INTRODUCTION
Pressure on land and soil resources due to rapidly growing populations is a serious problem across sub-Saharan Africa (SSA). In some regions, arable land has become scarce and soil degradation continues due to the shortening of the fallow period. Consequently, agricultural intensification and sustainable nutrient management practices are essential (Giertz et al., Reference Giertz, Steup and Schönbrodt2012; Hauser et al., Reference Hauser, Nolte and Carsky2006; Musinguzi et al., Reference Musinguzi, Ebanyat, Tenywa, Basamba, Tenywa and Mubiru2016; Vanlauwe and Giller, Reference Vanlauwe and Giller2006).
In peri-urban and urban vegetable production conditions in SSA, farmers tend to apply high quantities of mineral fertilizers to obtain high fresh biomass yields (Chianu et al., Reference Chianu, Chianu and Mairura2012; Drechsel and Dongus, Reference Drechsel and Dongus2010). Increasing economic costs and environmental concerns related to nitrogen (N) fertilization have made N use efficiency critically important for sustainable vegetable production while minimizing environmental degradation. However, N requirements, N effects on growth and development characteristics of indigenous vegetables, such as jute mallow (Corchorus olitorius L.; Tiliaceae family) have not received much scientific attention, and there is a need of developing such specific studies for this species.
The critical N concentration (Nc) is the minimum N concentration required to achieve maximum crop growth. Taking wheat as an example, Justes et al. (Reference Justes, Mary, Meynard, Machet and Thelier-Huche1994) developed a statistical method to determine the critical plant N concentration at different times during the growth period. The method has been successfully applied in temperate climatic conditions for several C3 and C4 cereal crops (Lemaire and Meynard, Reference Lemaire, Meynard and Lemaire1997; Lemaire et al., Reference Lemaire, Jeuffroy and Gastal2008; Plénet and Lemaire, Reference Plénet and Lemaire2000; Van Oosterom et al., Reference Van Oosterom, Carberry and Muchow2001; Yue et al., Reference Yue, Meng, Zhao, Li, Chen, Zhang and Cui2012; Ziadi et al., Reference Ziadi, Bélanger, Claessens, Lefebvre, Cambouris, Tremblay, Nolin and Parent2010) and for other tropical field crops, such as potato (Solanum tuberosum L.) (Bélanger et al., Reference Bélanger, Walsh, Richards, Milburn and Ziadi2001), sugarcane (Saccharum officinarum L. (Oliveira et al., Reference Oliveira, Gava, Trivelin, Otto and Franco2013), tomato (Lycopersicon esculentum Mill.) (Tei et al., Reference Tei, Benincasa and Guiducci2002), cotton (Gossypium hirsutum L.) (Xiaoping et al., Reference Xiaoping, Wang, Wang, Wenqi and Zhiguo2007) and corn (Zea mays L.) (Ziadi et al., Reference Ziadi, Brassard, Bélanger, Cambouris, Tremblay, Nolin, Claessens and Parent2008). A species-specific critical N dilution curve can be used in models to predict the crop N response and fertilizer N requirements (Greenwood, Reference Greenwood2001). These models can then be used for optimizing N fertilization and reduce the risk of nitrate leaching (Le Bot et al., Reference Le Bot, Adamowicz and Robin1998).
Our study aimed to delineate the critical N curve on whole jute mallow plants in order to diagnose plant N status and crop N requirements and in turn improve fresh biomass yield and N use efficiency.
