HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Figures Only Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in ISI Web of Science Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (8) Citing Articles via Google Scholar Google Scholar Articles by Elings, A. Search for Related Content PubMed Articles by Elings, A. Agricola Articles by Elings, A. Related Collections Crop Growth and Development Maize Plant Analysis

CORN

Estimation of Leaf Area in Tropical Maize

The International Center for Maize and Wheat Improvement (CIMMYT), Lisboa 27, Apdo. Postal 6-641, 06600 Mexico, D.F., Mexico

a.elings{at}plant.wag-ur.nl

	ABSTRACT

TOP ABSTRACT INTRODUCTION Materials and methods Results Discussion Conclusions REFERENCES

 
Leaf area development at regular time intervals during the growing season is estimated by many crop growth models and is needed for competition studies. In plant breeding programs, observation is generally restricted to a single leaf area estimate or assessment of the plant type. Two procedures that build on earlier studies are presented to estimate total plant leaf area of maize (Zea mays L.). Leaf area development of six tropical maize cultivars grown in 1995 and 1996 in several tropical environments in Mexico (both favorable and moisture- and N-limited) was observed and analyzed. First, the validity of a bell-shaped curve describing the area of individual leaves as a function of leaf number was investigated. When individual cultivar–environment combinations were normalized for area of the largest leaf and for total leaf number, one parameter set described all combinations. It remained difficult, however, to estimate these parameters in advance, which limits predictive applications in crop growth models. Analytical application after flowering, when parameter values can be determined, is possible. Second, a method was developed to directly estimate total leaf area when total leaf number and area of the largest leaf are known. The method makes use of the facts that the area of the largest leaf relative to total plant leaf area is constant and that this constant is linearly related to total leaf number. This study has shown that approaches previously presented by others are applicable in modified form over a wide range of environmental conditions.

Abbreviations: BA, El Batán • LAI, leaf area index • LAI_g, green leaf area index • LAI_tot, total leaf area index • LAI_{tot, max}, maximum total leaf area index • PR, Poza Rica • TL, Tlaltizapán

	INTRODUCTION

TOP ABSTRACT INTRODUCTION Materials and methods Results Discussion Conclusions REFERENCES

 
LEAF AREA INDEX of a crop was defined by Watson (1947) as the one-sided area of green leaf tissue per unit area of land occupied by that crop. The capacity of the crop to intercept photosynthetically active radiation and synthesize carbohydrates for growth is a nonlinear function of LAI. Here, this measure is called green leaf area index (LAI_g), as distinguished from total leaf area index (LAI_tot), which includes green and senesced leaf area. Accurate estimation of LAI_tot and LAI_g is essential to many crop growth and crop competition studies and simulation models, and LAI_tot, LAI_g, and leaf weight are possible components of selection indices in crop improvement.

Dwyer and Stewart (1986) introduced a slightly skewed bell-shaped function to describe the relationship between leaf number and the area of mature leaves. Integration of individual leaf areas gives total plant leaf area. Muchow and Carberry (1989), Keating and Wafula (1992), and Birch et al. (1998) have shown that this bell-shaped function is superior to other equations that relate LAI_tot to leaf number. They also showed that the parameters of this function can be related to the total number of leaves, which can be estimated from thermal time from emergence to tassel initiation, well before LAI_tot is reached (Kiniry, 1991). Tassel initiation occurs at approximately the 7-to-12-visible-leaves stage, depending on the relative maturities of the varieties and the photoperiods to which they are exposed (G.O. Edmeades, personal communication, 1998). However, all mentioned studies were conducted under nonlimiting growth conditions. Furthermore, Birch et al. (1998), working with five cultivars, observed genetic variation in parameter values. The question remains, therefore, whether relationships between LAI_tot and the parameter set describing this bell-shaped curve are valid for wider sets of germplasm and growing conditions.

