Crop Science Journal of Natural Resources and Life Sciences Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

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 Related articles in Crop Science
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI 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 ISI Web of Science (7)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Correia, P. J.
Right arrow Articles by Martins-Loução, M. A.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Correia, P. J.
Right arrow Articles by Martins-Loução, M. A.
Agricola
Right arrow Articles by Correia, P. J.
Right arrow Articles by Martins-Loução, M. A.
Related Collections
Right arrow Other Crops
Right arrow Crop Models
Right arrow Plant Nutrition
Crop Science 42:1577-1583 (2002)
© 2002 Crop Science Society of America

CROP ECOLOGY, MANAGEMENT & QUALITY

Nutritional Diagnosis in Carob-Tree

Relationships between Yield and Leaf Mineral Concentration

Pedro José Correia*,a, Ilda Anastáciob, Maria da Fé Candeiasc and Maria Amélia Martins-Louçãod

a FERN, Univ. do Algarve, Campus de Gambelas 8000-117 Faro, Portugal
b AIDA Loteamento Industrial de Loulé, Apartado 302, 8100 Loulé, Portugal
c Direcção Regional de Agricultura do Algarve, Serviços de Apoio, Largo de Sto Amaro 8800 Tavira, Portugal
d Dep. de Biologia Vegetal, Faculdade de Ciências de Lisboa, Campo Grande, C2 Piso 4, 1749-016 Lisboa, Portugal

* Corresponding author (pcorreia{at}ualg.pt)


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSION
 REFERENCES
 
Reliable fruit production of carob-tree (Ceratonia siliqua L.) in the Mediterranean region may be linked to nutrient availability and nutrient status of the trees. The objectives of this study were to determine relationships between yield and leaf nutrient concentrations, determine limits of nutrient concentration required for maximum yield, and establish yield prediction models for carob-tree. The study was conducted in a mature carob-tree orchard established on an acid soil, with low organic matter content. Treatments consisted of two N rates (as ammonium nitrate), 0.3 and 0.9 kg N tree-1 yr-1 and three irrigation levels (0, 50, and 100% of Class A pan evaporation). Each tree also received 0.6 kg K after the second year of the experiment. The field trial was conducted during the 4 yr from 1992 to 1995, and leaf N, P, K, Ca, Mg, Fe, Mn, and Zn concentrations were analyzed in autumn (40 d after full bloom) and winter (90 d after full bloom). A nonbearing year (third year of experiment) was excluded from this study. Each nutrient concentration value obtained on both sampling dates was correlated with fruit production and subsequently several multiple regressions were tested. The best estimation model indicated that 92% of yield variation may be related to N, P, K, Mn, and Fe leaf concentration values. This model was validated against independent data. The optimal leaf nutrient limits were established corresponding to maximum yields obtained under the conditions of this site.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSION
 REFERENCES
 
THE EVALUATION OF TREE NUTRITIONAL STATUS in natural environments is particularly difficult due to the interference of climatic variables. Additionally, in fruit tree orchards, management practices such as fertilization and pruning can greatly influence nutrient concentration patterns.

Among evergreen species of the Mediterranean basin, mineral nutrition does not seem to have been extensively investigated, although some work has been done with different species of Quercus (Sabaté and Gracia, 1994; Oliveira et al., 1995; Robert et al., 1996) and Olea (Bouat, 1984; Jordão et al., 1994; Bouranis et al., 1999).

Leaf analysis provides a useful tool for assessing nutritional status of several plant species (Mengel and Kirkby, 1987), and in many cases is directly related to fruit production. Therefore, models have been proposed for defining the optimal nutrient concentration, corresponding to maximum yield in a certain type of soil (Guzmán and Romero, 1988; Moreno et al., 1998; MacKerron and Young, 1999). On the other hand, the comparison of leaf analysis with critical nutrient levels and yield may be used to improve fertilizer efficiency (Angus, 1995).