MATERIAL AND METHODS
Experimental site
The field experiments were conducted during the 2010 and 2011 dry seasons (from August to December) at the experimental research site of the University of Abomey-Calavi, 10 km north-west of Cotonou, Benin (06°25′N, 02°19′E, 15 m a.s.l.). The field used in 2011 was adjacent to that used in 2010. The site has a bimodal rainfall pattern characterized by one long wet season from April to mid-July and one short wet season from September to October, with an average annual rainfall of 1150 mm, and a mean annual temperature of 27.5 °C. The climatic conditions recorded during the study are shown in Figure 1. The soil at the site is classified as an acidic Acrisol (Ultisol) (FAO, 1998). Soil samples in the top 0–0.20 m layer were collected before transplanting, and analysed for soil-nutrient status (Dabin, Reference Dabin1967). Chemical properties and texture in 2010 and 2011 were respectively: pH in H2O = 5.48 and 5.00; N total (Dumas) = 0.053% and 0.070%; organic carbon = 0.83% and 0.74%; clay content (%) = 21 and 9; silt content (%) = 3 and 3; sand content (%) = 77 and 88. In 2010, available phosphorous (Olsen modified by Dabin) was 47.6 ppm, and cation exchange capacity (hexamine cobalt chloride) was 11. The soil was coarse sandy soil derived from coastal sedimentary formations (Terre de barre) with very low total N content (Gaiser et al., Reference Gaiser, Igue, Weller, Herrmann, Graef, Lawrence and von Oppen2000). Before the trials, broadleaved weeds (Hyptis suaveolens and Schrankia leptocarpa) and sedges (Cyperus spp.) had been growing in the experimental plots for more than 1 year.
Figure 1. Monthly average rainfall and temperature in 2010 and 2011 at Abomey-Calavi, South Benin.
Experimental design and crop management
The plots were planted with a popular West African indigenous jute mallow cultivar (Eleti ehoro). The plant is an erect annual broadleaf that grows up to 2 m tall and is usually highly branched, with a tap root system; the stems are reddish, fibrous and tough, the leaves are ovate, the inflorescence is made up of yellow solitary flowers and the fruits are cylindrical capsules (Supplementary figure S1 available online at https://doi.org/10.1017/S0014479717000230). The crop reaches maturity between 3 and 4 months and harvest dates are not associated to maturity. Fresh leaves are cut (harvested) around 28–60 days from sowing.
A randomized complete block design was applied with four replications and six levels of N fertilization: 0, 30, 60, 120, 180 and 240 kg N ha−1. The fertilizers were incorporated at a depth of 5 cm around the plants till 15 days after transplanting (DAT), and at the same depth between the rows for the other N applications. More specifically, N was top dressed in the form of urea and incorporated into the soil at 8, 18, 28 and 38 DAT, 25% of the total N amount to each supply. P2O5 (20 kg ha−1) as triple superphosphate (46% P2O5), and K2O (33 kg ha−1) as K2SO4 (50% K2O), were surface broadcast and incorporated once at transplanting. Each plot size was 6.12 m² (5.1 × 1.2 m).
The seeds were immersed for 10 h in water at 75 °C, then in water at 25 °C for 10 min to suppress dormancy. In both years, 24 day-old seedlings 100–150 mm tall at the 3–5 leaf stage were transplanted in the plots at a density of 11 plants m−1 (i.e. 0.30 × 0.30 m plant spacing). Transplanting and a sparse plant density were used to ensure maximum plant growth. The sowing dates were September 20 in 2010 and August 26 in 2011.
The irrigation management was done according to the evapotranspiration (ETP) using Penman–Monteith approach (Allen et al., Reference Allen, Pereira, Raies and Smith1998) during the dry season, i.e. 3.5–4.0 mm on average. The daily water supply was around 6 mm, split twice, because of the sandy texture of the soil in order to avoid water stress and maximize N uptake. Weeds were removed by hand and insects were controlled by applying the insecticide Cypercal® (cypermethrin) once a month at a rate of 20 mL in 16 L of water.
Data collection
Measurements were conducted weekly on five plants in the central row of each plot, at 14, 21, 28, 35, 42 and 49 DAT. After 49 DAT, measurements were not taken as leaf production decreased significantly and plant lignification increased. Whole plants were cut at ground level using pruning shears, and separated into stems, leaves and reproductive organs. Fresh biomass was weighed and subsamples were oven dried. After drying to constant weight, subsamples were ground to a fine powder for total N analysis using the Kjeldahl method (Bremner, Reference Bremner, Evans, Ensminger, White and Clark1965). In 2010, the leaf area was measured using an LAI (leaf area index) meter (Li3000, Licor, Lincoln NE, USA).
