Agroforestry Systems 44: 21–36, 1999. 1999 Kluwer Academic Publishers. Printed in the Netherlands.
Tithonia and senna green manures and inorganic fertilizers as phosphorus sources for maize in Western Kenya C. N. GACHENGO1, 2, C. A. PALM1, *, B. JAMA3 and C. OTHIENO2 1
Tropical Soil Biology and Fertility Programme (TSBF), P.O. Box 30592, Nairobi, Kenya; Moi University, P.O. Box 1125, Eldoret, Kenya; 3 International Centre for Research in Agroforestry (ICRAF), P.O. Box 30677, Nairobi, Kenya (*Author for Correspondence: E-mail:
[email protected])
2
Key words: biomass transfer, decomposition, mineralization, nitrogen Abstract. Efforts to overcome declining soil fertility on small holder farms in western Kenya must be consistent with the reality of low utilization of inorganic fertilizers. Likewise organic inputs alone cannot supply adequate nutrients. The use of two organic resources, Tithonia diversifolia (tithonia) and Senna spectabilis (senna) leaves, and their combination with inorganic P for improving soil fertility and maize yields was investigated on a P limiting soil in Western Kenya. Treatments included: 1) control, no inputs; 2) 5 t ha–1 (dry matter) tithonia leaves; 3) 5 t ha–1 senna leaves; 4) 5 t ha–1 tithonia leaves + 25 kg P ha–1 as triple superphosphate (TSP); 5) 5 t ha–1 senna leaves + 25 kg P ha–1 (as TSP); and 6) 25 kg P ha–1 of TSP. Maize was used as a test crop. Decomposition and P and N release of tithonia and senna leaves were determined in a litterbag study. Tithonia + TSP applications tripled maize yields compared to the control, senna + TSP and tithonia sole application doubled yields, while senna sole applications did not increase yields substantially. A large residual yield was produced in the tithonia treatments in a subsequent crop. These yield results were consistent with the higher quality and faster release of N and P from the tithonia leaves compared to senna. The tithonia biomass transfer system can improve yields in the short term but has limitations because of the large amount of biomass and the associated labor requirements.
Introduction Soil fertility has decreased in large areas of sub-Saharan Africa over the past 30 years due to the continued removal of nutrients in crop harvests. Nutrients removed by harvest are usually greater than the amount returned as fertilizers. In Kenya, for example, there is an average annual net mining of 42 kg N ha–1, 3 kg P ha–1 and 29 kg K ha–1 from the soil (Smaling, 1993). There is an urgent need to increase the use of external inputs to reverse the negative nutrient balance and increase food production. This nutrient replenishment may be achieved through use of mineral fertilizers or organic inputs, or a combination of both depending on their availability and the level of nutrient deficiency. Average fertilizer use for 35 crops, including cash crops, in Kisii district of western Kenya are currently less than 17 kg N ha–1 year–1 and 12 kg P ha–1 year–1 (Smaling, 1993) compared to recommendation of 75 kg N ha–1 year–1 and 25 kg P ha–1 year–1 (FURP, 1994). Organic inputs can have fertilizer equivalency values of 50 to 100 kg N ha–1 (Ladha et al., 1988) and may meet
22 crop demands for N but are not likely to meet P demands because of the low P concentration in most organic materials (Palm, 1995). In addition to a net negative P balance, P is one of the major nutrients limiting crop production (Onim et al., 1990) and about half of the soils in the Kenyan Highlands have moderate to high P-adsorption capacity (Braun et al., 1997). Thus, a portion of the small amounts of P currently added as mineral fertilizer would be adsorbed by the iron and aluminum oxide clay minerals, decreasing availability of the fertilizer for uptake by crops. Increasing P availability on these soils would require addition of large amounts of fertilizer, an expensive venture not affordable by most resource limited farmers in the area. Although organic inputs cannot meet crop P requirements they can increase P availability by reducing the P adsorption capacity of the soil (Iyamuremye and Dick, 1996; Nziguheba et al., 1998). Organic materials differ considerably in their ability to supply nutrients to the soil and crop. These differences relate to the decomposition and nutrient release rates and patterns, which are in part controlled by the resource quality of the materials. Organic resource quality is a function of the nutrient concentration, the relative proportions of lignin, metabolic carbohydrates, cellulose, and presence of modifiers, such as polyphenolics (Swift et al., 1979; Palm and Rowland, 1997). Research over the past ten years has attempted to identify organic amendments that release nutrients in amounts similar to and in synchrony with crop nutrient demand (Giller and Cadisch, 1997). Traditional organic inputs such as crop residues and animal manures are generally of insufficient quantity and are of low quality and thus cannot meet crop nutrient demands. Due to the low P in organic materials and low rates of mineral fertilizer P applied by farmers, it may be necessary to use inorganic fertilizers to supplement organic materials in order to achieve better crop yields. The combination may lead to an increase or decrease in available nutrients in the soil depending on the quality of the organic material (Palm et al., 1997). Agroforestry programs often introduce trees in an attempt to increase biomass quality and production on farms. In western Kenya, a readily available source of plant biomass already exists in hedges bordering farms and roadsides (Bradley, 1988). Farmers prune these hedges once or twice a year to avoid competition between the hedges and crops, but only one quarter of the farmers use the prunings as soil amendments (Carter et al., 1998). Prunings from these hedges could therefore potentially serve as green manure for soil fertility management. Trees and shrubs commonly found in the border hedges include, Senna spectabilis, Markhamia lutea, Lantana camara, Tithonia diversifolia, Euphorbia tirucalli, and Acanthus pubescens (Bradley, 1988; Carter et al., 1998). Field and pot studies were conducted in western Kenya using two of these plants, Tithonia diversifolia (tithonia) and Senna spectabilis (senna) to, 1) quantify the effects of organic input quality and their combination with mineral P fertilizer on soil P availability and maize yield; and 2) to relate the
23 decomposition and nutrient release patterns of organic inputs to maize yields and available nutrients in soil.