Plant-testing methods based only on leaf analysis pose some drawbacks such as dilution effects, and new approaches such as the use of flowers instead of leaf analysis have been developed for highly profitable crops like Citrus, peach, and pear trees (Palazzo et al., 1993; Sanz et al., 1994; Belkhodja et al., 1998). In less profitable crop species such as carob-tree, an evergreen species that has become more economically important during the last decade, the information concerning the relationships between leaf nutrient variation, soil nutrient availability, and yield is absent. The economic importance and sustainability of this crop in the whole Mediterranean area is dependent on the reliability of fruit production. Thus, it is important to clarify whether nutrient availability and consequently nutrient status of trees have direct effects on yield.

In a previous paper, leaf nutrient variation related to phenological events has been studied in a carob-tree grove established on an infertile soil and submitted to different N and irrigation treatments (Correia and Martins-Loução, 1997a). The results presented here refer to concurrent studies at the same field trial. The aim of this work was (i) to analyze possible relationships between yield and leaf nutrient concentrations, (ii) to determine the limits of nutrient concentration required for maximum yield, and (iii) to establish yield prediction models for the carob-tree.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 CONCLUSION
 REFERENCES
 
Experimental Site
The experimental orchard was located in southern Portugal. When the experiment started, trees were 30 yr old (cv. Mulata) and had been planted on a sandy loam soil at 12- by 12-m spacing. Three soil samples, taken at depths of 0- to 40-cm, were mixed and a representative composite soil sample was oven-dried for 48 h at 30°C and passed through a 2-mm sieve. Soil N was determined by the Kjeldahl method (Kjeldahl, 1883), and K and P were extracted using a solution of ammonium lactate and acetic acid (Riehm, 1958). The P content in the extracts was quantified colorimetrically, and K by flame photometry (Isaac and Kerber, 1971). Soil pH was determined in soil-water (1:2.5) suspension and organic carbon by oxidation of dicromate (Walkley and Black, 1934). A correction factor (1.724) was used to convert organic C in soil organic matter content (Walkley and Black, 1934). Soil chemical analysis of the composite sample is presented in Table 1 . The trees were subjected to three irrigation and two fertilization regimes during the four consecutive years (1992–1995). The three irrigation regimes were 0% (precipitation only), 50%, and 100% of water loss from a Class A pan. The fertility regimes were two N application levels (0.3 kg N tree-1 yr-1 and 0.9 kg N tree-1 yr-1). Fertilizer was applied in April and irrigations were applied from June to August. Nitrogen fertilizer contained 15% Ca and 20.5% N with equal amounts of NO-3 and NH+4. In the second year of the experiment, potassium sulphate was also applied (0.6 kg K tree-1 yr-1) to all treatments. No P was applied since a previous leaf analysis made before the application of fertilizers did not indicate a deficiency in this element. Each of the six N x water combinations (treatments) was replicated three times with four trees per replication, distributed in a completely randomized design.


View this table:
[in this window]
[in a new window]
  		Table 1. Chemical analysis of a composite soil sample obtained at 0- to 40-cm depth at the beginning of the experiment.

 
Leaf Sampling and Leaf Mineral Concentrations
In each treatment, two to three trees were selected for leaf nutrient analysis. Mature leaves (1.5 yr old) were taken at random from the outer edge of the crown in all directions, at 1.8- to 2.0-m height. Leaf sampling was conducted every two months from May 1991 until December 1994 and consisted of 30 to 40 mature leaves per tree. Carob-tree flowering occurs during a relatively long time interval, and it is not possible to specify full blooming with accuracy. However, under our experimental conditions, a good correlation between inflorescence number registered at the beginning of October (first week) and yield (r2 = 0.90; P < 0.001; Correia and Martins-Loução, 1997b) was obtained. Autumn and winter samplings of leaves were considered 40 and 90 d after such date.

Leaf N concentration was determined by the Kjeldhal method (Kjeldhal, 1883); P was analyzed spectrophotometrically; and K, Ca, Mg, Mn, Zn, and Fe by atomic absorption spectroscopy (Pye Unicam, Cambridge, UK) according with the Association of Analytical Chemists (1990). Concentration values used for model determination were expressed as g kg-1 and mg kg-1.

Fruit Harvesting
Pods were harvested every year in September and yield was expressed as kg ha-1. In 1994 (the third year of the experiment), there was a decrease in fruit production for all N x water combinations (Correia, 1996). This year was, therefore, considered as a nonbearing year and was not included in the yield curve determinations.