Mathematical relationships between length, width, and area of maize leaf blades can serve as a basis for direct leaf area estimation. Montgomery (1911) found that the area of a maize leaf blade can be estimated as its length multiplied by its maximum width multiplied by 0.75. In a modification of this method, Lal and Subba Rao (1951) proposed a logarithmic transformation of leaf length and width. Total plant leaf area can therefore be obtained by measuring the lengths and widths of all leaves of a representative sample of leaves and plants (Stewart and Dwyer, 1999). As these are laborious methods, McKee (1964) also proposed that total leaf area may be estimated with less precision by multiplying the sum of the lengths of all leaves on a plant by 6.67, which would exclude measurement of leaf widths. Francis et al. (1969), on the basis of seven single crosses and their 14 parent inbred lines from the USA, suggested multiplying the area of the seventh leaf from the top by a leaf area factor of 9.04 to obtain total plant leaf area. They noted, however, that this coefficient varied widely for inbred populations grown at different densities, and that such cases would require separate leaf area factors. In another study, Pearce et al. (1975) proposed multiplying the area of the eighth leaf from the top by 9.39. Methods based on such mathematical relations are not time-consuming if the number of cultivars and plots involved is limited. However, coefficients have been developed only for temperate Corn Belt cultivars growing under favorable conditions (Fakorede et al., 1977), and not for tropical or subtropical cultivars growing under tropical or subtropical conditions.

The first objective of this study was to determine whether a relationship based on the bell-shaped curve proposed by Dwyer and Stewart (1986) would be valid for diverse sets of tropical germplasm and growing conditions. A second objective was to develop an alternative, widely valid method to directly estimate, from length and width of a particular leaf, total leaf area of tropical lowland maize cultivars differing in leaf number and maturity, and growing under optimal and suboptimal conditions. For that purpose, leaf area development of a set of five tropical maize cultivars grown under a variety of tropical growing conditions was observed and analyzed.

	Materials and methods

TOP ABSTRACT INTRODUCTION Materials and methods Results Discussion Conclusions REFERENCES

 
Experimental Details
Five maize cultivars adapted to the lowland tropics, varying in maturity and therefore leaf number, were evaluated in several environments: cv. Pool 16 C₂₀ ("C" stands for cycle of selection), cv. PR 8330, cv. Across 8328 BN C₆, cv. La Posta Sequía C₄, and cv. CML247 x CML254. The tropical highland hybrid cv. CML246 x CML242 was also evaluated in one experiment (Table 1) .

The cultivars were evaluated in PR at low, medium, and high levels of soil N availability (Table 2) , and P was applied at 60 kg P₂O₅ ha^-1 in all experiments. The low and medium N treatments were conducted on continuously cropped fields that had not received chemical N for 9 and 6 yr, respectively, whereas the high N treatment received 75 kg N ha^-1 at sowing and 125 kg N ha^-1 one month after sowing. The 1995 experiment at TL received 150 kg N ha^-1 at sowing and 50 kg N ha^-1 at anthesis, resulting in a medium level of N availability (Table 2). Different levels of soil N availability resulted in relatively low, medium, and high levels of leaf N concentration (Elings et al., 1997). The cultivars were evaluated at TL in the dry season of 1996 at low, intermediate, and adequate levels of water availability during flowering and grain filling. Water was applied by furrow irrigation. Duration of irrigation and fraction of irrigated furrows was based on long-term experience regarding the build-up of soil moisture deficit. For the low moisture level, irrigation was halted about 4 wk before anthesis and resumed midway through grain filling; for the medium moisture level, irrigation was halted about 1 wk before anthesis and not resumed; and for the adequate moisture level, irrigation was applied as needed (about every 10 d) to ensure near-optimal growth, based on long-term field experience. Sowing dates of cultivars were varied in an attempt to make flowering dates of all cultivars coincide in time and with a given level of soil moisture deficit that was created for the entire experiment. Those three experiments received 150 kg N ha^-1 at sowing. Evaluations at BA were conducted at adequate levels of water and N availability (175 and 50 kg N ha^-1 at sowing and one month after sowing, respectively). All N was applied as urea.