Materials and methods Study site The study was located on Julius farm (Jama et al., 1997) in West Bunyore Division, Vihiga District, Western Kenya (lat. 0°06′ N, long. 34°34′ E, 1460 m asl). The area receives bimodal rainfall of 1800 mm per year with peaks in April through May (long rains) and September through October (short rains). The major soil of the area is kaolintic, isohyperthermic Kandiudalf. Some of the initial soil chemical properties in the surface 15 cm are as follows: pH (water) 5.3; total soil carbon 12.3 g kg –1 soil; total nitrogen 1.6 g kg –1 soil; exchangeable acidity, 0.38 cmolc kg–1 soil; KCl extractable Ca, 3.8 cmolc kg–1 soil; KCl extractable Mg, 0.83 cmolc kg–1 soil, bicarbonate-EDTA extractable K, 0.07 cmolc kg–1 soil, and bicarbonate-EDTA extractable inorganic P, 1.8 mg kg–1 soil. The soil is considered moderately P fixing according to the method of Fox and Kamprath (1970), with a soil P concentration of 0.2 mg L–1 corresponding to 310 mg P kg–1 adsorbed by the soil (Nziguheba et al., 1998). The loamy soil contains 30% clay, 24% silt, and 46% sand. Phosphorus is had been determined as the primary limiting nutrient in the soil of this farm (Jama et al., 1997). Quality characteristics of plant materials Tithonia and senna were selected for this study because of their abundance around the farms in the area. Leaves of the two species were characterized for quality parameters according to the recommendations of Palm and Rowland (1997). Fresh leaves were picked from the plants and air-dried. The leaves included the petioles, and in the case of senna they also included the rachis since senna has compound leaves. Samples were ground to pass through a 0.5 mm sieve and analyzed for total N, P and K by Kjeldahl digestion with concentrated sulfuric acid (Anderson and Ingram, 1993). Nitrogen and P were determined colorimetrically (Parkinson and Allen, 1975) and K by flame photometry (Anderson and Ingram, 1993). Lignin was determined according to the acid detergent fiber (ADF) method of van Soest (1963). Total extractable polyphenols were analyzed from air-dried material by extraction using 50% aqueous methanol. The plant to extractant ratio was 0.1 g /50 ml and phenols were analyzed colorimetrically using the Folin-Ciocalteu reagent as described by Constantinides and Fownes (1994).