Data analysis
The dilution curve for the Nc was determined as described by Justes et al. (Reference Justes, Mary, Meynard, Machet and Thelier-Huche1994) and Xiaoping et al. (Reference Xiaoping, Wang, Wang, Wenqi and Zhiguo2007). For each experiment and sampling date, amounts of above-ground dry matter (DM) and N concentrations were compared among the N treatments. DM content data were subjected to the analysis of variance (ANOVA) using the Student–Newman–Keuls test (SAS Institute, 2001). The LSD values were presented when there was a significant treatment effect (p ≤ 0.50). For each date, the theoretical critical-N concentration point was defined as follows: (i) each N treatment was specified by two variables, DM and total N concentration; (ii) the data on N deficient growth conditions were fitted by a simple linear regression; (iii) the data on N excess growth conditions were used to obtain the maximum DM; (iv) the theoretical critical-N concentration point was obtained by the calculated maximum aboveground biomass and its total N concentration as the ordinate of the maximum aboveground biomass in a simple linear regression; (v) the power regression equation that fitted these theoretical critical-N concentration points was determined.
The Nc during the vegetative and flowering plant growth stage is accurately represented by a potential allometric function (Equation 1) and is referred to as a critical N dilution curve (Greenwood et al., Reference Greenwood, Lemaire, Gosse, Cruz, Draycott and Neeteson1990; Greenwood et al., Reference Greenwood, Gastal, Lemaire, Draycott, Millard and Neeteson1991; Justes et al., Reference Justes, Mary, Meynard, Machet and Thelier-Huche1994):
$$\begin{equation}
{N_c} = {\rm{ }}a{W^{ - b}}
\end{equation}$$
where W is the total DM expressed in Mg DM ha−1; Nc is the total N concentration in shoots expressed in g kg−1 DM, and a and b are estimated parameters. Parameter a represents the N content in the total shoot biomass for 1 Mg DM ha−1, and parameter b represents the coefficient of dilution describing the relationship between the N concentration and shoot biomass (Ziadi et al., Reference Ziadi, Bélanger, Claessens, Lefebvre, Cambouris, Tremblay, Nolin and Parent2010). For each year and each sampling date, data points for which N does not limit shoot growth or is not in excess were selected. The highest shoot biomass (p ≤ 0.05) obtained with any rate of N fertilization and the corresponding N concentrations were identified and selected (Greenwood et al., Reference Greenwood, Lemaire, Gosse, Cruz, Draycott and Neeteson1990). If the highest shoot biomass was obtained with two or more N rates, the lower was selected. These date points were then used to determine the relationship between Nc and shoot biomass using an allometric function.
Equation 1 can be rearranged to define the N nutrition status of a crop as an N nutrition index (NNI) (Fletcher and Chakwizira, Reference Fletcher and Chakwizira2011) (Equation 2):
$$\begin{equation}
NNI = Nt/Nc
\end{equation}$$
where Nt is the observed N concentration in the crop. NNI > 1 means endogenous N supports maximum biomass production, suggesting that the N supply at root level is in excess of the plant demand. NNI < 1 means that N uptake limits plant growth, probably because insufficient N is available at the root level. NNI can thus be used as a basis for taking decisions about field N-fertilization, e.g. N fertilizer should only be applied when NNI < 1 (Lemaire et al., Reference Lemaire, Jeuffroy and Gastal2008). Further, NNI can be used for site-specific N fertilizer recommendations.
RESULTS AND DISCUSSION
Shoot dry biomass and critical N concentration
Mean DM of jute mallow ranged from 0.05 to 5.58 Mg ha−1 depending on the N level, DAT and year (Table 1). DM differences across the years were highly significant (p < 0.01) at each harvest date except at 21 DAT. DM was more than 1.5 as high in 2011 than in 2010 except at 21 DAT, and the highest DM was observed at 49 DAT in both years (Table 1). At 49 DAT, high DM was recorded in the 120 N and 180 N treatments, and then dropped in 240 N. For DM and N concentration in DM, there were no significant interactions for year versus N rates for all sampling dates. N concentrations in DM decreased steadily during the growing season, and a higher N application rate generally resulted in higher plant N concentration in 2010 and 2011 (Figure 2).