Models
Fruit production was related to leaf nutrient concentration determined at different leaf collection times, and only during the fruit-bearing years by using several regressions. To validate the models under different nutrient availability, fertilization, and irrigation, treatments were combined and analyzed as one. Thus, fruit production registered in 1992, 1993, and 1995 (combined yield) was separately correlated with leaf N, K, P, Ca, Mg, Mn, Zn, and Fe concentrations obtained in autumn 1991, 1992, and 1994 (autumn sampling), and those obtained in winter 1992, 1993, and 1995 (winter sampling). Spring and summer leaf nutrient concentrations were excluded from this study since the vegetative growth reaches a maximum during these seasons (Correia and Martins-Loução, 1995) and might dilute nutrient concentration.

Iron and Zn concentration values were transformed (the square root of the values was used) in order to obtain linear trends; data linearization indicated that the effects of the independent variables (nutrients) were additive (data not shown). Hence, the variates N, P, K, Mn, Fe, and Zn were regressed on yield (dependent variable) and several multiple regression models of the form: y = a0 + a1x1 + a2x2 +...+ anxn (Ott, 1992) were obtained.

Model Validation
One of the aims of this study was to establish a yield prediction model, which consisted of calculating values of the response variable using a regression equation. This is achieved by using the equation that maximizes the coefficient of determination (Legendre and Legendre, 1998). Therefore, the best model was chosen based on the adjusted coefficient of determination, which is dependent on sample size and on the significance level, and subsequently tested on independent data. The validation procedure consisted of the following steps: (i) Mature carob-trees (10–20 yr old, mainly Mulata) were selected in southern Portugal and leaves were collected for chemical analysis. The leaf sampling and analytical procedures were the same as described above. These trees were established in a calcareous, unirrigated soil, and submitted to four different fertilization regimes: N, K, N plus K, and an unfertilized control (for details see Correia et al., 1999). In the validation trial, fertilizers were applied in April and leaf sampling was done in autumn of the same year. In the following year, pods from each fertilization regime were harvested and weighed (observed yield). This procedure was done in two consecutive years; (ii) Using the model, yield was estimated for each treatment corresponding to eight validation points. Each point was the mean of three replicates; (iii) and finally, observed values were correlated to estimated values.

Statistical Analysis
The data analysis was conducted with SAS procedures (SAS Institute, 1989).


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 CONCLUSION
 REFERENCES
 
The production of carob during four consecutive years is presented in Table 2 . Significant differences were observed among treatments after the first application of water and fertilizer in the spring of 1991. These differences were particularly clear between treatments with low and high N treatments. In 1994, yield decreased and reached the values registered in 1991 (data not shown), that is, the pretreatment year. On the basis of this 4-yr trend, 1994 was thus considered a nonbearing year. Fruit production increased again in 1995 and trees fertilized with low N and without irrigation showed significantly lower yield compared with high N treatments.


View this table:
[in this window]
[in a new window]
  		Table 2. Carob-tree fruit production for two N fertilizer application rates at three irrigation levels. 1994 was considered a nonbearing year and was not included in the subsequent regression analysis.

 
The relationships between yield and leaf nutrient concentration determined in autumn and winter are shown in Table 3 . For Ca and Mg, regressions were not significant and these two nutrients were excluded from subsequent data analysis, namely the elaboration of the final model. Phosphorus and Zn concentrations determined in autumn were not related to yield. Regressions presenting the highest correlation coefficients are shown in Fig. 1 and 2 .


View this table:
[in this window]
[in a new window]
  		Table 3. Regressions between carob-tree fruit production (Y) and leaf nutrient concentration (x) determined in autumn and winter.

 


View larger version (18K):
[in this window]
[in a new window]
  		Fig. 1. Regressions between carob-tree yield and leaf N (A), K (B), and P (C).

 


View larger version (19K):
[in this window]
[in a new window]
  		Fig. 2. Regressions between carob-tree yield (kg ha-1) and leaf Mn (A), Zn (B), and Fe (C).