24 Experimental designs and treatments. 1) Field experiment A field trial was established to compare maize yields and soil nutrient availability resulting from applications of green manures of tithonia and senna, and their combination with mineral P fertilizer. Both green manures consisted of leaves collected from farm hedges. A randomized complete block design with four replications was used. The treatments were: 1) a control with no inputs; 2) 5 t ha–1 (dry matter) tithonia leaves; 3) 5 t ha –1 senna leaves; 4) a combination of 5 t ha–1 tithonia leaves with 25 kg P ha–1 applied as triple superphosphate (TSP); 5) a combination of 5 t ha–1 of senna leaves with 25 kg P ha–1 applied as TSP; and 6) 25 kg P ha–1 of TSP. The plant materials and the TSP were broadcast and incorporated into the soil manually with a hoe to 15 cm depth in 3 m by 3 m plots. For the first crop, treatments were applied in September 1994, the beginning of the short rainy season, and maize (Hybrid 511) sown at a spacing of 0.25 m by 0.75 m. For the subsequent two crops planted in March 1995 and September 1995, maize (Hybrid 614 and 511, respectively) was sown with no application of organic materials or TSP in order to observe residual effects. Crop 2 was planted two weeks after harvesting crop 1 while crop 3 was planted two weeks after harvesting crop 2. Hand weeding was done twice in each growing season. Crop residues for crop 1 and 2 were removed at harvest to reduce confounding effects from additional organic inputs of different qualities. Soil was sampled to 15 cm depth at three and 16 weeks after application of treatments at four locations in each plot. These samples were bulked and a subsample obtained for analysis of resin extractable P (Sibbesen, 1978) and KCl extractable ammonium (Anderson and Ingram, 1993) and nitrate (Hilsheimer and Harwig, 1976). Soils were sampled only during the first crop. Maize grain and stover yields and nutrient uptake were determined for the three crops harvested in January 1995, August 1995 and January 1996. A total of 18 plants were harvested per plot in an area of 3.37 m2 leaving one border row on all sides. Total fresh weight of the cobs and stover per plot was taken in the field. Subsamples were taken to the laboratory for oven drying at 65 °C to constant weight to determine the dry matter content of the samples. Yields were estimated by multiplying the total fresh weight per plot by the dry matter factor and the results expressed per unit area. In order to compare the treatment effects of the different seasons, yields were converted to relative increase compared to the control: Yield increase (%) =
(Yieldtreatment – Yieldcontrol) × 100 Yieldcontrol
Nutrient uptake by the crop was determined as the product of the grain and stover yields multiplied by the nutrient concentration of the respective components. Samples of grain and stover for the first crop were analyzed for
25 total N and P by methods described above. The nutrient concentrations for grain and stover obtained for the first crop were also used to determine nutrient uptake for the second and third crops. It had been established earlier in a similar experiment in the same area that nutrient concentrations in the grain and stover did not differ significantly among treatments or seasons. Nutrient recovery was determined as shown below: Nutrient recovery (%) =
(Nutrient uptaketreatment – Nutrient uptakecontrol) × 100 Amount of nutrient applied to first crop
2) Decomposition experiment Decomposition and nutrient release of the leaves of tithonia and senna were determined in the field using the litterbag technique. The decomposition trial was located adjacent to the field trial. Fresh leaves equivalent to forty-five grams on a dry weight basis were placed in rigid nylon litterbags (dimensions 30 cm by 30 cm with mesh size 7 mm). This rate is equivalent to 5 t ha–1 on a dry weight basis. The leaves were evenly spread in the litterbags and the bags were buried horizontally in the soil (15 cm depth) in 1 m by 1 m plots. The plots were distributed in a randomized complete block design with five blocks and two treatments (tithonia and senna). Each plot contained five litterbags. One litterbag was collected from each plot at one, two, four, eight and 16 weeks to follow dry matter and nutrient loss. The plant materials remaining in the litterbags at each time were separated from soil and organic debris by hand and oven dried at 65 °C to constant weight. In order to correct for contamination by the mineral soil, subsamples were ashed at 450 °C for four hours. The difference between the dry weight of the decomposing leaves and its ash content gave the dry weight of leaves remaining on an ash-free basis. Subsamples of the initial plant materials (time 0) and plant materials remaining at each sampling time were analyzed for total N and P as described above. The amount of nutrients remaining in the litterbag at each sampling time was determined by multiplying the mass of leaves remaining by their N and P concentration. It has been assumed that the nutrients released or immobilized at each time is the difference between the amount of nutrients contained in the initial plant materials and the amounts in the materials at the given sampling time. The decomposition and nutrient loss constants, k, were determined by the negative, single exponential model, X = e–kt, where X is the proportion of initial leaf mass (or nutrient) remaining at each time, t, in years (Wieder and Lang, 1982). Half life (t50), the time when 50% of the leaves have decomposed (or half of the nutrients have been released) was calculated as: t50 =
–ln(0.5) k
26 3) Pot experiment After initial observations that there was little crop response to application of TSP without other nutrients, a pot trial was designed to determine which nutrient(s) among N, P and K were limiting or co-limiting maize growth on this soil. The pot experiment was a completely randomized design with four replications. Soil from the top 15 cm of the experimental field site was used. The soil was collected from outside the experimental plots since treatments had already been applied. Two and a half kilograms of soil were placed in pots into which was added the following treatments: 1) control with no inputs; 2) NPK; 3) NP; 4) NK; and 5) PK. Nitrogen was added as urea at a rate equivalent to 189 kg ha–1, P as TSP at a rate equivalent to 16.5 kg ha–1, and K as KCl at a rate equivalent to 227 kg ha–1. These rates are equal to the amounts contained in 5 t ha–1 (dry matter basis) of tithonia leaves. Maize (Hybrid 511) was sown at a rate of 3 seeds per pot. Deionized water was used to keep the pots moist throughout the study period. Plants were thinned to two per pot following germination and to one plant one month after germination. The weight of the thinned plants was taken. At two months, the aboveground parts of the maize plants were harvested, dried to constant weight at 65 °C and combined with the earlier thinnings. Data analysis Analysis of variance was conducted using the ANOVA procedure of SAS (SAS Institute Inc., 1995) to determine the effects of treatments on crop yields, nutrient uptake and recovery, and soil nutrient availability. Standard error of the difference (SED) was used to make treatment comparisons. T-tests were used to compare decomposition and nutrient loss rate constants (k) between the two plant materials. Statistical significance refers to alpha of at 0.05.
