African Crop Science Journal African Crop Science Society ISSN: 1021-9730 EISSN: 2072-6589 Vol. 8, Num. 3, 2000, pp. 263-272

African Crop Science Journal, Vol. 8. No. 3, pp. 263-272

TREE LEGUMES IN MEDIUM-TERM FALLOWS: NITROGEN FIXATION, NITRATE RECOVERY AND EFFECTS ON SUBSEQUENT CROPS

C.S. Wortmann* and C.K. Kaizzi¹
*Formerly with CIAT; current address: Department of Agronomy and Horticulture, University of Nebraska-Lincoln, 279 Plant Science, NE 68583, USA
¹Kawanda Agricultural Research Institute, P.O. Box 6247, Kampala, Uganda

(Received 15 December, 1999; accepted 14 August, 2000)

Code Number: CS00028

INTRODUCTION

Legumes tend to be successional species which are competitive on low N soils, but become replaced by non-legume climax species once the soil N deficiency is alleviated (Tothill, 1985). This feature of plant ecology is applicable to farming systems using the legumes in improved fallows. Nitrogen deficiencies may be alleviated by rotating legumes with crops through N₂ fixation and redistribution of N to the surface soil which is beyond the typical rooting depth of the common crops. Nitrogen accumulation for densely planted legume tree saplings ranges up to 500-600 kg N ha^-1 yr^-1 (Giller and Wilson, 1991).

The feasibility of fallows as a practice to manage soil fertility is often questioned due to increasing land shortages. However, fallow is practiced in some densely and moderately densely populated areas. In the highlands of southwestern Uganda, two independent studies found over 20% of the arable land was in fallow (excluding grazing, tree and non-cultivatable land) (Grisley and Mwesiga, 1994; Linblade et al., 1998). In parts of eastern and central Uganda, as well, approximately 20% of the arable land was found to be in fallow (Wortmann and Kaizzi, 1998).

The probable advantage of legumes over grass cover crops is their ability to fix atmospheric N through their symbiotic relationship with Rhizobium spp., usually associated with the host’s root system. Often information is inadequate to distinguish the proportions of plant N derived from the atmosphere, from deep in the soil profile, and from surface soil layers.

Capacity for biological N₂ fixation varies with species, environment and management (Tothill, 1985). It is difficult to measure N₂ fixation in the field as tree root systems are likely to go very deep into the soil (Giller and Wilson, 1991). They summarised a number of estimates of N derived from the atmosphere which ranged from 0 – 100%; the median estimate was 60% nitrogen. Cropping systems benefit from this accumulation of N if a major proportion is derived from the atmosphere or from deep horizons of the soil profile, leached beyond the typical rooting depths of other crops in the rotation.

Nitrate is adsorped on positively charged surfaces of widely occurring acidic, low organic matter sub-soils high in kaolinite and Fe and Al oxides. Legume tree (Sesbania sesban L.) fallows have been found to be efficient in recovery of nitrates from the 1 to 2 m depth and deeper (Hartemink et al., 1996; Mekonnen et al., 1997). Natural fallows can be as efficient as the tree fallows in nitrate recovery in some situations (Mekonnen et al., 1996) although Jama et al. (1998) found natural fallow to be less effective in improving soil productivity than sesbania fallow.

Deep rooting fallows may contribute inadequate P on low P soils to meet the needs of the subsequent crops; fallow plants acquire primarily crop available P since little P is available in the sub-soil of most tropical soils managed under low input conditions (Buresh and Tian, 1998). Nutrients released by decomposing legume plant material lacks a N:P balance relative to the needs of crops. On P deficient soils, P must be supplied from other sources to enable a crop to efficiently use N supplied by the fallow (Jama et al., 1998).