Table 1. Effect of different N application rates on jute mallow dry matter (Mg ha−1) on different sampling dates in number of days after transplanting (DAT) in 2010 and 2011.
Note: DF = degree of freedom; significance level: (p < 0.01) = highly significant, and (p ≤ 0.05) = significant.
Figure 2. Changes in N concentration (g kg−1 DM) at different sampling dates (in number of days after transplanting) and at different N application rates in 2010 (a) and 2011 (b). The vertical bars represent LSD values (p ≤ 0.05) at each sampling date.
The decline in N concentration with increasing biomass observed is due to the fact that the N concentration in plants usually decreases as plants grow taller due to the N dilution phenomenon. The fraction of total plant N associated with photosynthesis decreases with a concomitant increase in the N fraction of structural and storage components (Ziadi et al., Reference Ziadi, Brassard, Bélanger, Cambouris, Tremblay, Nolin, Claessens and Parent2008).
Among all sampling dates and years, only six data points out of 12 met the previously defined statistical criteria and provided an estimation of the Nc for a given DM (Table 1). The DM ranged from 0.17 to 6.01 Mg ha−1 for the combined leaf, stems and reproductive organs representing the whole plant (Table 2). These points were used to estimate the parameters of the critical N dilution curve for pooled year (2010–2011) as the interactions between N rates and years were not significant for DM and N concentration in whole plant (Figure 3). The model shown in Figure 3 accounted for 99% of the total variance.
Table 2. Dry biomass of jute mallow leaves, stems and reproductive organs at the FSA site in 2010 and 2011.
Different letters indicate statistically significant differences between N application rates according to the F test (p ≤ 0.05).
LSD = least significant difference.
ns = not significant.
*p ≤ 0.05.
**p < 0.01.
Figure 3. Critical N dilution curve in jute mallow. Pooled data (2010–2011).
The shape of the critical N dilution curve obtained in this study was specific to whole plant jute mallow. Although jute mallow is a C3 plant (Palit and Bhattacharyya, Reference Palit and Bhattacharyya1984), the a and b coefficients were lower than the values of the generic C3 species reported by Greenwood et al. (Reference Greenwood, Lemaire, Gosse, Cruz, Draycott and Neeteson1990) (a = 5.70 and b = 0.50). The lower a value may be due to the fact biomass of less than 1 Mg ha−1 during the early stage of growth was included when plotting the critical N dilution curve. The low value of parameter b could be associated with the relatively low biomass partitioning to the stem (Debaeke et al., Reference Debaeke, Van Oosterom, Justes, Champolivier, Merrien, Aguirrezabal, González-Dugo, Massignam and Montemurro2012). The lower N concentration in relation to other C3 species obtained in this study may have been due to the fact that the DM used to plot the critical curve included the stem tissues in addition to the leaves, with an increasing proportion of structural and storage materials that contain little nitrogen. Indeed, the proportion of cell wall constituents (cellulose, lignin) increases during plant growth and the protein and related N concentration decreases compared to that of leaves (Justes et al., Reference Justes, Mary, Meynard, Machet and Thelier-Huche1994). The dilution curve parameters obtained in this study were different from those determined for other C3 crops (a = 4.57 and b = 0.42 for potato by Bélanger et al. (Reference Bélanger, Walsh, Richards, Milburn and Ziadi2001); a = 4.53 and b = 0.42 for sunflower by Debaeke et al. (Reference Debaeke, Van Oosterom, Justes, Champolivier, Merrien, Aguirrezabal, González-Dugo, Massignam and Montemurro2012); a = 4.69 and b = 0.53 for linseed by Flénet et al. (Reference Flénet, Guerif, Boiffin, Dorvillez and Champolivier2006); a = 4.07 and b = 0.38 for annual ryegrass by Marino et al. (Reference Marino, Mazzanti, Assuero, Gastal, Echeverria and Andrade2004); a = 4.30 and b = 0.13 for cotton by Xiaoping et al. (Reference Xiaoping, Wang, Wang, Wenqi and Zhiguo2007)). This difference in critical N dilution curves seems to be the result of specificities in the morphology and N uptake of jute mallow. In this study, the critical N dilution curve was defined for the main cultivar of jute mallow on extended soil called Terre de barre formed on tertiary sediment along the Gulf of Guinea in the degraded coastal savanna agroecology of West Africa stretching from Accra (Ghana) to Lagos (Nigeria). It will be necessary to investigate whether the same curve can be used for other jute mallow varieties, and under different pedological, climatic and cropping conditions.