Results Organic resource quality Tithonia leaves were of higher resource quality than senna leaves based on their chemical characteristics (Table 1). The N concentrations of the leaves of both materials were higher than the critical level of 20.0 to 25.0 mg g–1, below which point net N immobilization from the soil would be expected. The P concentration of senna leaves however was below the critical level of 2.5 mg g–1 and P immobilization would be expected (A. Kwabiah, per. comm.). Lignin concentrations were less than 10% and polyphenols less than 2% for both materials, levels that would not significantly reduce decomposition rates (Palm et al., 1997).
27 Table 1. Some chemical properties of green manures used in the field trial in Western Kenya. N
Tithonia diversifolia Senna spectabilis
P
K
Mg
Ca
Lignin
Polyphenol (TAE)
———— mg g–1 plant material ————
———— % ————
37.8 29.8
6.48 9.17
3.30 2.01
45.5 16.0
4.18 1.94
18.32 16.37
1.59 1.24
TAE = Tannic Acid Equivalents.
Field experiment Grain yields Maize grain yields for the first crop and sum for the three crops were in the order: tithonia + TSP > senna + TSP = tithonia ≥ senna > TSP = control (Table 2). The control plots yielded a total of only 2.7 t ha –1 for the three crops compared to 8.6 t ha–1 for the tithonia + TSP treatment. Tithonia + TSP application tripled yields compared to the control while senna + TSP and tithonia applied alone doubled crop yields. It is important to note that TSP applied without the addition of other nutrients only increased yields above that of the control during the second crop. Yields obtained with tithonia green manure were at least 0.5 t ha–1 greater than with senna green manure for each crop, though this difference was not significant for the first crop. The combined application of tithonia + TSP increased yields by 1 t ha–1 above that of the sole application of tithonia for each of the first two crops. Yield from the combination of senna + TSP was higher than senna applied alone only for the second crop. The tithonia + TSP treatment gave higher yields than senna + TSP for the first crop but the two treatments resulted in similar grain yields in the second crop. The average grain yield for all treatments for the second crop, 2.88 t ha–1, Table 2. The effect of organic input quality and combined organic and inorganic nutrient sources on maize grain yields in Western Kenya. Yields for Crops 2 and 3 are residual effects from one time application of treatments. Treatment
Crop 1
Crop 2
Crop 3
Total yield
–1
———————————— t ha ———————————— Control Senna Tithonia Senna + TSP Tithonia + TSP TSP SED
0.8 1.5 2.0 2.0 3.2 0.8 0.3
1.4 2.0 3.3 4.1 4.3 2.2 0.3
SED = Standard error of the difference in means.
0.5 0.6 1.2 0.8 1.1 0.5 0.2
2.7 4.1 6.5 6.8 8.6 3.5 0.5
28 was higher than that of the first and third crops, 1.72 t ha–1 and 0.78 t ha–1, respectively. The second crop grew during the long rainy season with 1025 mm of rain compared to 680 mm and 600 mm of rain for the first and third crops, respectively. Tithonia + TSP application produced the highest initial and residual effect with relative yield increases of 300%, 200% and 120%, for crops 1, 2 and 3 (Figure 1) respectively. Tithonia and senna applications show the same declining trends in residual yields, though tithonia had greater residual effects for both crops than did senna. Both the senna + TSP and sole TSP treatments showed a greater increase in relative yield for crop 2 compared to crop 1 but small increases for crop 3. Grain yields showed a response to the amount of P added from the combination of the organic and inorganic sources (Figure 2). The amount of P added explained over 80% of the yield variation. In general, the points representing tithonia applications, with or without TSP, fall above the regression lines while those of senna fall below the lines. Nutrient uptake and recovery by the maize crop The N and P concentrations in the grain and stover did not differ among treatments although the N concentrations tended to be higher in the combined organic materials + TSP treatments. The average nutrient concentrations were 14.8 g N kg–1 and 2.2 g P kg–1 in the grain and 13.3 g N kg–1 and 1.3 g P kg–1 in the stover. Phosphorus and N uptake trends among treatments were generally similar to that of crop yields, being significantly greater for tithonia compared to senna and increasing with additions of P (Table 3). Nutrient recovery patterns differed for N and P (Table 3). Phosphorus recovery in the first crop ranged from as low as 1% for the TSP treatment to
Figure 1. Crop yields for the first and residual crops, expressed as the percent yield increase relative to the control in Western Kenya.