The area of each individual leaf blade (Y, in square centimeters) was computed as length x maximum width x 0.75. The value of 0.75 was considered an acceptable average of reported values: 0.72 (Keating and Wafula, 1992), 0.73 (McKee, 1964; Dwyer and Stewart, 1986), 0.74 (Stewart and Dwyer, 1999), 0.75 (Montgomery, 1911), and 0.79 (Birch et al., 1998). Plant leaf area (Y_p, in square centimeters) was computed as the sum of the areas of a plant's individual leaf blades.

Bell-Shaped Function
Dwyer and Stewart (1986) introduced a slightly skewed bell-shaped function to describe the relationship between leaf number and the area of mature leaves:

(1a)

where Y is the mature leaf area of individual leaves, Y₀ is the mature leaf area of the largest leaf, X is the leaf number counted from the bottom of the plant, X₀ is the leaf number of the largest leaf and the point of inflection of the curve, and a and b are empirical constants. Parameter a quantifies the kurtosis of the curve, with low values of a resulting in sharply rising and falling curves. Parameter b controls the degree of skewness of the curve, with negative and positive values of b resulting in curves that are skewed towards the left and right, respectively (Keating and Wafula, 1992). As Y₀ varies among cultivars and environments (Table 3) , the amplitude of the bell-shaped curve also varies. Similarly, as X_tot (and therefore X₀) varies, the breadth of the curve varies. This variation in curve shape and related function parameters, which makes application of the function to leaf area estimation difficult, can be reduced by standardizing amplitude and breadth. This is done by dividing the areas of individual leaves by Y₀, and dividing the leaf numbers by X_tot (Dwyer et al., 1992). If leaf area data for each cultivar–experiment combination are normalized for their corresponding values of Y₀, then leaf area of mature leaves normalized for Y₀ is described by

(1b)

where , ₀, , and result from the fit of the normalized values (Dwyer and Stewart, 1986). If data for cultivar–experiment combination are also normalized for the total number of leaves, then leaf area of mature leaves normalized for Y₀ and X_tot is described by

(1c)

where and ₀ result from normalization for total leaf number, and , , and result from normalization for leaf area of mature leaves and total leaf number. This latter type of normalization is referred to here as double normalization.

1. Select an Eq. [1c] parameter set to be validated (e.g, the one derived for all PR data) and an observed data set to validate against (e.g, a cultivar–environment combination at TL).
   
2. Multiply Eq. [1c] values (of PR) by observed X_tot (of TL), giving estimated X values.
   
3. Multiply Eq. [1c] values by observed Y₀, giving estimated Y values. Estimated X and Y values will differ to some extent from the observed ones.
   
4. From the resulting curve, read the estimates of X₀ and Y₀.
   
5. Determine and sum the estimates of all leaf blade areas to obtain an estimate of Y_p.
   
6. Fit Eq. [1a] to the cultivar–environment combination to validate against, which provides another estimate of X₀, Y₀, and Y_p.
   
7. Compare both estimates of X₀, Y₀, and Y_p.
   
8. This method has been followed in this study; alternatively, estimated Y_p can be compared with the average observed Y_p.
   

Direct Estimation
Francis et al. (1969) and Pearce et al. (1975) proposed directly estimating Y_p by multiplying the area of a particular leaf of that plant by a leaf area factor. Their leaf area factors were used in this study to estimate Y_p for each cultivar–environment combination. As this proved unsatisfactory (see Results), a new leaf area factor F was determined. Whereas Francis et al. (1969) and Pearce et al. (1975) related Y_p to the area of a leaf with a fixed leaf number, it was hypothesized that Y_p could be related to the area of the largest leaf, in an attempt to account for genotypic and environmental effects on total leaf number. Therefore, leaf area factor F was defined as

(2)

Since the number of the largest leaf, counted from the base of the plant, depends on the total number of leaves, F also depends on X_tot. In linear form:

