| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Universidad de la República, Facultad de Agronomía, Garzón 780, Montevideo, Uruguay CP12900
* Corresponding author (chperdom{at}fagro.edu.uy).
 |
ABSTRACT |
|---|
 TOPNOTES ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: AVI, absolute volume increase • CONTV, control volume • LNC, leaf nitrogen concentration • MAP, months after planting • MAXV, maximum volume • NAI, nitrogen availability indices • NMIN, soil mineralizable nitrogen • RV, relative volume • SOM, soil organic matter
|
INTRODUCTION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Several approaches to identify nutrient deficiencies in eucalypts and guide fertilizations have been used in other countries. One of the most extensively used methods is tissue analyses based on critical ratios among nutrients (Herbert, 1996; Judd et al., 1996; Knight and Nicholas, 1996). Unlike single nutrient concentrations, nutrient ratios tend to remain relatively constant with age (Beaufils, 1973). Nutrient ratios also are less affected by the indeterminate growth habit of eucalypts, plants that have the capacity to grow at the expense of nutrient retranslocation, causing nutrient dilution (Cromer, 1996; Knight and Nicholas, 1996). The N/P ratio in foliage has been extensively used to guide N and P fertilization in eucalypts. In the case of N, it is assumed that response probability increases with ratio decrease from a defined critical value. Reports from Australia (Judd et al., 1996) and New Zealand (Knight and Nicholas, 1996) have identified the value 15 as the critical N/P ratio for E. globulus and E. nitens. This index is recommended for before and after planting in Australia and only at planting in New Zealand (Knight and Nicholas, 1996). The N/P ratio has also been used to detect N deficiencies in E. grandis, at values ranging from 13 in Australia (Judd et al., 1996) to 18 in South Africa (Herbert, 1996).
A more sophisticated methodology based on critical ratios among many nutrients is the Diagnostic and Recommendation Integrated System (DRIS; Beaufils, 1973), which has been widely used in Australia (Cromer, 1996) and New Zealand (Knight and Nicholas, 1996) to predict eucalypt response to fertilizers. Knight and Nicholas (1996) have questioned the validity of the DRIS sufficiency levels defined for New Zealand, however, and they have suggested that foliar analysis interpretation by this methodology is only tentative.
Although the nutrient ratio method is probably the most used guide for nutrient applications in eucalypt plantations, single nutrient concentrations have also been evaluated. Herbert (1996), working with E. grandis in South Africa, found that N response was not only related to the N/P ratio but also to the single N foliage concentration, for which he estimated a critical value of 28 g kg–1. Herbert (1996) also reported that N mineralization capacity and topsoil organic C content were promissory indices of N response. Shedley et al. (1995), working with 9-wk E. globulus seedlings in a glasshouse experiment, determined a critical leaf N concentration of 26 g kg–1 and an adequate range of 26 to 35 g kg–1. They also reported a similar range of values for a field experiment with 7-mo-old plants, suggesting that the range in leaf N concentration of this species does not significantly change during the first months of growth.
In other countries, there seems to be a lack of local information on plant and soil indices to predict response to N. According to reports from Argentina of Dalla-Tea and Marcó (1996) and from China of Huoran and Wenlong (1996), fertilizer recommendations in these countries are seldom based on soil or plant tests. Chilean researchers have reported that the foliar critical concentrations proposed for other countries are not always appropriate for their conditions, and suggest that local nutritional standards for the most common eucalypt species in the country should be determined (Prado and Toro, 1996).
In Uruguay, local nutritional standards to guide postplanting N applications also should be established. This need is based not only on the obvious economic benefits associated with efficient N use, but also on environmental concerns about the deleterious effects of N excess. Some of these effects are the well-documented water pollution with NO3 (Hamilton and Helsel, 1995) and increasing emission of N2O from soils. The latter has been linked to the worsening greenhouse effect and ozone layer destruction (Bouwman, 1990). Excessive N fertilization also tends to increase soil acidity and the amount of exchangeable Al, thus damaging soil quality (Ringrose and Neilsen, 2005). The objectives of this work were (i) to identify NAI based on single nutrient concentrations when N is applied to eucalypts in the first year after planting, and (ii) to estimate critical values for these indices to separate responsive from unresponsive sites. To accomplish these objectives, we conducted field experiments and evaluated the volume response to N of the two prevailing eucalypt species planted in Uruguay, E. grandis and E. globulus.