29
Figure 2. Maize grain yields in relation to P added from organic and inorganic sources for crop 1 and the sum of three crops in Western Kenya. T = tithonia application, T+ P = tithonia + TSP application, S = senna application, S + P = senna + TSP application.
a high of 31% in the tithonia treatment. The total P recovered for the three crops ranged from 4% for TSP to 79% for tithonia. Percentage recovery of the P added as tithonia green manure was almost twice that of the P added as senna green manure. Addition of TSP to the organics had little effect on P recovery when added with senna but there was a decrease in percent P recovered when added to tithonia. There were significant treatment differences in nitrogen recovery among treatments. Nitrogen recovery for the first crop ranged from 13% for senna to 50% for tithonia + TSP and 27% to 100% total recovery for the three crops. Recovery of nitrogen added as tithonia alone was almost twice that of senna alone. Addition of P to the organics more than doubled nitrogen recovery in the maize crop when added to senna and also increased nitrogen recovery, but less so, when added to tithonia. Resin extractable P and inorganic N changes in the soil The application of tithonia or senna alone did not increase resin P at week 3 compared to the control (Table 4). The combined TSP and tithonia or senna treatments gave higher resin P than the organic materials applied alone and than the control at both sampling times. The TSP applied alone had similar resin P to that of the combined applications and higher resin P than the sole organic treatments at week 3. By week 16 the TSP treatment still had higher resin P than the organic treatments but lower resin P than the combined treatments, despite lower removal of P by the maize grown in the sole TSP treatment. At week 3, soil extractable N was almost 50 kg N ha –1 higher in the tithonia
30 Table 3. Nutrients added, total aboveground nutrient uptake, and nutrient recovery by the first maize crop and sum of three crops in Western Kenya. Treatment
Control Senna Tithonia Senna + TSP Tithonia + TSP TSP SED
Nutrients added (kg ha–1)
(Nutrient uptake (kg ha–1)
Nutrient recovery (%)
Crop 1
Total 3 crops
Crop 1
Total 3 crops
N
P
N
P
N
P
N
P
N
P
000 149 189 149 189 000
00 10 17 35 42 25
039 059 086 081 126 039 012
04 06 09 08 12 04 01
117 157 236 256 311 136 028
11 15 24 25 31 13 03
NA 013 025 028 046 NA 007
NA 019 031 012 020 000 008
NA 027 063 093 103 NA 017
NA 037 079 038 046 004 019
NA = Not applicable. SED = Standard error of the difference in means.
31 Table 4. Effects of treatments on soil extractable resin P and inorganic nitrogen during crop 1 in Western Kenya. Week 3 Treatment
N
Week 16 P
N
P
–1
——————————— kg ha ——————————— Control Senna Tithonia Senna + TSP Tithonia + TSP TSP SED
27.38 40.58 75.54 43.61 63.80 33.91 09.57
05.00 06.36 08.61 13.76 14.41 17.51 03.06
27.40 21.31 24.49 32.03 22.74 22.08 04.51
06.49 06.72 06.62 11.91 10.48 08.69 00.84
SED = Standard error of the difference in means.
treatment compared to the control, while the senna treatments had 13 kg N ha–1 more than the control though this was not significant. At week 16, there were no significant treatment differences in soil extractable N. Addition of TSP with or without organics did not affect the soil inorganic N levels. Decomposition experiment Tithonia decomposed rapidly and released nutrients significantly faster than senna (Figure 3). Half of the tithonia and senna decomposed (t50) in 1.1 and 4.3 weeks, respectively. Decomposition of tithonia was accompanied by immediate release of both P and N, while that of senna was accompanied by immediate release of N but immobilization of P occurred during the first two weeks. Although both plant materials released N immediately, the pattern of release from senna was slower (Figure 3). At three weeks about 80% of N and P had been released from tithonia leaves and only 40% from senna leaves. The differences between tithonia and senna in nutrient concentrations and nutrient release rates resulted in large differences in total P and N released through decomposition during the period of crop growth, if we assume that nutrients lost from the litterbags represent mineralization and that they were available for plant uptake. Using the kN and kP nutrient release constants to calculate nutrients provided by week 3 through mineralization of tithonia and senna, tithonia would have provided an equivalent of 162 kg N ha–1 and 14 kg P ha–1 while senna would have provided 61 kg N ha–1. Net immobilization of 2 kg P ha–1 from the soil occurred in the senna treatment at week 3 (Figure 3). Between weeks 3 and 16 an additional 26 kg N and 2 kg P were provided through decomposition of tithonia and 76 kg N and 7 kg P from senna.