(3)

Plot averages of cultivar–environment combinations were used in curve fitting. For each cultivar, with the exception of the highland hybrid that was represented by only three data points, and for all cultivars combined, Eq. [3a] was fitted, and slopes and intercepts of the regressions were tested for heterogeneity with the PROC GLM method of the SAS statistical package (SAS Inst., 1985). Also, Eq. [3a] was fitted for PR and for TL data. The latter fits were validated against cultivar–environment combinations from TL and BA, and PR and BA, respectively (cross-validation). Validation involves the following steps:

1. Select an Eq. [3a] parameter set to be validated (e.g., the one derived for all PR data) and an observed data set to validate against (e.g., a cultivar–environment combination at TL).
   
2. Establish the value of F on the basis of observed X_tot (of TL) and the solution of Eq. [3a] (for PR).
   
3. Multiply the observed area of the largest individual leaf (of TL) by F, resulting in an estimate of Y_p.
   
4. Compare estimated and observed Y_p.
   

	Results

TOP ABSTRACT INTRODUCTION Materials and methods Results Discussion Conclusions REFERENCES

 
Bell-Shaped Function
Fitting Eq. [1a] to cultivar–environment combinations resulted in variable estimates of curve parameters. Differences in growing conditions caused great differences in the leaf area of the largest leaf. Estimates of Y₀ varied between 153 cm² for Pool 16 C₂₀ in Environment 6 and 940 cm² for CML247 x CML254 in Environment 11 (Table 3). Averaged over environments, Y₀ was smallest for Pool 16 C₂₀ with a value of 509 cm², and greatest for CML247 x CML254 with a value of 703 cm² (not considering CML246 x CML242 that was only cultivated once under good growing conditions). Estimates of X₀ varied between 9.23 leaves for Pool 16 C₂₀ in Environment 6 and 16.04 leaves for CML247 x CML254 in Environment 10. Averaged over environments, X₀ was smallest for CML246 x CML242 and the early-maturing cultivar Pool 16 C₂₀, with values of 11.09 and 11.78 leaves, respectively, and greatest for the late-maturing cultivar CML247 x CML254, with a value of 14.93 leaves. Values of parameters a and b also varied among cultivars and environments. Parameter b was negative in some instances, particularly in the case of CML247 x CML254 and Environment 4. Correlations between the four curve parameters and total leaf number (n = 46) over all cultivar–environment combinations were

In general, coefficients of determination were too low to justify estimation of curve parameters from total leaf numbers. Whereas Y₀, X₀, and a were weakly associated with X_tot, variables b and X_tot were independent, and the correlation between Y₀ and X₀ was weak. Curvilinear relationships also did not offer adequate descriptions.

Data normalization with respect to only Y₀ and fitting of Eq. [1b] naturally considerably reduced variation of ₀ in comparison with Y₀, with estimates of ₀ varying mostly between 0.998 (only two values were lower) and 1.010 (Table 4) . The overall average of ₀ was 1.001. However, other curve parameters maintained considerable variation (data not presented).

View this table: [in this window] [in a new window]   Table 4 Range and average in parameter values for Eq. [1c] for each cultivar over all environments, after normalization for largest leaf area and leaf number. 0 and 0 are the normalized values (dimensionless) of the mature leaf area and leaf number, respectively, and and are empirical constants

Although fitting of Eq. [1a] through all combined original data was not possible due to the wide data scatter, it was possible for Eq. [1c] through all data normalized with respect to both Y₀ and X₀. This resulted in parameter values of , and (Table 4). Fitting Eq. [1c] for PR data only resulted in parameter values of 0.980, 0.669, -10.19, and 0.85, respectively, and for TL data only in parameter values of 0.978, 0.651, -9.68, and 1.84, respectively (Table 4). The three fitted curves differ very little in their shape (Fig. 1) . The 95% confidence interval of the curve through all data points was too narrow to be displayed in Fig. 1.