Uganda has phosphate rock deposits of low reactivity. The Busumbu P rock deposit lies at 0° 50-55’ N and 34° 14-18’ E and contains deeply weathered carbonatite of a calcite-magnetite-apatite-phlogopite assemblage (Davies, 1956). It consists of soft rock which is composed of fine, soft, earth material (5-7% P), and hard rock ore (13% P) where secondary apatite has cemented particles together (Van Kauwenbergh, 1991). In earlier on-farm research in Uganda, unprocessed Busumbu soft P rock was not reactive in a maize-bean rotation over four seasons, although agronomic efficiency was comparable to P fertiliser on an acid soil (pH 4.5) deficient in P and Ca in Rwanda (Nyabuhunga, pers. comm., 1999). Reactivity may be improved in the active rhizosphere of vigorously growing legumes where pH may be reduced due to N fixation (Loomis and Connor, 1992, pp. 209). Organic acids and CO₂ produced during the decomposition of green manure may induce increased acidity. Legumes, however, typically have higher ash alkalinity than grass crops (Pierre and Banwart, 1973) due to the cation/anion imbalance caused by biological N fixation; if so, decomposition of legume plant materials may reduce acidity. Negatively charged organic compounds, such as citrates and oxalates, produced by hydrolysis of organic material, may chelate Ca²⁺ ions and thus lower Ca²⁺ ions in the soil solution providing a driving force for the dissolution of P rock (Hammond et al., 1986). Reactivity of the P rock may be improved by mixing it with triple superphosphate (TSP). Upon application, acidification occurs at the surface of granules of TSP which may improve the reactivity of P rock adhering to the granules. TSP is also expected to stimulate rapid root growth of the legumes, producing more rhizosphere to interact with the P rock.

The objectives of this research were to determine: legume species efficiency in the use of P from Busumbu P rock; the amount of atmospheric N₂ fixed by the legumes; and the effects of legumes on the subsequent maize and bean intercrops.

MATERIALS AND METHODS

Field experiments were conducted in Uganda at Senge Farm near Kawanda Agricultural Research Institute over four cropping seasons in 1997 and 1998. The site has a bi-modal rainfall distribution of about 1200 mm yr^-1 with wet seasons occurring from March to June, with a peak in April, and from August to December, with a peak in November. Samples of the deep, well-drained, red sandy clay loam (Rhodic Kandhapludalf) soil taken to 20 cm depth revealed a pH of 5.0, organic matter concentration of 3.3%, and 7 mg kg^-1 available P and exchangeable K 0.39 cmol_c kg^-1.

Two factors were evaluated including legume species and P application. The legumes included sesbania (Sesbania sesban), tephrosia (Tephrosia vogelii), mucuna (Mucuna pruriens), and pigeonpea (Cajanus cajana). The P treatments included: 21 kg P ha^-1 applied as triple super phosphate (TSP) each season; 86 kg P ha^-1 applied as Busumbu P rock (8.6% P) in a single application at the beginning of the first season; 86 kg P ha^-1 applied as 875 kg of Busumbu rock phosphate and 54 kg of TSP per hectare in a single application; and no P applied. Phosphate was broadcast and incorporated. The Busumbu P rock was obtained from a quarry at the southern part of the deposit and was of the soft rock type (van Straaten, 1997).

A split plot design was used with the legumes as main plots and P treatments as sub-plots. There were four replications. Sub-plot size was 4.8 m x 5 m with 1.5 m between sub-plots, and the harvested sub-plot area was 3.5 m x 4 m.

The dates of various operations are presented in Table 1. Legumes were sown in rows spaced at 0.60 m with 3 plants m^-1 for pigeonpea, sesbania and tephrosia, and two plants m^-1 for mucuna. Maize-bean intercrop was produced in the third and fourth seasons. Maize was sown in rows spaced at 0.75 m with two plants per hill and 0.60 m between hills. Bean was sown in single rows between the rows of maize with 0.10 m between plants. No rhizobium inoculum was applied to any of the legumes.

Sesbania and tephrosia grew for two seasons. Mucuna matured at the end of the first season, but the biomass, including seed, was left in place and it reseeded and voluntarily established a full stand the second season. Pigeonpea also matured at the end of the first season, but the seed was stolen and a natural fallow established for the second season. Prior to planting for the third (1998A) season, all plants were cut near the soil surface. The aboveground standing biomass and woody (> 15 mm diameter) material were weighed in the field, and sub-samples were taken. The sub-samples were sun-dried and, later oven-dried at 60o C for two days. The dry weight of whole plant and woody biomass produced were determined. Biomass dry weight was also determined at 150 days after sowing from 1 m² plots.

Woody material of tephrosia and sesbania was removed from the plots. The remaining plant material was surface applied but some mixing of plant matter with surface soil occurred during tillage. Non-woody and woody plant materials were analysed for P content (Murphy and Riley, 1962).