|
MATERIALS AND METHODS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Twenty experiments were conducted in Uruguay from 2001 to 2003 to study the volume response of E. globulus and E. grandis to N applied 6 or 12 MAP. Site-specific soil types and chemical properties are described in Table 1 . The experimental sites were located in the southeast, the midwest, and the northeast of the country, in commercial plantations of E. grandis and E. globulus. The sites represent the variability of eucalypt growth conditions in the country as far as region, soil type, previous crop, and crop management (Tables 1 and 2 ). Total annual rainfall during most of the experimental period was above normal and therefore soil water availability was not a limiting factor in this research.
|
|
Treatments and Experimental Design
The N treatments and method of application sometimes differed across these on-farm trials (Table 2). The N fertilizer rates (granulated urea) applied for sites fertilized 6 MAP were 0, 14, 28, and 41 g N plant–1, except for Sites 5, 6, and 7. These three sites were established in the second year of the research and received higher N rates because we had to compromise with plantation companies' wishes. They thought that first-year N rates were too low, especially at two sites established on reforested plantations (Sites 5 and 6), where higher amounts of decomposing plant residues are usually found on the soil surface. In experiments fertilized 12 MAP, N rates were higher than at most sites fertilized 6 MAP (except for Site 7) because older plants were expected to have higher N requirements and could be less affected by any potential negative effects of high N rates than younger plants.
At most sites, the urea was broadcast by hand over a radius of approximately 1 m around each tree, and was not incorporated into the soil. This is the most common application method used in Uruguay plantations. At three sites (Sites 15, 19, and 20), the N fertilizer was applied by the plantation companies using their equipment and normal practices. In Site 15, N was machine applied using a centrifugal-type spreader and two N sources were evaluated, urea and NH4NO3, because they wanted to compare these sources when broadcasting N on the soil surface without incorporation. At Sites 19 and 20, the urea was applied to one side of each crop line and incorporated into the soil by disking.
A randomized complete block design with three replicates per treatment was used at all sites. Each plot consisted of 36 plants arranged in six rows of six plants. Exceptions were the sites fertilized by the cooperating companies (Sites 15, 19, and 20), where plots consisted of 60 plants arranged in six rows of 10 plants. To minimize border effects, only the middle 16 plants of each plot were measured.
Soil and Plant Sampling
Just before applying the treatments, composite soil samples (eight cores) were randomly collected from each block at two depths (0–20 and 20–40 cm) from an area close to the tree plantation line, <30 cm from the tree center at each side. To characterize the original topsoil chemical properties, additional soil samples (0–20 cm) (composites of eight cores per block) were also collected from an area <30 cm from the center of the interrow, because this area was the least disturbed (in eucalypt plantations, generally only the planting line is plowed). At this time, leaf samples were also collected from each block; each sample consisted of a composite of the 25 youngest and fully developed leaves taken randomly from the whole crown (6-mo-old plants) or its lower third (12-mo-old plants). To evaluate N treatment effects on LNC, 6 mo after establishing the experiments 25 leaf samples were collected from the lower third of the crown of each plot.
Laboratory Analyses
Soil samples were dried in a forced-air oven at 40°C, and ground to pass a 2-mm sieve. Samples collected from the tree line were analyzed for exchangeable NH4, NO3, NMIN, and organic matter. Exchangeable NH4 was extracted with 2 M KCl using a solution/soil ratio of 5:1, and determined by colorimetry using the modified indophenol blue method described by Mulvaney (1996). In this method, sodium salicylate is used instead of phenol. Nitrate was extracted with a saturated solution of CaSO4 using a solution/soil ratio of 2.5:1, and determined with a specific activity ion electrode (Model 48680–00, Hach Co., Loveland, CO). A solution of Al2(SO4)·18H2O, H3BO3, Ag2SO4, and NH2HSO3 adjusted the ionic strength and eliminated interferences (Gelderman and Beegle, 1998). Mineralizable N was estimated as the increase in soil NH4 content after 7 d of anaerobic incubation at 40°C, as described by Keeney (1982). Soil organic matter (SOM) was determined by digestion with K2Cr2O7 and H2SO4 without external heat (Walkley and Black, 1934), and the Cr+3 produced was measured by colorimetry (Sims and Haby, 1970). Soil pH was measured in water with a combined specific activity ion electrode (Orion Model 91–0, Thermo Fisher Scientific, Waltham, MA). Phosphorus was determined by the Bray-1 method (Bray and Kurtz, 1945) because this is the standard method used in eucalypt plantations in Uruguay. Exchangeable bases (Ca, Mg, K, and Na) were extracted with NH4OAc buffered at pH 7 (Thomas, 1982), Ca and Mg were determined by atomic absorption spectrometry, and K and Na by emission spectrometry. Leaf samples were dried in a forced-air oven at 60°C, ground to pass a 1-mm mesh sieve, and analyzed for total Kjeldahl N (Bremmer and Mulvaney, 1982).