View larger version (30K): [in this window] [in a new window]   Fig. 1 Equation [1c] fitted to all cultivar–environment combinations (ALL), and to those at Poza Rica (PR) and Tlaltizapán (TL). Because of the closeness of the lines, only part of the curves are given. The inset shows the complete ALL curve with the 95% prediction interval for five new observations

View larger version (19K): [in this window] [in a new window]   Fig. 2 Comparison of observed and estimated total plant leaf area (Yp) for all cultivar–environment combinations. Bell curve and direct data points represent Yp estimations with Eq. [1c] and [3a], respectively. TL&BA data points represent Yp estimations for Tlaltizapán and El Batán that are estimated on the basis of the parameter set developed for Poza Rica. Similarly, PR&BA data points represent Yp estimations for PR and BA that are estimated on the basis of the parameter set developed for TL. The line indicates a 1:1 relationship

View larger version (23K): [in this window] [in a new window]   Fig. 3 Two illustrations of validation of Eq. [1c]. (a) Leaf area of Pool 16 C20 in Environment 10 at TL (three replicates) was closely estimated with the curve fitted for PR, whereas (b) leaf area of CML247 x CML254 in Environment 6 at PR (three replicates) was rather poorly estimated with the curve fitted for TL. The dotted lines represent the fits of Eq. [1a]

(4b)

View larger version (28K): [in this window] [in a new window]   Fig. 4 Regression for all cultivar–environment combinations of the ratio of the area of the largest leaf blade to total plant leaf area (F) on total leaf number (Xtot), for all values of Xtot, and for 17 < Xtot < 23. Confidence and prediction intervals (95%) are valid for the regression for all values of Xtot

Cross-validation proved satisfactory. For example, in spite of the increasing scatter of F vs. X_tot at smaller leaf numbers, LAI_{tot, max} was accurately estimated for all cultivars in TL and BA environments on the basis of Eq. [3a] parameterized for all cultivar–environment combinations in PR (Fig. 2). As cultivar-specific parameter sets of Eq. [3a] were not significantly different, and as cross-validation gave similar results, it is concluded that for estimation purposes, the combined relationship of Eq. [3b] can be used. Its 95% prediction intervals for five observations are given in Fig. 2. More observations only marginally increased prediction accuracy.

Averaged over all cultivars and environments, areas of the leaves located one and two positions below the largest leaf were 93% and 86%, respectively, of that of the largest leaf. Corresponding values of the leaves located one and two positions above the largest leaf were 93% and 84%, respectively. This indicates a relative error of 7% and 15% respectively if a leaf one or two positions away from the largest leaf is accidentally used in computations.

	Discussion

TOP ABSTRACT INTRODUCTION Materials and methods Results Discussion Conclusions REFERENCES

 
Bell-Shaped Function
The bell-shaped function (Eq. [1a]) proved its robustness by adequately describing a wide range of combinations of cultivars and environments. Cultivars varied in maturity, environmental adaptation, and growth vigor. Environments included N-limited, moisture-limited, and favorable growing conditions. The leaf area of the largest leaf was associated with these growth restrictions, and variation in photoperiod and temperature during the photoperiod-sensitive phase of the maize plant led to variation in total leaf number (Table 5) within cultivars. Parameter a (breadth of the curve) was smaller (narrower) for late-maturing germplasm than for early-maturing germplasm, but was not related to growing conditions (Table 3). Parameter b (skewness of the curve) was negative (curve skewed to the left) in 11 cases that were not consistently associated with particular cultivars or environments (Table 3). Data normalization for the area of the largest leaf (Eq. [1b]) was effective in reducing variation in Y₀. However, applicability of both Eq. [1a] and [1b] is restricted to a particular genotype–environment combination, as genotype x environment interaction for leaf number was not taken into account in this analysis. This problem was solved by normalization for both the area of the largest leaf and total leaf number (Dwyer et al., 1992). The resulting Eq. [1c] could be applied to individual genotype–environment combinations and to all experiments at one location, but also proved capable of describing all double-normalized data jointly with one parameter set. Cross-validation showed that Eq. [1c] can be used to estimate individual values of Y₀, X₀, and Y_p (Fig. 2). This procedure is specified in the Materials and Methods section. Prediction on the basis of observations on at least five plants is sufficiently accurate. Some accuracy is lost in comparison with Eq. [1a], which is the inevitable but acceptable price to be paid for greatly reducing the size of the original data set. The degree of accuracy required will determine whether this is acceptable.