32
Figure 3. Decomposition and N and P release patterns for tithonia and senna in Western Kenya. kM, kN, and kP refer to the decomposition and N and P release constants per year, respectively. * indicates significant differences between species.
Pot Experiment The results of the pot study confirm that P was the primary limiting nutrient in the soil, followed by N (Table 5). Yields obtained from the treatment with N and K but without P (Treatment 4) were similar to that of the control in which no nutrients were added (Treatment 1). When P and K were supplied but not N (Treatment 5), maize yields were higher than those of the control treatment but less than those of treatments where N and P (Treatment 3) or N, P and K (Treatment 2) were supplied indicating that N was the second most limiting nutrient. The absence of K (Treatment 3) did not show a reduction in maize biomass compared to the NPK treatment. Table 5. Dry weights of two month old maize plants grown on soil treated with various combinations of inorganic N, P, K in Western Kenya. Numbers in parentheses indicate the amount of nutrient added as an equivalent kg ha–1. Treatments 1. Control 2. NPK (189:16.5:227) 3. NP (189:16.5) 4. NK (189:227) 5. PK (16.5:227) SED SED = Standard error of the difference in means.
Weights (g pot–1) 1.09 4.58 3.91 1.17 2.33 0.44
33 Discussion and conclusions Comparison of crop yields, nutrient recovery, and soil extractable P and N obtained in the field trial reflect the differences in quality of the organic materials added as green manures. The higher yields for the first crop with applications of the higher quality tithonia are due to a combination of more P and N added and faster release patterns of P and N from tithonia as compared to senna. The short-term immobilization of P and slower provision of N resulting from application of the lower quality senna no doubt negatively affected the early growth of maize in that treatment. The higher relative yield of crop 2 for the senna + TSP treatment compared to crop 1 suggests that nutrients became available later and resulted in a delayed crop response, although the same trend was not observed for senna alone. The lack of crop response to TSP alone in crop 1 is not easily explained given that P was shown to be the primary limiting nutrient. One possibility is that low soil moisture during the early stage of crop growth may have inhibited diffusion of P while the treatments with fresh organics added would have had higher soil moisture facilitating diffusion. This explanation is supported to some degree by the higher crop response to TSP alone in crop 2 when rainfall was twice that of crop 1 or 3. These trials with organic inputs illustrate the difficulty in interpreting such data because organics add several nutrients (Bouldin, 1988) and the amount, ratio and release of nutrients added from the different organic materials vary (Palm et al., 1997). It is therefore difficult to relate differences in yields to one nutrient alone unless all other nutrients are added in unlimiting quantities. The fact that crop growth in this soil was limited by P and then N rather than just P as previously determined (Jama et al., 1997) also makes it difficult to compare yields and nutrient recovery from the organic treatments, that supply both N and P, with the TSP treatment that supplies only P. The maize crop recovered a higher percentage of the P and N added from tithonia green manure compared to senna. This observation supports the idea that the quality of the organic input, not just the amount of nutrients added, affects nutrient availability patterns and crop growth. Less P and N were added with senna compared to tithonia but also the crop recovered a lower percentage of both nutrients from the senna green manure. This lower recovery of the P and N added as senna green manure in crop 1 might be a result of the late release of P and slower release of N from senna, that resulted in a lack of synchrony between crop nutrient demand and supply by the soil. The P recovery values from the treatments with organic inputs are high compared to recovery levels of 10% for the first crop and 20-30% after several crops as reported by van der Eijk (1997) for inorganic fertilizers. This high P recovery is most likely related to the low levels of P added, on the steep part of the fertilizer response curve, where more P is taken up per unit P added. The high recovery might also be an effect of the organic additions on P availability on this moderately P fixing soil. Nziguheba et al. (1998) found