 
Figure 1A shows the relationship between fruit production and leaf N concentration determined at the autumn sampling. Leaf N concentration explained 78% of yield variation, and maximum fruit yields (2600 to 2700 kg ha-1), were obtained with the highest leaf N concentrations (between 24.0 g kg-1 and 25.0 g kg-1 of N leaf dry weight). Similar trends were observed for yield vs. leaf K (Fig. 1B), but only 63% of yield variation is linked to K concentration, determined in the same sampling dates as for N. Maximum fruit production (Fig. 1B) was obtained with 12.5 g kg-1 leaf K and did not increase with higher leaf K. On the contrary, increasing leaf P concentration was associated with decreasing fruit production (Fig. 1C). This trend was highly significant (r2 = 0.58; P = 0.0016) and maximum fruit production was recorded at 0.8 g kg-1 leaf P.

Fruit yield showed a quadratic response to leaf Mn concentration from winter sampling (Fig. 2A). The highest yields were obtained within a range of 130 to 140 mg kg-1. Zinc at winter sampling (Fig. 2B) and iron at autumn sampling (Fig. 2C) were related to yields according to multiplicative models presenting a negative slope, where greater fruit production was achieved at lower nutrient concentration. Nutrient concentrations resulting in maximum yields were 9.5 to 11 mg Zn kg-1 and 42 to 50 mg Fe kg-1.

Multiple regression analysis data are presented in Table 4 , which shows that adding successive variables, particularly micronutrients, improves the model. Apparently, the best fitted model includes Naut, Pwin, Kaut, Mnwin, Feaut, and Znwin levels. Nitrogen, K, and Fe determined in winter, and P, Mn, and Zn determined in autumn were excluded in this variable testing due to lower correlation coefficients, as shown in Table 3. The analysis of residuals (data not shown) for this model shows that the residuals are increasing as the independent variables decrease. It should be noted that the absence of N in the models largely decreases the coefficient of determination. Therefore, the final equation used for yield estimates is the following:


View this table:
[in this window]
[in a new window]
  		Table 4. Dependent variables of regression models for carob-tree yield vs leaf nutrient concentration. Coefficients of determination and statistical significance of the models are presented. Subscripts in the models indicate the sampling date (aut = autumn, win = winter).

 
In order to validate the model, leaf N, P, K, Mn, Fe, and Zn concentrations recorded in the validation orchard were used with the final equation to estimate yield. The estimated yields were plotted against the observed yields (Fig. 3) . Each fertilization regime showed a large standard error in the observed yield, indicating a large difference in fruit production among trees that received the same fertilizer treatment. The trees that receive N plus K were the exception with a standard error of 30 kg ha-1. Variation in estimated yield values was also large. Accurate yield estimation can be approached, but a correction of estimated values should be made. As shown in Fig. 3, the model underestimated the real yield values. The correction of a certain estimate value should be at least increased to a value established by the bottom dashed line (Fig. 3), or to a maximum indicated by the top dashed line.



View larger version (23K):
[in this window]
[in a new window]
  		Fig. 3. Regression between observed and estimated carob-tree yield obtained by the six-parameter model presented in Table 4. Standard errors of the means are shown for both variables. Fertilization regimes used in the validation trial are: N (solid circles); K (open triangles); N plus K (solid squares); and unfertilized control (open circles). Upper and lower limits for the observed yield are represented by two dashed lines. Both lines were obtained by regression analysis.

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
CONCLUSION
 REFERENCES
 
The low soil fertility of the site (Table 1) and the absence of cultural practices during pretreatment years decreased maximum potential yield, since the greatest observed value after treatment (Table 2) was lower than other reported values for carob-tree (Esbenshade and Wilson, 1986; Lloveras and Tous, 1992). However, and as with other fruit trees, namely olive trees (Huguet, 1984; Jordão et al., 1994), higher yields were associated with an increase in N (Fig. 1A), K (Fig. 1B), and Mn levels (Fig. 2A). Maximum N (25.0 g kg-1) and K (12.5 g kg-1) leaf concentrations were greater than the values reported by Cabrita and Martins-Loução (1991) for mature trees under field conditions, and were greater than those reported by El-Gazzar et al. (1981) for potted carob-tree.