The need to estimate Y₀, X₀, a, and b in advance remains, if this approach is to be of use in, for example, crop simulation models. Muchow and Carberry (1989), Keating and Wafula (1992), and Birch et al. (1998), who conducted experiments under nonlimiting growth conditions, suggested that these parameters can be estimated from the total leaf number; they found significant linear relations between total leaf number and the four curve parameters. This is an attractive solution, as total leaf number can be determined well before flowering by destructive sampling and dissection of growing leaves. Total leaf number can also be estimated from thermal time accumulated from emergence to tassel initiation (Kiniry, 1991), or visible leaf number at tassel initiation (Russell and Stuber, 1984). Keating and Wafula (1992) were careful not to imply that their relationships were valid for growth-limiting conditions, and Birch et al. (1998) observed genetic variation in their relationships. This study shows that such linear relationships do not hold for a wider range of growing conditions that are characterized by limited availability of moisture and nutrients, and that curvilinear relationships do not offer adequate descriptions where there is wide data scatter. This study therefore did not lead to a good substitute for parameter estimation on the basis of total leaf number.

Direct Estimation
The relationship (Eq. [3b]) that is presented here builds on the proposal by Francis et al. (1969) and Pearce et al. (1975) that total leaf area per plant can be related to the area of a single leaf blade that represents a relatively large proportion of the total leaf area. Application of their leaf area factors to the five tropical cultivars of this study resulted in serious underestimation of total plant leaf area, and variation in calculated leaf area factors from present experimental data averaged across cultivars was unacceptably large. Equation [3b] corrects for the consequences of variation within and among cultivars in leaf number on the relative area of the largest leaf. Equation [3b] was developed for four tropical lowland open-pollinated cultivars, one tropical lowland hybrid and one tropical highland cultivar that varied in leaf number and that were grown at production levels varying between 0.7 and 8.8 t grain ha^-1 (Table 1). Variation in photoperiod and temperature led to variation in total leaf number within cultivars (Table 5). The cultivars that were used in this study differ in photoperiod-by-temperature sensitivity (Edmeades et al., 1994), but this did not significantly alter regression coefficients. Phenotypic variation in leaf characters (e.g., leaf thickness), as a consequence of variation in soil N or soil moisture availability, was also accounted for. Equation [3b] was developed for a plant density of 5.3 plants m^-2, and as other relationships have proven to be density-dependent (Fakorede et al., 1977; Francis et al., 1969), care must be taken when applying this relationship to substantially different plant densities.

Pool 16 C₂₀ caused especially wide data scatter at X_tot < 17. Without observations on other cultivars with similarly few leaves, it is not possible to determine whether this is a cultivar effect or whether it indicates a fundamental limitation of the technique. And although validation results remained satisfactory, restricted application of the technique is recommended in instances of low total leaf numbers.

This method of estimating leaf area is explained in Materials and Methods. Prediction on the basis of observations of at least five plants is sufficiently accurate. The largest leaf in the experiments was usually one of the three leaves below the ear leaf, or the ear leaf itself. If the largest leaf can not be identified with reasonable certainty, it may be necessary to measure all four of them, which adds considerably to the labor required. If accuracy is required, this is a worthwhile investment, as the average error is relatively large at 7% and 15% if the selected leaf is one or two positions away from the largest leaf, respectively. However, as a general rule it can be assumed that the leaf with the greatest length is also the one with the greatest area. Measurement of the largest area of the mature leaf could be done during grain filling but before the beginning of senescence, which through shriveling influences the leaf blade area. Establishment of maximum plant leaf area using Eq. [3b] requires the total number of leaves, which can be counted reliably if leaf numbers 5 and if necessary 10 are marked early in the season so that lower leaves are accounted for before they senesce and are lost.