Nitrogen Availability Indices
Soil and plant samples collected just before N application were used for NAI estimation as follows: one NAI was based on plant samples (LNC only) and four NAI on soil samples taken from the tree line: soil organic matter (SOM), NMIN, NO3, and NH4 plus NO3. All NAI were averaged by site (means of three replications), and subsequently only NAI means were used.
Volume Yield Determination
The diameter at breast height (dbh) over bark, measured at 1.2 m from the ground, and the height of each tree (h) located within the growth evaluation area were recorded 12 mo after N application. The volume of each tree (v) was then estimated as
 | [1] |
This formula is normally used by other researchers in Uruguay to estimate volume, and 0.45 is a suitable shape factor for eucalypts (R. Methol, INIA, Uruguay, personal communication, 2002). Finally, the total yield volume per plot (V) was computed as the v summation of all trees effectively present within the growth evaluation area (of the 16 originally planted) and expressed in cubic meters per hectare.
Estimation of Predicted Maximum Volume and Volume of Control Plots
Four regression models were evaluated at each site-year to describe the relationship between V (Y) and applied N (X). Two models were segmented (linear-plateau and quadratic-plateau), and the other two were the linear and the quadratic models. The model with the highest coefficient of determination (R2) value was selected and then used to estimate the predicted maximum volume (MAXV). When one of the segmented models was selected, MAXV was equated to the plateau of the model. When the quadratic model was selected, MAXV was estimated by equating the first derivative of the response equation to zero, solving for X, substituting this value of X into the response equation, and solving for V. When the linear model was chosen, MAXV was estimated by taking X as the maximum N rate applied (to each particular site) and solving for V. If none of the segmented models converged or if the R2 of any of the four tested models was <0.25, MAXV was estimated as the mean V of all fertilized plots within each site-year. These models were estimated using available data from the three replications. The volume of control plots (CONTV) was always estimated as the average volume of unfertilized plots within each site-year.
Evaluation of Volume Response to Postplanting Nitrogen Applications
Two expressions were used to assess volume response to N: absolute volume increase (AVI) and relative volume (RV):
 | [2] |
 | [3] |
The rationale for using two response expressions was the uncertainty concerning the most appropriate expression (Blackmer et al., 1989) and the differences in the way each expression characterizes responses. Absolute volume increase is expressed in volume units (m3 ha–1) and could be used to assess the economic significance of the observed response (Smethurst et al., 2001). This index, however, could be affected by the uncontrolled factors that influence volume at each site. An equivalent expression (absolute yield increase) has been used by Mallarino and Blackmer (1992) to estimate critical levels for P in corn (Zea mays L.). The relative volume, however, is independent of these uncontrolled factors, and its equivalent expression (relative yield) has been extensively used as an index of yield response (Blackmer et al., 1989; Fox et al., 1989; Binford et al., 1992; Mallarino and Blackmer, (1992); Meisinger et al., 1992; Klausner et al., 1993). This index tends to minimize yield variability due to factors other than N availability when data from different site-years are pooled (Binford et al., 1992).
Estimation of Critical Levels
One critical level was determined for each NAI by relating each NAI (X) with RV. In turn, two segmented models (linear-plateau and quadratic-plateau) were evaluated to describe these relationships, and the model with the highest R2 was selected. The critical level was defined as the X value associated with the intersection of both segments of the selected model. The ability of this critical value to separate responsive from unresponsive sites was used as the criterion for selecting the most promising NAI. The critical value was then plotted on the graph relating the NAI being evaluated with AVI. Because this relationship had not been used to estimate the critical value, the capability of this value to separate responsive from unresponsive sites on the new plot was used as an additional criterion to select NAI.