34 that applications of 5 t ha–1 (15 kg P) of tithonia increased the sum of resin + NaOH inorganic extractable P when compared to additions of similar amounts of P added as TSP. The increase in P availability was related to a decrease in the P-adsorption capacity of the soil with additions of tithonia. This effect of the organics on reducing P adsorption may explain the higher resin P at week 16 in the organic + TSP treatments compared to the sole TSP treatment, despite lower P removal through crop harvest in the TSP treatment (Table 4). Nitrogen recovery values from tithonia green manure were high compared to senna and higher than the 20% or less reported for most organic inputs (Giller and Cadisch, 1995). The N recovery for the three crops, as high as 100% for tithonia + TSP, was also high compared to other studies (Ladd et al., 1983; Sisworo et al., 1990). The difference method of calculating nutrient recovery often overestimates nutrient recovery and that may, in part, explain the high values in this study. Also, the fact that P was the primary nutrient limiting crop growth and nutrient uptake may also account for the high N recoveries. Once the P limitation was overcome the crop readily took up N. This effect is shown by the dramatic increase in N recovery when P was added to the organics (Table 4). The soil extractable P and N values at week 3 of crop 1 (Table 4) generally support the above discussion on patterns of nutrient availability and yields for crop 1 with a few exceptions. The tithonia treatments that produced higher yields generally had higher levels of N in the soil at week 3 compared to the senna treatments though the resin P values were not different. The high resin P in the sole TSP treatment did not result in high maize yields (the possible reasons have been mentioned). Though the exact amounts of N and P released from decomposition by week 3 are not found in the soil, the trends in N and P availability in the soil correlate well with the patterns from the litterbag study. Immobilization of P that was ascertained through the litterbag method was not detected by a decrease in soil resin P. The large residual yields for crop 2 compared to the control are not reflected by differences in the soil N and P measured at the end of crop 1 (crop 2 was planted two weeks later). Tithonia + TSP had higher yields but lower soil extractable N and P levels at the beginning of crop 2 compared to senna + TSP, and tithonia alone had almost 1.5 the yield of the control but similar levels of N and P. This lack of correlation of crop 2 yields to soil resin P and inorganic N at the beginning of the crop suggests that other soil fractions, such as NaOH extractable P, mineralizeable N, or light fraction N and P might help improve the correlation to crop yields in low input systems (Maroko et al., 1998). In conclusion, nutrient availability in the right quantities, ratios, and at the time plants need them is important for good crop yields. Green manure from the high quality tithonia was able to provide P and N in quantities and rates sufficient to double crop yields compared to the no input control for three consecutive crops following a single application. The green manure of senna
35 which was of lower quality, based on P and N concentrations and release rates as determined by the litterbag method, increased yields but less so than tithonia and for only two crops. Addition of inorganic P fertilizer with the green manures was necessary to increase yields substantially on this P-limiting, moderately P-fixing soil. This study demonstrates that the traditional hedges in western Kenya contain an under-utilized resource in the form of plant biomass for soil fertility management. There are however limitations to the extent and sustainability of this biomass transfer system. The amounts of tithonia green manure that can be produced on farm boundaries are insufficient to cover large areas and the labor required in cutting, transporting, and incorporating the green manures is substantial. In addition, this biomass transfer system is merely transferring nutrients within the farm and not actually redressing the nutrient depletion in the soils, therefore it may not be sustainable in the long term. Acknowledgements We would like to thank Roland Buresh for his help in the design of this experiment, and Paul Smithson and others at the ICRAF Laboratory for helping in soil and plant analysis and. Nicholas Kung’u and Margaret Muthoni of KARI Muguga Laboratory for their help in the initial characterization of the plant materials. References Anderson JM and Ingram JSI (1993) Tropical Soil Biology and Fertility: A Handbook of Methods. CAB International, Wallingford, England, 221 pp Bouldin DR (1988) Effect of green manure on soil organic matter content and nitrogen availability. In: Green Manure in Rice Farming: The Role of Green Manure Crops in Rice Farming Systems, pp 151–163. International Rice Research Institute, Los Banos, Philippines Bradley PN (1988) Survey of woody biomass on farms in Western Kenya. Ambio 17: 40–48 Braun AR, Smaling EMA, Muchugu EI, Shepherd KD and Corbett JD (1997) Maintenance and improvement of soil productivity in the highlands of Ethiopia, Kenya, Madagascar and Uganda. African Highlands Initiative Technical Report no. 6, ICRAF, Nairobi, Kenya Carter S, Palm C, Murwira H, Gichuru M and Swift M (1998) Bioavailability of nutrients and farmer practice. 