world trade. To answer the questions of when, where and how much, we require a monitoring technique to evaluate the N status of corn. Ideally the technique will be fast and inexpensive and allow for on-the-spot decision making.
One tool has been suggested by Blackmer at Iowa State University. In their procedure the nitrate concentration in the top foot of soil is quantified when the plants are six inches tall. They have shown that most Iowa soils containing at least 21 ppm nitrate do not require additional N fertilizer. This research was the basis for the N test kit developed and offered by Iowa State University and the Hach Chemical Company during the 1990 growing season. The process is faster than sending soil samples to a lab, but it is still labor intensive and slow. It requires several hours to get a single result. The test kit has met with mixed, but overall positive, reviews by most growers.
Another possible tool is the Minolta SPAD-520 chlorophyll meter. The meter determines the relative amount of chlorophyll present in plant leaves. Chlorophyll amount in plant leaves is closely related to the N content of the leaves so if the relative chlorophyll level is known then the relative N content is known. Correlations of R=0.91 between the measured SPAD value and leaf N concentration of rice have been reported in Minolta literature. The SPAD-520 meter is very fast. A value can be obtained in less than one minute. The SPAD-520 meter is used currently in Asia for N management of rice.
While the SPAD-520 may be a valuable tool for on-farm decision making, several
critical questions remain however. First, we do not know if the SPAD-520 can
be used earlier in the season, which is essential if it is to be used for side
dress decisions. Second, we do not know how the meter will respond to different
environments and in particular different soil types. Third, we do not know if
the meter needs to be calibrated for individual hybrids. The objective of this
research was to evaluate the Minolta SPAD-520 chlorophyll meter as a rapid on-farm
decision making aid for N management in corn in Illinois.
The field component of this project was conducted on a cooperating farms in
Illinois and at the University of Illinois experimental farms at Urbana, Brownstown,
and Dekalb. The experiment was conducted for four years (1991-1994) and each
year we used only three of the above locations. This provided for a total of
twelve location-years which we have labeled as environments for this work. For
each environment the statistical design was a randomized complete block with
three replications. Individual plots were six rows wide by 50 feet long. At
each location ten different open pedigree hybrids were over planted and then
after emergence thinned to a final population of about 26,000 plants per acre.
Hybrids and environments were considered as random components for the analysis
of this work. Nitrogen rates were 0, 80, 160, and 240 lb N/acre and applied
as 28% solution injected into the soil at the first leaf stage. All sampling
was conducted on the middle two rows. SPAD meter readings and soil sampling
were conducted throughout the season. At each sampling time leaf disk punches
were taken in order to analyze for total N in the leaf tissue. Final yield was
obtained by harvesting the entire length of the middle two rows with a plot
combine and adjusting to 15.5% moisture.
The results for 1994 were substantially different from those obtained in the earlier years. Thus, for this report we will concentrate of the data pooled over environments. This is appropriate since environments were considered random and our desired inference space is all future years we might expect to experience in the state of Illinois. Spad readings and leaf punches were taken at various times during each season, but for the sake of space we will only discuss three sample times which were common to all environments. Those times are: seventh leaf stage (V7), midgreensilk (R1) and mid-grainfill (R3).
An abbreviated ANOVA and main effect mean values are shown in Table 1. With the exception of the H*N and E*H*N interactions, all main effects and other interactions are significant. Thus, for all measures hybrids responded to nitrogen fertilizer rate (i.e. H*N) and this relationship was not affected by environment (i.e. E*H*N). Hybrids did respond differently to environments (i.e. E*H) and the effect of nitrogen rate was affected by environment. These results are not unexpected and are typical of large data bases such as this.
As previously mentioned there are reports of very close correlations (R=0.91) between Spad readings and leaf N content for small data bases. However, in most cases these involve a limited number of cultivars or even a single cultivar within a single environment. We saw similar correlations for subsets of this data base. However, when examining large numbers of hybrids and environments we see the relationship is still highly significant, but not as strong. For example linear correlation between Spad readings and Kjeldahl N determination at V7, R1, and R3 (fig 1) shows a progressively better correlation. At V7 the Pearson correlation coefficient was only 0.34, but increased to 0.65 at R1 and 0.80 at R3. The latter values are very impressive relationship when considering that over 1400 observations were used in the calculation. Thus, this research shows that by mid-greensilk we could use the Spad meter on any hybrid an obtain a reasonably good estimate of leaf N content without having to resort to laboratory analysis.