Statistical Analysis
Analyses of variance were performed for data from each site and across sites (when possible) by using the GLM procedure of the SAS statistical package (SAS Institute, 1990). The treatment sum of squares was partitioned into three nonorthogonal contrasts—"zero vs. the rest," linear, and quadratic—at most sites. Linear and quadratic regression models were fit by using the GLM procedure of the SAS statistical package (/). Nonlinear regression models (linear-plateau and quadratic-plateau) were fit by using the SOLVER procedure (Wraith and Or, 1998) in the tools of Microsoft Excel 2002 software (Microsoft Inc., Redmond, WA).
|
RESULTS AND DISCUSSION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Effect of Nitrogen Rate at Planting on Volume Response to Nitrogen
The N rates applied at planting by plantation managers (before these experiments were installed) varied from 0 to 50 g N plant–1 (Table 2) as a consequence of the variability in N management that exists among the producing companies in Uruguay. We evaluated the possible effect of differential N fertilization at planting on growth and growth response to N applied 6 or 12 MAP by analyzing the graphic relationships and correlations between the N rate at planting and two plant indices that were not affected by the N treatments, CONTV and LNC. We also analyzed the relationship between the N rate at planting and the two expressions of volume response (RV and AVI). The only statistically significant result found was in the E. grandis group (all sites included), where the correlation coefficient between N rate applied at planting and AVI was high and significant (r = 0.7385, P = 0.0580). This result was highly influenced by the two sites that received no N fertilizer at planting (Sites 1 and 2, Table 2); they showed the highest AVI but low LNC values. This relatively low impact of N rates at planting on N response was expected; at most sites, these N rates were similar and low (exceptions were Sites 1, 2, 9, and 10) (Table 2).
Evaluation of Volume Response to Nitrogen by Analyses of Variance
In four afforested E. grandis sites fertilized 6 MAP with the same rate (Sites 1–4), V increased significantly due to N fertilization in only one of the four sites (Site 2) (Table 3
). A combined ANOVA across these sites did not show a significant V increase due to N, although the significance of the "zero vs. the rest" contrast was close to the probability limit of 0.10. Site 7, another afforested E. grandis site fertilized 6 MAP, was analyzed separately because it received higher N rates than the previous sites, but it did not respond significantly to N either (Table 3). The same lack of response was observed in Sites 5 and 6, which were installed in reforested E. grandis plantations and fertilized 6 MAP with higher N rates than the afforested sites (Table 3). Therefore, the lack of response of these three sites fertilized 6 MAP with higher rates seems to indicate that the lack of response observed in the other four E. grandis sites was not caused by N rates that were too low to increase volume, as we thought during the first research year.
|
The results of a combined ANOVA across sites, including all E. grandis and E. globulus sites that received the same set of treatments, showed a statistically significant volume response to N (Table 3). In this analysis, "zero vs. the rest" was the most significant contrast, implying that the shape of the average response to N in both species fertilized 6 MAP was of the linear-plus-plateau type. A model of this type fitted to the volume data (not shown) also indicates that the mean optimal rate (critical level) was very close to the first applied rate. This result is important, because it shows that the range of rates used in these experiments was adequate to explore the potential N response of both species at this growth age.
The only E. grandis experiment with N applied 12 MAP (Site 8) did not show a significant volume response to N (Table 4 ). A similar lack of response was observed in three E. globulus sites fertilized 12 MAP (Sites 16, 17, and 18; Table 4). These sites were also evaluated together, but the result of this combined ANOVA was not significant (Table 4). In three other E. globulus afforested sites fertilized 12 MAP, N was incorporated into the soil or broadcast on the soil surface as urea or NH4NO3, with the equipment normally used in some commercial plantations. Irrespective of the N source or application method, however, no significant volume response was observed on these sites (Table 4). Apparently, the lack of response observed in the previous experiments was not caused by N application techniques that differed from those normally adopted in some commercial plantations.