The above procedures lead to estimation of total plant leaf area. The value of LAI_{tot, max} can be derived from Y_p and plant density. Plant density is known, or can be determined easily. Even though plant density in all experiments was 5.3 plants m^-2, given the validity of the parameter sets over a wide range of growing conditions and values of Y_p, it is expected that the parameter sets are valid for a relatively wide range of plant densities. They must be applied with care at very low or very high densities.

Green leaf area index is more important than LAI_tot for computation of crop growth rate, as LAI_tot includes photosynthetically inactive, dead leaf area. Greatest LAI_g in tropical maize is usually reached at flowering, after which it may stabilize for several weeks, depending on growing conditions, and then declines at an increasing rate as green leaf tissue senesces. A moderate level of senescence normally occurs before flowering, so greatest LAI_g is usually smaller than LAI_{tot, max}, and can be computed from LAI_{tot, max} and the fraction of senesced leaf area at flowering. The development of LAI_g over time remains a problem. In many plant breeding programs, observation of leaf area development is too labor-intensive. Breeders therefore often resort to observation of the plant type, which incorporates the amount of green leaf area, around or just after flowering. This is much less laborious than direct measurements of LAI with electronic devices (Denison and Russotti, 1997). Competition and detailed crop growth studies require temporal LAI_g estimates. If estimates are required beforehand, this will require some form of model calibration, depending on the equations and growing conditions. If estimates are required afterwards, LAI_g at any stage can be estimated by determining the fraction of dead leaf area over time and relating this to LAI_{tot, max}.

	Conclusions

TOP ABSTRACT INTRODUCTION Materials and methods Results Discussion Conclusions REFERENCES

 
The bell-shaped curve adequately described a wide range of tropical cultivars under water- and N-limited and favorable growing conditions. Data normalization for the area of the largest leaf and total leaf number of individual cultivar–environment combinations resulted in one parameter set that described all combinations. Predictive application in, for instance, crop growth models remains restricted as advance estimation of Y₀, X₀, a, and b remains problematic. Analytical application after flowering, when those values can be determined, is possible. If the total number of leaves and area of the largest leaf are known, total plant leaf area can also be estimated directly, making use of the facts that the area of the largest leaf relative to total plant leaf area is constant, and that this constant is linearly related to total leaf number. Prediction on the basis of observations of at least five plants is sufficiently accurate. This study has shown that approaches previously presented by others are in modified form applicable over a wide range of environmental conditions.Edmeades Chapman Lafitte 1994; SAS Institute 1985

	ACKNOWLEDGMENTS

 
The leaf area measurements made by L. Castañeda, J.C. Bahena, S. Rivas, and their teams are gratefully acknowledged. Drafts of the manuscript were reviewed by G.O. Edmeades, D.A. Hartkamp and various anonymous reviewers. J. Crossa and J. Withagen provided statistical advice.

Received for publication April 27, 1999.

	REFERENCES

TOP ABSTRACT INTRODUCTION Materials and methods Results Discussion Conclusions REFERENCES

 