Fertilizer application during spring and summer increases vegetative growth (Correia and Martins-Loução, 1995) and leaf N retranslocation percentage from senescing leaves (Correia and Martins-Loução, 1997a), causing an improvement in tree nutritional status, particularly N concentration in leaves, as previously shown (Correia, 1996). Apparently, leaves sampled in autumn and winter, during vegetative rest, may be a useful tool to assess nutritional status in carob-tree. Expressing concentration values on a leaf area basis (Correia and Martins-Loução, 1997a) may allow the evaluation of other sampling periods by avoiding dilution effects. This procedure, however, has not yet been established as routine plant analysis.

Inflorescence emergence, pollination, and fruit setting occur during autumn and winter. In the carob-tree, nutrient concentrations of the inflorescences have been rarely studied (Cruz et al., 1988; Cabrita and Martins-Loução, 1991), therefore nutrient demands of reproductive structures, such as flowers and peduncles, are not fully understood. The presence of inflorescences in autumn and growing fruits in winter may be critical since they are expected to behave as active sinks, as happens in other plant species (Karmacharya and Singh, 1992; Drossopoulos et al., 1996; Bouranis et al., 1999). Apparently, high N demand by both flowers and fruits were observed in both irrigated and nonirrigated carob orchards (Cruz et al., 1988). A high demand of K by flowers was only noticed under dry conditions, but fruits represent an important sink for K under any orchard conditions (Cruz et al., 1988). It is possible that a spring fertilizer application is sufficient to supply the metabolic sinks since a positive correlation between leaf N content and inflorescence number was observed (Correia and Martins-Loução, 1993).

It is known that immediately after pollination there is an increase in P transport towards the young developing seeds (Mengel and Kirkby, 1987), and this fact may explain the P pattern (Fig. 1C). However, the information about flower nutrient demand is scarce (Cabrita and Martins-Loução, 1991). It is reasonable to assume that P was translocated from leaves to reproductive structures during winter (Cabrita and Martins-Loução, 1991), as occurs in other legumes (Drossopoulos et al., 1994), resulting in a negative correlation between yield and P concentration (Table 3). The lack of exogenous P during the experimental period did not likely fulfill nutrient demand, and stored P could have been consumed, apparently inducing a nutritional imbalance. However, throughout leaf and fruit development P levels remain at constant low values (Cruz et al., 1988), but similar to concentration levels of other Mediterranean plants (Kruger, 1985; Oliveira et al., 1995).

Leaf Mn (Fig. 2A) increased with increasing yield, but the concentration values were above the range indicated for Olea (Fernandez-Escobar, 1997). Since no toxicity symptoms were observed in the field, we assume that these values are adequate for carob-tree. The pattern followed by Zn (Fig. 2B) is not as clear, with only 34% of yield variation explained by Zn concentration during winter sampling and yield declining with increasing Zn. Those values are below the range recommended for Olea (Benton-Jones et al., 1991) and below the values registered in drip-irrigated carob-tree (San Juan, 1999, unpublished results). This may suggest some deficiency in this element, and therefore the sensitivity to yield variations was lost.

The correlation between yield and Fe observed in the autumn sampling may give indications about flower nutrient concentration in carob-tree. As observed in peach trees (Belkhodja et al., 1998), Fe concentration is normally higher in flowers than in leaves, and in the olive tree there is a clear demand of this element near full bloom (Bouranis et al., 1999). If flowers behave as active sinks for Fe, a negative correlation between yield and Fe leaf concentration would be expected and verified (Table 3, Fig. 2C). It should be pointed out that symptoms of Fe deficiency in young leaves were not observed at low Fe concentrations (Fig. 2C), which are in accordance with soil pH (Table 1). In these soil conditions (pH = 5.9), it is expected that Fe will be assimilated by the plant (Grusack et al., 1999). Only alkaline soils are normally associated with Fe chlorosis in several tree species (Koseoglu, 1995).

The absence of correlation between Ca and yield is probably associated with Ca static behavior (Table 3). Apparently, Ca is more important during carob vegetative growth, promoting stress tolerance to high temperature and high salinity (Martins-Loução, 1991, unpublished data), than during fruit development (Cabrita and Martins-Loução, 1991). Similarly, Mg concentration values do not provide any information concerning fruit production trends, due to the absence of correlations (Table 3). Carob tree does not show any relevant variation dynamics of leaf Mg concentrations (Cruz et al., 1988).