|
In some experiments, the lack of V response to N could also have been produced by the high experimental variability, which is shown by the high CV values obtained at many sites, which is not unusual with eucalypts (Tables 3–5
). As an example, Site 14 had the highest V increase in its group (Table 3) but a high CV, thus there was no significant response to N. According to this interpretation, an increase in the number of replications could have produced significant ANOVA results at some sites; this possibility is supported by the highly significant result obtained for the combined ANOVA across both species fertilized 6 MAP (Table 3). Also, irrespective of statistics, in 16 out of 20 sites, CONTV was numerically lower than both the mean yield of the fertilized treatments (data not shown) and MAXV (Table 4).
|
Variability of Nitrogen Availability Indices among Sites
In contrast to most agricultural soils, NH4 concentrations were higher than NO3 concentrations in all but three experiments (Table 5). This result was expected because NH4 is the prevalent form of mineral N in soils under eucalypts (Adams and Attiwill, 1986; Turnbull et al., 1996). It has also been suggested that eucalypts prefer to absorb NH4 over NO3 (Shedley et al., 1995; Garnett and Smethurst, 1999; Garnett et al., 2003). These two mineral forms were present in low concentrations at most sites, suggesting that they would not be promising NAI for eucalypts. Besides, other evaluated NAI, such as LNC or NMIN, showed greater differences among sites.
Clear LNC differences were observed between the two species, with E. grandis generally showing higher values than E. globulus in trees of similar age (Table 5). This result coincided with previous reports indicating that E. grandis usually has the highest leaf macronutrient concentrations of all eucalypt species (Haag et al. [1976], cited by Silveira et al., 2000). It should also be noted that the LNC values observed in our experiments for both species were at the high end of the range reported by Judd et al. (1996). This can be an indication of the relatively higher fertility of the soils used for eucalypt plantations in Uruguay compared with most other soils used for eucalypt production.
Relationship between Nitrogen Availability Indices and Volume Response Expressions
Five NAI were initially evaluated, but the following discussion is restricted to the NAI selected on the basis of their relationship with N response indices. All possible relationships, however, including those not shown here, can be recreated with the information supplied in Table 5.
Eucalyptus grandis Fertilized Six Months after Planting
A reasonably good relationship was observed among LNC and RV in this group of sites (Fig. 1a
). The R2 value of the linear-plateau model fitted to the data was 0.57; this value could be considered acceptable when the empirical nature of the model is taken into account. This value was also within the range 0.26 to 0.52 obtained by Mallarino and Blackmer (1992) for the relationships between corn relative yield and several P soil tests; the linear-plateau and quadratic-plateau models were used to describe the relationship. This critical value correctly separated responsive from unresponsive sites, except for the site that was marginally below the nonresponse zone in spite of having a LNC value greater than the critical level (Fig. 1a). Moreover, this critical value correctly separated responsive from unresponsive sites when plotted on top of the relationship between LNC and AVI (Fig. 1b), which had not been used to estimate the critical value. This result reconfirms LNC aptitude to screen sites according to soil N availability. The critical value of 34.6 g N kg–1 was higher than 28.0 g N kg–1, a value reported by Herbert (1996) as the optimum foliar N concentration for E. grandis in South Africa.
|
Our results suggest that LNC is a promising index to evaluate soil N-supplying capability in E. grandis plantations fertilized 6 MAP. Nevertheless, the critical level estimated in this work should be confirmed on additional sites containing areas fertilized 12 MAP, for evaluation of LNC's ability to predict N response at this different N application timing.
Eucalyptus globulus Fertilized Six Months after Planting
For this group of sites, only NMIN was related to RV and AVI. The relationships were linear, however, which could mean that all NMIN values were lower than that of a hypothetical critical level for this group. These relationships are shown in Fig. 2
with those corresponding to E. globulus sites fertilized 12 MAP. The linear regressions are not shown, but both were statistically significant. From this result, all sites should have responded to N, which partially disagrees with ANOVA results by site (only two of the six sites responded significantly to N) but agrees with the significant result of ANOVA across sites (Table 3).
|
|
Eucalyptus globulus Fertilized Six and Twelve Months after Planting: Uncorrected Indices
We also tried to identify NAI that were independent of the timing of N application. When all E. globulus sites were grouped, a good relationship between NMIN and RV was found, except for one site (Fig. 2). When this site was eliminated, the estimated critical value was 109 mg N kg–1, clearly separating responsive from unresponsive sites (Fig. 2a and 2b).