Soil nitrate. Soil samples were collected from sub-plots on 14 March, 1998 and fractionated to depth increments of 0-15, 15-30, 30-50, 50-70 cm. Samples were combined for main plots and dried at approximately 40oC, crushed, and sieved. Nitrate plus nitrite was determined by CD reduction (Dorich and Nelson, 1984), with subsequent colorimetric determination of nitrite. As nitrite was probably small relative to nitrate, the values were simply reported as nitrate.

Nitrogen fixation. Ammonium sulfate was applied to micro-plots of 2.25 m² to supply 20 and 100 kg N ha to the legumes and reference crop (sorghum "cv Seredo"), respectively. The ¹⁵N atom excess in the N fertiliser was 5 and 1% for the legumes and sorghum, respectively. The fertiliser was dissolved in water and applied as a solution to ensure uniform distribution. The labeled fertiliser was applied on 9 May, 1997. The micro-plots were harvested on 16 September, 1997 by cutting the plants 10 cm above the ground. Fresh biomass weight was recorded. Sub-samples were oven dried at 60oC for dry weight determination and for subsequent N-15 analysis. The N-15 dilution method was used to calculate the extent of biological nitrogen fixation in the legumes.

Analyses of variance were conducted using Statistix V.2.0 (Analytical Software, 1998).

RESULTS AND DISCUSSION

Legume growth and nutrient acquisition. The aboveground standing biomass harvested for tephrosia and sesbania (26 and 31 Mg ha^-1) was the product of ten months of growth with more than normal rainfall during the last 5 months due to the El Nino effect (Fig. 1). In comparison, where soil P was lower and tree plant density less than in this study, Jama et al. (1998) harvested 22 and 28 Mg ha^-1 for sesbania at two locations in Kenya following 3 or 4 seasons of growth.

Tephrosia and sesbania had more aboveground biomass at the end of the second season than the mucuna and the pigeonpea-weedy fallows (Table 2). The standing biomass and plant nutrients for sesbania and tephrosia were similar in quantity, however, a greater proportion of the sesbania biomass was woody material. Jama et al. (1998) reported a much higher percentage of woody material than measured here, but they considered only the leaves and pods to be non-woody and the sesbania in their study was 18 months old at the time of cutting. Plant biomass and nutrients at the end of the second season were similar for the mucuna and pigeonpea-weedy fallows. Substantial amounts of nutrients, especially potassium, were removed in the woody biomass of sesbania and tephrosia. The amounts of nutrients returned to the soil by the shrub fallows, however, greatly exceeded that of the herbaceous fallows.

Maize and bean performance. Maize yield was greatest following sesbania and tephrosia in both seasons but the beneficial rotation effect of these fallow species was much reduced in the second successive season (Table 3). Maize performance was better following mucuna than the pigeonpea-weedy fallow rotation, but this advantage was also reduced in the second cropping season. Crop performance and response to the different fallow treatments in the 1998B season may have been constrained by soil water deficits in November and December (Fig. 1) which was the time of flowering and early grainfill. Bean yield in 1998A was similar following all legumes, but higher following sesbania and tephrosia fallows in 1998B. Bean yield in the 1998A season was presumably constrained by competition with the associated maize.

P uptake from the Busumbu P rock. P uptake by the legumes was not increased by P application, neither as TSP nor as P rock (data not presented). Mean bean yield across the two seasons was higher with TSP applied, but it was not affected by application of P rock. Maize did not respond to applied P, presumably due to adequacy of P released from the accumulated fallow biomass and to mineralisation of soil phosphorus.

Nitrogen fixation. Approximately 50% of plant N was estimated to be derived from the atmosphere for sesbania and tephrosia (Table 4). If this rate of fixation persisted over the 10 month period of growth, these two species fixed 200 to 300 kg N ha^-1. Pigeonpea had much less N₂ fixation and obtained most of its N from the soil. Species effect was not significant for the proportion of applied N recovered because of a significant species by P interaction effect. Pigeonpea recovered more N from fertiliser with TSP applied, while N recovery and N₂ fixation by tephrosia and sesbania was not affected by P application.