The resulting yield curves show that all the studied nutrients per se, with the exception for Ca and Mg, may be used to predict fruit production in the following year (Fig. 1 and 2) except in an off season production year. A quadratic equation was, in most cases, the best model (Fig. 1A, B, C, and Figure 2A). This equation has been used in wheat (Puente and Belda, 1999) and in potato tubers (MacKerron and Young, 1999) with good results. Assuming that all the studied nutrients interact on an additive basis, and providing that some variables should be numerically transformed, yield can be estimated from multiple regression equations. The best estimation indicates that 92% of the yield variation may be linked to N, P, K, Mn, Fe, and Zn leaf concentration values (Table 4) in spite of some less clear relationships, namely Mn and Zn. However, and as stated before, the best prediction model relies on the highest R2 (Legendre and Legendre, 1998). Residual plot analysis (data not shown) does not seem to suggest a higher-order model, but it should not be assumed that a significant correlation with yield should have a causal impact on yield. Potassium is positively related with yield (Fig. 1B), however, the sign of the model coefficient for K is negative. The independent variables often contain some redundant information and vary together, making it difficult to separate the effects of the different independent variables on the dependent variable. This multicollinearity, which was also shown by the large standard error of the K estimate (data not shown), does not affect the usefulness of a regression equation for prediction of new observations (Glantz and Slinker, 1990).

Validation of the model was based on data obtained from trees with different age, size, fertilization, and management practices and also established in different soil. However, these differences did not affect the validation of the model, since most of the leaf concentration values registered in the validation trees were within the same range determined in the experimental orchard. Since this is one of the conditions to overcome collinearity effects (Glantz and Slinker, 1990), the model should reasonably estimate fruit production. Since yield was higher in the validation orchard, even in the absence of irrigation (Correia et al., 1999), and for the same leaf nutrient concentration ranges, estimated yield values were lower than the observed values (Fig. 3), which caused the dispersion of results. Nevertheless, it is possible to draw two trend lines indicating an upper and lower level for the observed yield values. Thus, for a given estimated value of fruit production, a maximum and a minimum value can be calculated. The difference between those limits, {approx}1200 kg ha-1, may indicate an estimation of error of the model. The low fruit production values obtained (Table 2), in spite of fertilizer addition, are associated with the low soil fertility of the site (Table 1) and with the fact that above some leaf nutrient concentrations (N ranging from 24.0 to 25.0 g kg-1 dry weight, and K ranging from 12.5 to 14.0 g kg-1 dry weight), no yield increase is obtained.


   CONCLUSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSION
REFERENCES
 
Leaf N, P, K, Mn, Fe, and Zn concentrations during autumn and winter may be used to evaluate fruit production of the carob-tree. According to the model presented, optimal nutrient ranges of such elements can be defined. Model validation still needs to be tested for different soil conditions, plant development, and cultivars. However, this first yield model for carob-tree represents an important approach towards an improvement of carob-tree productivity.


   ACKNOWLEDGMENTS
 
We thank Algarverde for field facilities. This work was partially supported by INTERREG No. 20/REGII/6/96 and PRAXIS No 3/3.2/HORT/2168/95. We thank J. Osório and M. Pestana for their helpful remarks.

Received for publication June 26, 2001.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSION
 REFERENCES
 


Related articles in Crop Science:

This issue in Crop science

Crop Science 2002 42: 1393-1395. [Full Text]  




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 Related articles in Crop Science
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI 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 ISI Web of Science (7)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Correia, P. J.
Right arrow Articles by Martins-Loução, M. A.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Correia, P. J.
Right arrow Articles by Martins-Loução, M. A.
Agricola
Right arrow Articles by Correia, P. J.
Right arrow Articles by Martins-Loução, M. A.
Related Collections
Right arrow Other Crops
Right arrow Crop Models
Right arrow Plant Nutrition

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Vadose Zone Journal
Journal of Natural Resources
and Life Sciences Education		 Soil Science Society of America Journal
Journal of Plant Registrations Journal of
Environmental Quality		 The Plant Genome