• Bell M.A., van Keulen H. Soil pedotransfer functions for four Mexican soils. Soil Sci. Soc. Am. J. 1995;59:865-871.[Abstract/Free Full Text]
• Birch C.J., Hammer G.L., Rickert K.G. Improved methods for predicting individual leaf area and leaf senescence in maize (Zea mays). Aust. J. Agric. Res. 1998;49:249-262.
• Denison R.F., Russotti R. Field estimates of green leaf area index using laser-induced chlorophyll fluorescence. Field Crops Res. 1997;52:143-149.
• Dwyer L.M., Stewart D.W. Leaf area development in field-grown maize. Agron. J. 1986;78:334-343.[Abstract/Free Full Text]
• Dwyer L.M., Stewart D.W., Hamilton R.I., Houwing L. Ear position and vertical distribution of leaf area in corn. Agron. J. 1992;84:430-438.[Abstract/Free Full Text]
• Edmeades G.O., Chapman S.C., Lafitte H.R. Photoperiod sensitivity of tropical maize cultivars is reduced by cool night temperatures. Agron. Abstr. 1994;86:146.
• Elings A., White J.W., Edmeades G.O. Options for breeding for greater maize yields in the tropics. Eur. J. Agron. 1997;7:1-14.
• Fakorede M.A.B., Mulamba N.N., Mock J.J. A comparative study of methods used for estimating leaf area of maize (Zea mays L.) from nondestructive measurements. Maydica 1977;22:37-46.
• Francis C.A., Rutger J.N., Palmer A.F.E. A rapid method for plant leaf area estimation in maize (Zea mays L.). Crop Sci. 1969;9:537-539.[Abstract/Free Full Text]
• Keating B.A., Wafula B.M. Modelling the fully expanded area of maize leaves. Field Crops Res. 1992;29:163-176.
• Kiniry J.R. Maize phasic development. In: Hanks J., Ritchie J.T., eds. Modeling plant and soil systems. Agron. Monogr. 31. Madison, WI: ASA, CSSA, and SSSA, 1991:55-70.
• Lal K.N., Subba Rao M.S. A rapid method of leaf area determination. Nature (London) 1951;167:72.
• McKee G.W. A coefficient for computing leaf area in hybrid corn. Agron. J. 1964;56:240-241.[Free Full Text]
• Montgomery E.G. Correlation studies in corn. Neb. Agric. Exp. Stn. Annu. Rep. 1911;24:108-159.
• Muchow R.C., Carberry P.S. Environmental control of phenology and leaf growth in a tropically adapted maize. Field Crops Res. 1989;20:221-236.
• Pearce R.B., Mock J.H., Bailey T.B. Rapid method for estimating leaf area per plant in maize. Crop Sci. 1975;15:691-694.[Abstract/Free Full Text]
• Russell W.K., Stuber C.W. Prediction of tassel initiation in corn. Crop Sci. 1984;24:507-510.[Abstract/Free Full Text]
• SAS Institute. SAS user's guide: Statistics. Cary, NC: SAS Inst, 1985.
• Stewart D.W., Dwyer L.M. Mathematical characterization of leaf shape and area of maize hybrids. Crop Sci. 1999;39:422-427.[Abstract/Free Full Text]
• Watson D.J. Comparative physiological studies on the growth of field crops: I. Variation in net assimilation rate and leaf area between species and varieties, and within and between years. Ann. Bot. (London) 1947;41:41-76.

This article has been cited by other articles:

				H. F. Oraby, C. B. Ransom, A. N. Kravchenko, and M. B. Sticklen Barley HVA1 Gene Confers Salt Tolerance in R3 Transgenic Oat Crop Sci., September 23, 2005; 45(6): 2218 - 2227. [Abstract] [Full Text] [PDF]

				K. D. Subedi and B. L. Ma Ear Position, Leaf Area, and Contribution of Individual Leaves to Grain Yield in Conventional and Leafy Maize Hybrids Crop Sci., September 23, 2005; 45(6): 2246 - 2257. [Abstract] [Full Text] [PDF]

				J. Yang and M. Alley A Mechanistic Model for Describing Corn Plant Leaf Area Distribution Agron. J., January 1, 2005; 97(1): 41 - 48. [Abstract] [Full Text] [PDF]

This Article Abstract Figures Only Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in ISI Web of Science Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (8) Citing Articles via Google Scholar Google Scholar Articles by Elings, A. Search for Related Content PubMed Articles by Elings, A. Agricola Articles by Elings, A. Related Collections Crop Growth and Development Maize Plant Analysis