N₂ fixation was not measured for mucuna in this trial, but in a nearby trial, N derived from the atmosphere, averaged over two seasons, was 67% of plant N. Higher levels of fixation might have been achieved for pigeonpea with inoculation: the numbers of effective rhizobia may have been low because pigeonpea had not been produced on this land for at least 30 years. However, pigeonpea belongs to the promiscuously nodulating "cowpea" rhizobia cross inoculation group, although some evidence suggests it is slightly specific among that group (Giller and Wilson, 1991).

The problem of using sorghum, with its relatively shallow root system, as a reference crop for estimating N₂ fixation by young trees was recognised. However, the trees were only 5.5 months of age when sampled for ¹⁵N content and a major proportion of their N was probably acquired from the surface horizon. If the trees did acquire significantly more N from lower soil profiles than did sorghum, the probable implication is that the estimates of N derived from the atmosphere are low.

Soil nitrate. Surface soil nitrate levels were high at 43 days following the cutting and surface application of fallow biomass (Table 5). Soil nitrate levels were highest with tephrosia and lowest with mucuna in the surface soil. Presumably, all legumes were efficient in acquisition of nitrate in the surface soil, but more nitrate may have been released from senescent nodules and leaves of tephrosia than for other species. There was more nitrate following mucuna at the 50-70 cm depth than after tephrosia and sesbania, indicating that mucuna was less efficient than sesbania and tephrosia in acquisition of deep nitrate, possibly due to fewer deep roots with mucuna.

The high levels of nitrate in the surface soil was unexpected as the growing vegetation should have extracted most of the nitrate. However, over 40 days elapsed between the time of cutting the vegetation and the sampling for nitrates. During this period, 125 mm of rain fell with about 60 mm as small rainfall events of less than 10 mm each and one event of over 40 mm. With this periodic wetting and drying, significant amounts of nitrate may have been released through the mineralisation of surface soil organic matter, as well as the decomposition of the lower C:N components of the fallow plant material with subsequent nitrafication. The rainfall events were not sufficient to cause significant leaching, implying that most of the newly formed nitrate remained in the surface soil.

Nutrient release from legume biomass. Nutrient mineralisation and immobilisation were not measured for the legume biomass. As the plant materials were surface applied, the soil:plant material interface was much less than if the material were incorporated probably resulting in slower rates of decomposition and mineralisation of nutrients.

The legume biomass generally had sufficient N concentration and low C:N ratios (<= 20) for net N mineralisation to occur shortly after cutting (Loomis and Connor, 1992; Yadvinder-Singh et al., 1992; Jarvis et al., 1996). Only sesbania had a C:N value which may be near the critical level for significant net immobilisation to occur.

Mineralisation of P is often more important than N supply on tropical soils (Sanchez et al., 1989). Generally, P concentrations and C:P ratios were within the ranges where temporary net immobilisation of soil P at the soil:plant residue interface is expected to occur (Table 6). The critical level for net P mineralisation is estimated to be between 0.2 and 0.3% P (Yadvinder-Singh et al., 1992) and all species in the current study fell within this range. Assuming a biomass C content of 40%, the C:P ratio for most samples ranged from 200-400 while the critical C:P ratio for P mineralisation to occur may be as low as 55 to as high as 300. The moderately high C:P ratio may have reduced the effectiveness of decomposing legume plant material in improving the solubility of the P rock (Zaharah and Bah, 1991).

Plant biomass is composed of components including leaves, branches and stems which vary in nutrient contents and C:nutrient ratios (Fischler et al., 1999). The higher quality components are likely to mineralise quickly with a positive mineralisation-immobilisation turnover and an increase in nutrient availability. These rapidly released nutrients probably contributed to the nutrient supply for the subsequent maize-bean intercrop and nourished micro-organisms feeding on the lower quality plant components. The effect essentially follows a two component model for nutrient mineralisation described by Yadvinder-Singh et al. (1992) using two simultaneous first-order reactions.

Nutrient recovery. Crop yields were highest following tephrosia and sesbania fallows. The additional yield accounted for an added uptake of 18 kg N and <1 kg P per hectare over the two seasons as compared to the pigeonpea-weedy fallow. This assumes that concentrations of N for the mature maize and bean plants were 0.75 and 2.58%, respectively. However, the average N content of the tephrosia and sesbania fallows exceeded that of the pigeonpea-weedy fallow by approximately 380 kg ha^-1. Thus, recovery of N was low following the tephrosia and sesbania fallows. The fate of the remaining N was not determined, but most may have been lost. Giller and Cadisch (1995) reviewed several studies relevant to tropical agricultural systems and estimated N recovery in the first season to range from 6-28% and 9-12% of N in green manures and leguminous tree prunings, respectively. They estimated that recoveries of legume N in second and subsequent seasons were generally small (2-15%), and information on long-term recovery is lacking. Nitrogen losses, however, may well exceed the gains from N₂ fixation resulting in a negative N balance.

