Maize response to doubled-up legumes , compost manure and inorganic fertiliser on smallholder farms in Ntcheu district of Malawi

A study was carried out in 2011/2012 and 2012/2013 growing seasons in Kandeu and Manjawira Extension Planning Areas (EPAs) in Ntcheu district, Central Malawi, to determine maize response to crop residue incorporation from legume cropping systems and compost manure. In 2011/2012 growing season, maize with or without compost manure, sole and intercropped legumes (pigeon pea, groundnut, soyabean and cowpea) were planted. In 2012/2013 growing season, maize was planted as a test crop to assess its response to residues from legumes after harvest and N fertiliser. The experiment was laid out in a randomised complete block design, replicated 38 times. Maize grain yields following incorporation of legume crop residues were 1000 kg higher than from continuous sole cropped maize in both Kandeu and Manjawira EPAs, (p <0.001). There was no significant difference in maize grain yield following sole and doubled-up legumes. Grain yield of sole-cropped unfertilised maize, maize with inorganic fertiliser and compost manure were significantly different (p<0.001) across farms in Kandeu EPA, with an average of 3159 kg for fertilised (92 kg N ha -1 ) maize. Grain yield following sole groundnuts and top dressed with 23 kg N ha -1 was higher (3542 kg) compared to maize fertilised with 92 kg N ha -1


INTRODUCTION
Soil fertility is one of the main factors that influence agricultural productivity (Valencia et al., 2001).Soils in Sub-Sahara Africa, including Malawi, have shown to be worsening in their nutrient balance.In Malawi, soils show *Corresponding author.E-mail: joseph161980@gmail.com.
Author(s) agree that this article remain permanently open access under the terms of the Creative Commons Attribution License 4.0 International License a loss of 48kg ha -1 year -1 of nitrogen (N), 7 kg ha -1 year -1 of phosphorus (P) and 37 kg ha -1 year -1 of potassium (K) due to continuous cultivation with low inputs of inorganic fertilizer of about 14.6 kg ha -1 (Henao and Baanante, 1999).The decline in soil fertility is due to continuous growing of crops that always take away nutrients from the soil, soil erosion which carries away top soil, low incorporation of crop residues in the soil resulting into less organic matter because in most cases crop residues are burnt or fed to animals, increase in human population, climate change and poor farming practices (Mwale et al., 2011;Ngwira et al., 2012a).
Legumes or compost manure can help to enhance soil fertility and productivity of maize systems (Ngwira et al., 2012a).Legumes are integrated in maize-based systems as either intercrops (cereal-legume intercrops or legumelegume intercrop/maize ration) or rotation systems.Studies have shown that there is substantial increase in maize grain yield when grain legumes like soyabean, pigeonpea and groundnut have been intercropped (doubled-up) and rotated with maize (Twomlow et al., 2001;Bogale et al., 2001).
Legume-maize rotation research done in Ethiopia, Uganda and Tanzania showed that maize-legume cropping system yielded higher (0 to 135%) more than unfertilised (Bogale et al., 2001).In Malawi, a study conducted by Twomlow et al. (2001) in Dedza district reported that integration of legumes in maize based systems consistently increased maize grain yield from almost 0 kg ha -1 in the 1997/1998 growing season to in excess of 2,000 kg ha -1 in 1999/2000 growing season due to legume crop residues.
Substantial increase in subsequent yield of maize have been reported with integrated soil fertility management involving legumes (pigeonpea, groundnut and soybean) and little doses of inorganic fertiliser (23 kg N ha -1 and 24 kg N ha -1 ) (Njira et al., 2012).Pigeonpea can be grown as an intercrop with maize or with other legumes like groundnut and soybean.In pigeonpea/maize intercrop, there was direct benefit to maize in terms of nutrients from the decomposing litter and biologically fixed nitrogen which result into increased maize grain yield of up to 7,000 ha -1 (Mwale et al., 2011).When pigeonpea is grown as an intercrop with groundnut or soybean, one of the benefits is bringing water to the shallow rooted legume (Ghosh et al., 2005) and a combined effect on nitrogen from biologically fixed nitrogen by both legumes and the decomposing litter (Mwale et al., 2011;Ngwira et al., 2012c, Njira et al., 2012).Despite such positive results, there was still need to focus much on comparing maize grain yields following some legume-legume intercrops with emphasis to different agroecological zones.The two study areas where this study was conducted represent two biophysical environments.
Kandeu EPA represents a medium to high rainfall environments while Manjawira represent the low rainfall environments.Testing the doubled up legumes of pigeonpea, groundnut, soyabean and cowpea would help evaluate the performance of these legume cropping systems in terms of their effects on maize grain yield of the subsequent maize in relation to the two biophysical environments.Further to this, an evaluation of maize response to compost manure in relation to legume cropping systems, supplemented with small doses of inorganic fertiliser has also not been widely done and there was need to add to the little knowledge that is existing on the same.
Compost manure is one of the organic fertilisers that can be used to improve soil fertility.Organic manure, which may include both livestock and compost manure with well decomposed organic material, improves soil chemical properties and thus provide the essential nutrients for crop production (Ngwira et al., 2013).Chilimba et al. (2005) in a survey of compost manure made by farmers in all Agricultural Development Divisions (ADDs) in Malawi reported the levels nitrogen, phosphorus and potassium in the ranges of 0.21 to 2.2%, 0.05 to 0.73% and 0.12 to 2.26%, respectively.
Another four-year study by Ngwira et al. (2013) in some EPAs in Malawi showed that a combination of urea plus compost made in Manjawira EPA resulted into increase in soil organic carbon (SOC) of 64%.In the same study application of compost only resulted into more SOC (1.00%) compared to SOC in soils applied with and without inorganic fertiliser (0.87% and 0.68%) (Ngwira et al., 2013).Hence, the study aims to determine maize response to crop residue with additions from various grain legume cropping systems and compost manure.