16th World Congress of Soil Science. ISSS, Montpellier, France Constantinides M and Fownes JH (1994) Tissue-to-solvent ratio and other factors affecting determination of soluble phenolics in tropical soils. Communications in Soil Science and Plant Analysis 25: 3221–3227 Fox RL and Kamprath EJ (1970) Phosphate sorption isotherms for evaluating the phosphate requirements of soil. Soil Science Society of America Proceedings 34: 902–907 FURP (1994) Fertilizer use recommendation for Kisii. Volume 1. FURP, Nairobi, Kenya Giller KE and Cadisch G (1995) Future benefits from biological nitrogen fixation: an ecological approach to agriculture. Plant and Soil 174: 255–277 Giller KE and Cadisch G (1997) Driven by nature: a sense of arrival or departure: In: Cadisch G and Giller KE (eds) Driven by Nature: Plant Litter Quality and Decomposition, pp 393–399. CAB International, Wallingford, England
36 Hilsheimer R and Harwig J (1976) Colorimetric determination of nitrate from meat and other foods: an alternative color reagent for the carcinogenic 1-napthylamine and an improved extraction method. Canadian Institute of Food Science and Technology Journal 9: 225–227 Iyamuremye F and Dick RP (1996) Organic amendments and phosphorus sorption by soils. Advances in Agronomy 56: 139–185 Jama B, Swinkles RA and Buresh RJ (1997) Agronomic and economic evaluation of organic and inorganic sources of phosphorus in Western Kenya. Agronomy Journal 89: 597–604 Ladd JN, Amato M, Jackson RB and Butler JHA (1983) Utilization by wheat crops of nitrogen from legume residue decomposing in soils in the field. Soil Biology and Biochemistry 15: 231–238 Ladha JK, Watanabe I and Saono S (1988) Nitrogen fixation by leguminous green manure and practices for its enhancement in tropical lowland rice. In: Green Manures in Rice Farming. The Role of Green Manure Crops in Rice Farming Systems. International Rice Research Institute, pp 165-183. Los Banos, Philippines Maroko JB, Buresh RJ and Smithson PC (1998) Soil phosphorus fractions in unfertilized fallowmaize systems on two soils in Kenya. Soil Science Society of America Journal (In press) Nziguheba G, Palm CA, Buresh RJ and Smithson PC (1998) Soil phosphorus fractions and adsorption as affected by organic and inorganic sources. Plant and Soil 198: 159–168 Onim JFM, Mathuva M, Otieno K and Fitzghugh HA (1990) Soil fertility changes and response of maize and beans to green manures of Leucaena leucocephala, Sesbania sesban, and pigeon pea. Agroforestry Systems 12: 197–215 Palm CA (1995) Contribution of agroforestry trees to nutrient requirements of intercropped plants. Agroforestry Systems 30: 105–124 Palm CA, Myers RJK, Nandwa SM (1997) Combined use of organic and inorganic nutrient sources for soil fertility maintenance and replenishment. In: Buresh RJ, Sanchez PA and Calhoun F (eds) Replenishing Soil Fertility in Africa: SSSA Special Publication Number 51, pp 193–217. SSSA, Madison, USA Palm CA and Rowland AP (1997) Chemical characterization of plant quality for decomposition. In: Cadisch G and Giller KE (eds) Driven by Nature: Plant Litter Quality and Decomposition, pp 379–392. CAB International, Wallingford, England Parkinson JA and Allen SE (1975) A wet oxidation procedure suitable for the determination of nitrogen and mineral nutrients in biological materials. Communications in Soil Science and Plant Analysis 6: 1–11 SAS Institute (1995) SAS/Stat Users Guide, Version 6.1, SAS Institute, Cary, NC, USA Sibbesen E (1978) An investigation of the anion exchange resin method for soil phosphate extraction. Plant and Soil 50: 305–321 Sisworo WH, Mitrosuhardjo MM, and Myers RJK (1990) The relative roles of N fixation, fertilizer, crop residues and soil in supplying N in multiple cropping systems in a humid, tropical and upland cropping system. Plant and Soil 121: 73–82 Smaling EMA (1993) An agro-ecological framework for integrated nutrient management, with special reference to Kenya. Doctoral Thesis, Agricultural University, Wageningen, The Netherlands Swift MJ, Heal OM and Anderson JM (1979) Decomposition in Terrestrial Ecosystems. Studies in Ecology. University of California Press, Berkeley, California 372 pp van der Eijk D (1997) Phosphate fixation and the response of maize to fertilizer phosphate in Kenyan soils. Doctoral Thesis, Agricultural University, Wageningen, The Netherlands van Soest, PJ (1963) Use of detergents in the analysis of fibrous feeds. II: A rapid method for the determination of fibre and lignin. Journal of the Association of Agricultural Chemists 46: 829–835 Wieder R and Lang G (1982) A critique of the analytical methods used in examining decomposition data in litterbags. Ecology 63: 1636–1642