Much N loss may have occurred due to volatilisation as the biomass was initially surface applied, especially for tephrosia and mucuna which had relatively high N concentrations. Some leaf material was incorporated when a light tillage operation was performed to remove weeds before planting the maize-bean intercrop, but significant N losses, as much as 30%, may have occurred from the high quality leaves in the interim (Yadvinder-Singh et al., 1992; Giller and Cadisch, 1995). Termites fed on the small branches which were not removed as woody material; the fate of the N in this material is not known. Similar losses may have occurred with incorporation, however. Jones et al. (1997) reported that the residual effects of surface applied high quality tree legume leaves exceeded that of incorporated leaves, suggesting greater losses with incorporation. Hagedorn et al. (1997) found that 25% of the N in tephrosia plant material (2.85% N) was mineralised within the first five days following rainfall, and that topsoil mineral N decreased by 50-70% during a two week period of high rainfall in Rwanda. They also found that N leaching at a depth of 20 cm was increased by 50% by incorporation of green and farmyard manure versus no application, suggesting that losses by denitrification and possibly leaching were exacerbated by incorporation.

Nitrogen use efficiency might have been improved by applying the tree biomass more extensively and at lower rates. The soil nitrate levels at the time of sowing the first maize-bean intercrop (Table 5) were sufficient to support a good crop without further mineralisation, suggesting that remaining legume biomass could have been better used on nearby land where legumes had not been produced. Removal of more of the plant material as wood, however, would have resulted in a greater N deficit.

CONCLUSIONS

Sesbania and tephrosia were most effective in improving soil productivity, probably due to high biomass production, high N₂ fixation and recovery of leached nitrates. These legume fallows contributed more than mucuna and pigeonpea fallows to the soil N balance, deriving about 50% of plant N from the atmosphere. Recovery of N was low, however, and losses may have been high, possibly in excess of that gained from the atmosphere. Busumbu P rock was of low reactivity in the rhizosphere of the legumes and it was not improved by mixing Busumbu P rock with TSP. Processing, but possibly only crude beneficiation, is required to utilise the P rock more efficiently. Alternatives for the use of the tree biomass need to be investigated to improve N recovery. Medium-term improved fallow might be produced in small blocks, or in bands across the field, with transfer of a major part of the leafy biomass to adjacent land which is continuously cropped.

ACKNOWLEDGEMENTS

The authors are grateful to: National Agricultural Research Organisation and the Director of Research for Kawanda Agricultural Research Institute for providing land, facilities and other assistance for this research; Ms. G. Nalukenge for providing technical assistance; Dr. P. Smithson of ICRAF for assistance with analyses of samples; and the Canadian International Development Agency and Swiss Development Cooperation for funds provided.