The relationship between Spad meter readings and grain yield were also significant at each reading (fig 2) and showed a similar pattern of improved correlation with time. The Pearson correlation coefficient was only 0.14 at V7, but increased to 0.41 at R1 and a respectable 0.69 at R3. Thus, the Spad meter provides a reasonable method of estimating crop performance. The quality of the information is clearly not perfect, but examination of the relationships between leaf N content (the current technique) and yield shows it is better than the Kjeldahl analysis we are using now. The Pearson correlation coefficient for the relationship between leaf N content (i.e. Kjeldahl analysis) and yield was only 0.07 at V7, 0.35 at R1, and 0.51 at R3 (fig 3). All of these correlations are inferior to those for the Spad meter at the same time.
The question that beckons to be answered is how high should a Spad reading be at a given time? In this experiment regression analysis indicated linear rather than curvilinear responses. The greater the Spad reading at R1 and R3, in particular, the higher the grain yield at the end of the season. Since the responses were not curvilinear (i.e. they did not plateau) it is not possible to calculate or suggest an optimum Spad value. Rather, the linear relationship suggests a critical value which a field should be at or above. Examination of fig 2 (stages R1 and R3) suggests that a corn crop should have a Spad reading of at least 60 by R1 to ensure maximum yields.
It is worth noting that these correlations are not perfect. For any given hybrid
in any given environment a Spad reading may or may not be informative and thus
must be used with caution. However, in general the Spad meter does appear to
be a useful instrument. It does not seem reasonable to suggest that separate
critical numbers be calculated for individual hybrids. Note that the spread
for leaf N (fig 3) is even larger than
that for the Spad meter (figs 2), but
we do not suggest unique critical N contents for individual hybrids.
This field experiment indicates that the Minolta Spad meter is superior to Kjeldahl analysis as an indicator of grain yield and a reasonably good estimator of leaf N content. It is suggested that corn in Illinois should have a Spad reading of at least 60 by R1 to ensure maximum yields. If readings are less than 60 producers should examine their production system with particular attention to N nutrition.
Table 1: An abbreviated analysis of variance tablefor 12 location-years (environments)
1Professor and Research Associate, Department of Agronomy, University of Illinois, Urbana, IL
Breitenbeck, G. A., and J. M. Bremner. 1986. Effects of various nitrogen fertilizers on emission of nitrous oxide from soils. Biology and Fertility of Soils, 2:195-199.
Bremner, J. M., and A. M. Blackmer. 1978. Nitrous oxide: Emission from soils -during nitrification of fertilizer nitrogen. Science, 199:295-296.
Bremner, J. M., G. A. Breitenbeck, and A. M. Blackmer. 1981. Effect of anhydrous ammonia fertilization on emission of nitrous oxide from soils. Journal of Environmental Quality, 10:77-80.
Burford, J. R., and J. M. Bremner. 1975. Relationships between the denitrification capacities of soils and total, water-soluble and readily decomposable soil organic matter. Soil Biology & Biochemistry, 7:389-394.
Eno, C. F., and W. G. Blue. 1957. The comparative rate of nitrification of anhydrous ammonia, urea, and ammonium sulfate in sandy soils. Soil Science Society of America Proceedings, 21:392-396.
Firestone, M. K. 1982. Biological denitrification. In: Nitrogen in agricultural soils (F. J. Stevenson et al., ed.). Agronomy Monograph 22. American Society of Agronomy, Madison, Wisconsin. pp. 289-326.
McInnes, K. J., and I. R. P. Fillery. 1989. Modeling and field measurements of the effect of nitrogen source on nitrification. Soil Science Society of America Journal, 53:1264-1269.
Mulvaney, R. L. 1994. Nitrification of different nitrogen fertilizers. In: 1994 Illinois Fertilizer Conference Proceedings (R. G. Hoeft, ed.). pp. 83-96.
Mulvaney, R. L., and C. W. Boast. 1986. Equations for determination of nitrogen-15 labeled dinitrogen and nitrous oxide by mass spectrometry. Soil Science Society of America Journal, 50:360-363.
Mulvaney, R. L., and L. T. Kurtz. 1982. A new method for determination of 15N-labeled nitrous oxide. Soil Science Society of America Journal, 46:1178-1184.
Norman, R. J., and L. T. Kurtz. 1986. Laboratory apparatus for preparing and dispensing nitrogen-15-labeled anhydrous ammonia. Soil Science Society of America Journal, 50:238242.
Norman, R. J., L. T. Kurtz, and F. J. Stevenson. 1987. Solubilization of soil organic matter by liquid anhydrous ammonia. Soil Science Society ofAmerica Journal, 51:809-812.
Ritchie, G. A. F., and D. J. D. Nicholas. 1972. Identification of the sources of nitrous oxide produced by oxidative and reductive processes in Nitrosomonas europaea. Biochemistry Journal, 126:1181-1191.
Stevenson, F. J. 1994. Humus chemistry. Genesis, composition, reactions. 2nd ed. John Wiley and Sons, New York.