For each plant component (leaf, stem and reproductive organs) the highest DM was recorded at 49 DAT in both years (Table 2). At 49 DAT, leaf dry matter (LDM) ranged from 0.71 to 2.06 Mg ha−1 in 2010 and 1.12 to 2.33 Mg ha−1 in 2011. Application of identical N rates resulted in different N concentrations in DM in 2010 and 2011 (Table 3). The leaf N concentration range was 3.3–6.23% in 2010 and 3.7–6.04% in 2011. The stem N concentration range was 0.5–3.3% in 2010 and 1.03–3.57% in 2011. There were significant interactions for year versus N rates for N concentrations in leaf and stem organs. For reproductive organs, the N concentration range was 1.08–3.40% in 2010 and 1.73–4.47% in 2011. No significant interactions were found for year versus N rates for N concentrations in reproductive organs.
Table 3. N concentration of jute mallow leaves, stems and reproductive organs at the FSA site in 2010 and 2011.
Different letters indicate statistically significant differences between N application rates according to the F test (p ≤ 0.05).
LSD = least significant difference.
ns = not significant.
*p ≤ 0.05.
**p < 0.01.
During the growing season in 2010, LAI ranged from 0.2 at 14 DAT to 5.9 at 49 DAT, depending on the N application rate. Higher N application increased the LAI during the study period, and the maximum LAI was recorded at 120 N (Figure 4).
Figure 4. Time course of LAI in jute mallow in relation to N supply. The vertical bars represent LSD values (p ≤ 0.05) at each sampling date, and ns indicates non-significant for F test.
LAI increased with increasing N rate up to 120 N, which is around the critical uptake level, and a slight LAI decrease was observed at higher N rates (180–240 N). At around the critical N concentrations, a linear relationship was noted between N uptake and LAI (Figure 5), which is further evidence of the close interaction between N availability, leaf development, light interception, growth and N uptake (Tei et al., Reference Tei, Benincasa and Guiducci2002). Leaf is an essential photosynthetic plant organ highly responsive to N availability (Evans, Reference Evans1989). During the first part of the crop cycle, the general pattern of DM, the relative growth rate and N availability are directly associated with the relative size of the leaf area and to some extent with the plant metabolic components (Plénet and Lemaire, Reference Plénet and Lemaire2000). For C3 cereal crops (rice and winter wheat), Yao and Ata-Ul-Karim et al. (Reference Yao, Ata-Ul-Karim, Zhu, Tian, Liu and Cao2014) and Yao and Zhao et al. (Reference Yao, Zhao, Tian, Liu, Ni, Cao and Zhu2014) showed that leaf based Nc dilution curve could be used for assessing N status in crop plants. It warrants further research to investigate leaf based Nc dilution curve for jute mallow and assess its applicability in estimating N nutrition level. Moreover, during the vegetative growth, the leaf is the centre of plant photosynthesis function; the partitioning of DM and N-containing nutrients among different plant parts (leaf and stem) is optimized to satisfy the growth demand of leaves. Under these circumstances, the shape of the leaf Nc dilution curve is stable (Yao and Zhao et al., Reference Yao, Zhao, Tian, Liu, Ni, Cao and Zhu2014b).
Figure 5. N uptake as a function of LAI for N-fertilizer rates of 60 and 120 kg/ha (the rates around the critical dilution curve).