The low SOM and NMIN values for Site 16 (the eliminated experiment) are expected for this soil type (very sandy, acidic, and with low cation exchange capacity). The LNC, however, was high (23.7 g N kg–1) and above the critical range estimated for E. globulus sites fertilized 12 MAP, coinciding with the negative volume response to N observed at this site. Therefore, the logical conclusion seems to be that regardless of the low SOM and NMIN values, Site 16 had an adequate N supply, as properly assessed by LNC analysis. The source of N supply in this case, however, is unknown but could be a result of an unintentional application of a high N rate at planting not reported by the plantation manager.
Because there was no clear difference in LNC values between the two age groups of E. globulus (mean LNC was 21.9 and 21.3 g N kg–1 in the 6 and 12 MAP sites, respectively), we also evaluated the use of this index as a common test for the two fertilization times. Leaf N concentration, however, did not discriminate between responsive and unresponsive sites in this new grouping, and the critical level was not determined (Fig. 4a ). This result was expected because sites fertilized 6 MAP already showed a poor relationship between LNC and N response.
|
The advantage of using LNC–NMIN is that it integrates the information of soil and plant analyses. Hence, we can expect that if two hypothetical sites have the same LNC, the chances for N response would be smaller in the site with a higher NMIN. The relationships between LNC–NMIN and the two relative indices of V response clearly improved with respect to when LNC was used. One indicator of this improvement was the double increase in the R2 value of the model selected to describe the relationships, compared with when LNC was used (Fig. 4c vs. 4a). Moreover, the LNC–NMIN critical value separated reasonably well the sites with and without N response, and it also divided responsive from unresponsive sites when AVI was used (Fig. 4d).
As expected, the only exception to this trend was Site 16, because NMIN had already given wrong information on N response probability. Therefore, Site 16 was not considered to compute LNC–NMIN or to estimate the model presented in Fig. 4c. A LNC–NMIN value was computed for Site 16, however, using the average of all other sites, and plotted with a different symbol (Fig. 4a–4c).
The effect of using LNC–NMIN instead of LNC was relatively more important for sites fertilized 6 MAP, the group of E. globulus sites where LNC had previously shown a very poor relationship with RV and AVI. Two of these sites showed little response to N in spite of their low LNC values, but presented high NMIN values. Therefore, LNC–NMIN increased relatively more at these two sites, thus improving the overall relationships with RV. Another effect of this correction was that some sites fertilized 12 MAP that tested high in both LNC and NMIN moved toward the right end of the x axis, correctly indicating their very low probability of N response. In fact, N response was negative in one of these locations. In short, the LNC–NMIN index was promissory and could be used as a common NAI for E. globulus sites fertilized 6 or 12 MAP. More sites would be needed, however, to verify these trends and confirm the critical range.
|
CONCLUSIONS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
The N supplying capability of the soil could be successfully evaluated by NAI. Of all NAI evaluated in this work, LNC showed the strongest relationship with volume response to N, both in E. grandis fertilized 6 MAP and in E. globulus fertilized 12 MAP. In both groups, the selected index correctly separated responsive from unresponsive sites. This result suggests that, at least in Uruguay, leaf analysis based on single N concentrations can guide postplanting N applications in eucalypts.
Another promising NAI identified in this work was NMIN, which, except for one site, showed a clear relationship with the volume response to N when all E. globulus sites were grouped, irrespective of their fertilization timing. In this group, however, LNC did not show a good relationship with N response, but a combined plant and soil N index, the LNC adjusted by NMIN, was also related to the volume response to N.
Overall, this study suggests that the use of selected NAI can improve the prediction of volume response to N applied after planting in both of the evaluated eucalypt species. Nevertheless, due to the low number of experiments in each group, the critical levels identified in this work should be confirmed in future studies.
|
ACKNOWLEDGMENTS |
|---|
This research was financially supported by CSIC (Scientific Research Council, University of the Republic of Uruguay), and by the following plantation companies: Eufores, Caja Bancaria, Colonvade, Forestal Oriental, and Grupo Chileno Copihue.
|
NOTES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Received for publication August 18, 2006.
|
REFERENCES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