REFERENCES

1. Buresh, R.J. and Tian, G. 1998. Soil improvement by trees in sub-Saharan Africa. Agroforestry Systems 38:51-76.
2. Davies, K.A., 1956. The Geology of Part of Southeast Uganda with Special Reference to the Alkaline Complexes. Geological Survey of Uganda, Memoir No. III. Government Survey Unit, Entebbe, Uganda.
3. Dorich, R.A. and Nelson, D.W. 1984. Evaluation of manual cadium reduction methods for determination of nitrate in potassium chloride extracts of soil. Soil Science Society of America Journal 48:72-75.
4. Fischler, M., Wortmann, C.S. and Feil, B. 1999. Crotalaria (C. ochroleuca G. Don) as a green manure in maize-bean cropping systems in Uganda. Field Crops Research 61:97-107.
5. Giller, K.E. and Wilson, K.J. 1991. Nitrogen Fixation in Tropical Cropping Systems. CAB International, Oxon, UK. 171 pp.
6. Giller, K.E. and Cadisch, G. 1995. Future benefits from biological nitrogen fixation: an ecological approach to agriculture. Plant and Soil 174:255-277.
7. Grisley, W. and Mwesigwa, D. 1994. Socio-economic determinants of seasonal cropland fallow decisions: smallholders in southwestern Uganda. Journal of Environmental Mana-gement 42:81-89.
8. Hagedorn, F., Steiner, K.G., Sekayange, L. and Zech, W. 1997. Effect of rainfall pattern on nitrogen mineralisation and leaching in a green manure experiment in South Rwanda. Plant and Soil 195:365-375.
9. Hammond, L.L., Chien, S.H. and Mokwunye, A.U. 1986. Agronomic value of unacidulated and partially acidulated phosphate rocks indigenous to the tropics. Advances in Agronomy 40:89-140.
10. Hartemink, A.E., Buresh, R.J., Jama, B. and Janssen, B.H. 1996. Soil nitrate and water dynamics in sesbania fallows, weed fallows and maize. Soil Science Society of America Journal 60:568-574.
11. Jama, B., Buresh, R.J. and Place, F.M. 1998. Sesbania tree fallows on phosphorus-deficient sites: maize yield and financial benefit. Agronomy Journal 90:717-726.
12. Jarvis, S.C., Stockdale, E.A., Shepherd, M.A. and Powlson, D.S. 1996. Nitrogen mineralisation in temperate agricultural soils: processes and measurements. Advances in Agronomy 57:187-235.
13. Jones, R.B., Snapp, S.S. and Phommbeya, H.S.K. 1997. Management of leguminous leaf residues to improve nutrient use efficiency in the sub-humid tropics. In: Driven by Nature: Plant Litter Quality and Decomposition. Cadisch, G. and Giller, K.E. (Eds.), pp. 239-250. CAB International, Oxon, UK.
14. Linblade, K.A., Carswell, G. and Tumuhairwe, J.K. 1998. Mitigating the relationship between population growth and land degradation: land-use change and farm management in southwestern Uganda. Ambio 27:565-571.
15. Loomis, R.S. and Connor, D.J. 1992. Crop Ecology: Productivity and Management in Agricultural Systems. Cambridge University Press, Cambridge, U.K..
16. Mekonnen, K., Buresh, R.J. and Jama, B. 1997. Root and inorganic nitrogen distributions in sesbania fallow, natural fallow and maize fields. Plant and Soil 188:319-327.
17. Murphy, J. and Riley, J.P. 1962. A modified single solution method for the determination of phosphate in natural waters. Analytical Chemistry Acta 27:31-36.
18. Pierre, W.H. and Banwart, W.L.1973. Excess base and excess base/nitrogen ratios of various crop species and plant parts. Agronomy Journal 65:91-96.
19. Sanchez, P.A., Palm, C.A., Szott, L.T., Cuevas, E. and Lal, R. 1989. Organic input management in tropical agroecosystems. In: Dynamics of Soil Organic Matter in Tropical Ecosytems : Coleman, D.C., Oades, J. M. and Uehara, G. (Eds.), pp. 125-152. University of Hawaii Press, Honolulu, Hawaii.
20. Statistix for Windows 1998. Analytical Software, Tallahassee, FL, USA.
21. Tothill, J.C. 1985. The role of legumes in farming systems of sub-Saharan Africa. In: Potentials of forage legumes in farming systems of sub-Saharan Africa. Haque, I., Jutsi, S. and Neate, P.J.H. (Eds.), pp. 162-185. ILCA, Addis Ababa, Ethiopia.
22. van Straaten, P. 1997. Geological Phosphate Resources in Central East Africa. Trip Report, 23 June to 21 July, 1997. University of Guelph.
23. van Kauwenbergh, S.J. 1991. Overview of phosphate deposits in East and Southeast Africa. Fertilizer Research 30:127-150.
24. Wortmann, C.S. and Kaizzi, C.K. 1998. Nutrient balances and expected effects of alternative practices in farming systems of Uganda. Agriculture Ecosystems and Environment 71:115-130.
25. Yadvinder-Singh, Bijay-Singh and Khind, C.S. 1992. Nutrient transformations in soils amended with green manures. Advances in Soil Science 20:237-309.
26. Zaharah, A.R. and Bah, A.R. 1991. Effect of green manures on P solubilization and uptake from phosphate rock. Agroforestry Systems 48:247-255.