At 120 kg ha−1, the highest LDM was 2.06 and 2.32 Mg ha−1 in 2010 and 2011, respectively. The total DM has been reported at 15.0 Mg ha−1 (Schippers, Reference Schippers2000), while the highest biomass obtained in the present study was 6.0 Mg ha−1. The difference may be due to the sparse planting density employed in this study to measure maximum plant N uptake capacity per plant. On the other hand, small-scale African farmers usually broadcast jute mallow seeds resulting in a dense stand with limited space between plants (Fondio and Grubben, Reference Fondio, Grubben, Grubben and Denton2004).
Nitrogen nutrition index (NNI)
The N application rate had a significant effect on N uptake from 35 to 49 DAT in 2010 but not at 42 DAT in 2011 (Figure 6). When N uptake was significant, it increased with increasing N rates (mainly 120–180 N). For both years, the highest N uptake was recorded at 180 N followed by 120 N.
Figure 6. Time course of N uptake (kg ha−1) in relation to N supply in 2010 (a) and 2011 (b). The vertical bars represent LSD values (p ≤ 0.05) at each sampling date.
The NNI values ranged from 0.55 to 1.30 during the two study years (Figure 7). The data points with N applications of 0, 30 and 60 kg N ha−1 were generally less than 1.0, indicating that N limited growth. At 120 kg N ha−1, the data points were generally near 1.0, indicating that N did not limit growth. At 180 and 240 kg N ha−1, the data points were generally well above the critical N curve, indicating excessive N nutrition in both years. Consequently, 120 kg N ha−1 appeared to be the optimum N application rate for jute mallow in the conditions that prevail in southern Benin, thus validating the critical N dilution curve.
Figure 7. The N nutrition index (NNI) on six sampling dates for jute mallow under different N application rates in 2010 (a) and 2011 (b). The solid line represents an NNI of 1.0. Values ≥1 indicate non-limiting supply of N, whereas values of NNI < 1 indicate N deficiency.
The NNI identified for jute mallow ranged from 0.55 to 1.30, which is in line with the NNI values of 0.50 to 1.40 reported for winter wheat (Justes et al., Reference Justes, Mary, Meynard, Machet and Thelier-Huche1994). This confirms the ability of jute mallow to take up more N than it requires for maximum growth. Excess N can be stored in various chemical forms, including simple inorganic and complex organic compounds in plant tissues (Brégard et al., Reference Brégard, Bélanger and Michaud2000; López-López et al., Reference López-López, Aguirre-Bravo and Reich2006; Xiaoping et al., Reference Xiaoping, Wang, Wang, Wenqi and Zhiguo2007). Luxury consumption occurs when N is supplied in excess, thus highlighting the importance of appropriate N application (Xiaoping et al., Reference Xiaoping, Wang, Wang, Wenqi and Zhiguo2007).
NNI is a useful indicator for N status assessment and N fertilization management, but frequent and time-consuming measurements are required, which would not be feasible for small-scale farmers. More practical, real-time, quick and non-destructive field methods, such as a leaf colour chart, chlorophyll meter, and digital photography need to be promoted for determining NNI.
CONCLUSION
A critical N dilution curve (Nc = 3.35*DM−0.18) was obtained for jute mallow under our experimental conditions. The NNI calculated from that curve can be considered as a reliable indicator of the level of N stress for jute mallow during the growing season. N dilution curves can be used as a tool to diagnose the status of N in jute mallow during the vegetative and early reproductive phases. The relationship can also be used in the parameterization and validation of growth models to predict the N response and/or N requirements of jute mallow.
Acknowledgments
The authors thank Eugène Allissou for help collecting data, and are grateful to Jacques Le Bot, INRA, Centre de Recherche Provence-Alpes-Côte d'Azur (France) for helpful comments on the manuscript. Financial assistance for this study was granted by the European Union and the IFAD-funded RAP project (COFIN-ECG-65-WARDA).
SUPPLEMENTARY MATERIALS
For supplementary material for this article, please visit https://doi.org/10.1017/S0014479717000230.