MATERIALS AND METHODS
This study was carried out in Kandeu and Manjawira Extension Planning Areas (EPAs), Ntcheu District Agricultural Office, under the Lilongwe Agricultural Development Division (ADD), Central Malawi.Ntcheu belongs to the mid latitude and plateau areas with altitude up to 1,100m above sea level.
Manjawira EPA is at a bearing of 15.01 0 S and 34.49 0E, and elevation of 723.17m above sea level (Google Earth, 2012).Soils of Manjawira are described as Lithosols, Ferric luvisols, Xanthic ferralsols and Ferrallitic soils.The mean annual rainfall for 2009/2010 to 2011/2012 cropping season was 633 mm per annum.Kandeu EPA, located at a bearing of 14.49 0 S and 34.36 0 E, elevation of 1404m above sea level.The mean annual rainfall of 958 mm per annum for 2009/2010 to 2011/2012 cropping season (Sakala et al., 2003;Google Earth, 2012).The soils are Lithosols, Ferrallitic soils and Ferruginous with lithosols which are favourable for crops like maize and groundnuts (MoAFS Land Resource Department, 2008).
A total of 38 farmers (19 farmers in each of the two EPAs) participated in the on-farm trials, each farm served as a replicate.Each farmer had 8 treatments, of legumes and maize.The legume cropping systems were sole cowpea (Manjawira EPA only), soya bean (Kandeu EPA only), groundnut and pigeonpea; pigeonpea intercropped with cowpea, or groundnut or soybean.Maize was planted at two levels of N from urea fertiliser (0, 24 and 92 kg N ha - 1 ) and maize with compost manure at a rate of 12.5 Mt ha -1 (Ministry of Agriculture and Food Security, 2012) in the second season.This rate of compost manure gave an equivalent amount of 92 kg N ha -1  In the 2011 to 2012 growing season, eight plots making one block per mother trial farmer were used to plant the crops.Each subplot measured 10m x 10m with a total of 12 ridges spaced at 90 cm.The same plots that had legume cropping systems and maize in 2011/2012 were planted with maize as a test crop in 2012/2013 growing season.In 2011/12 season (year one), maize with and without compost manure or inorganic fertiliser, sole and intercropped groundnut, pigeonpea, soya and cowpea were planted in the eight plots.All plots that had legumes and compost manure were top dressed with 24 kg N ha -1 from urea at four weeks after planting.Maize + 0 N in 2011/2012 seasons acted as control in 2012/2013 growing season.The crop residues for both legumes and maize from 2011/2012 were incorporated in the soil, as basal organic fertiliser for 2012/2013 growing season crops.The treatments were laid out in a randomised complete blocks design (RCBD) with farmers' fields forming blocks and treatments forming units.
Soil samples were collected from each plot in in October 2011 and 2012 using an auger.From each plot, 5 points were sampled in a zig-zag manner at two depths; 0 to 20 cm (top soil) and 20 to 40 cm (sub soil).Composite samples were made for the top soil and sub-soil.The soil was air dried and ground to pass a 2mm sieve.Compost manure made by farmers was also sampled for analysis.Compost manure was sampled by collecting about three to five sub-samples from each compost heap to form a composite sample.This was done so as to have a representative sample from each farmer's heap of compost.
Both soil and compost manure were analysed for inorganic nitrogen, inorganic phosphorous, organic carbon and pH.Soil texture was analysed on soil only.The Mehlich 3 method was used to analyse available P (Chilimba et al., 1999).Total N was analysed using the Kjedahl method while organic carbon was analysed using the Walkley-Black wet Oxidation method as described by Anderson and Ingram (1989).Inorganic nitrogen as nitrate-nitrogen (NO3-N) was determined calorimetrically as described by Anderson and Ingram (1989).Soil pH was determined by using electrometric method (pH water) in the ratio of 1:2.5 soil to water (Anderson and Ingram, 1989).
The first set of maize crop data were collected after emergence to the vegetative growth stage number 6 to 10 (VE, and V6-V10).Such data included stand count after emergence and chlorophyll.Chlorophyll readings were collected using a SPAD 502Plus chlorophyll meter (Minolta, 2009).The newest fully expanded leaf with an exposed leaf collar was sampled for chlorophyll measurement near the middle of the leaf blade (Argenta et al., 2004;Shapiro et al., 2006).One leaf per plant and thirty plants per plot were sampled to provide adequate readings for computing the average chlorophyll readings and provide a reasonable estimator for plant N (Shapiro et al., 2006).
Destructive sampling of maize was done at the same four weeks after planting for analysis of tissue nitrogen.This was done in order to compare tissue nitrogen that was accumulated in the maize in 2011/2012 season before legume cropping system and nitrogen accumulated in the maize plant tissue in 2012/2013 season following legume cropping system.Eight maize plants were sampled at random from every plot, except the plot with maize plus 0 N which was split into maize plus 0 N and maize plus 24 kg N ha -1 in 2012/2013 season.In this plot, four plants per plot were sampled.To analyse for nitrogen accumulation in the maize stover, the destructive samples were oven dried at a temperature of 75°C for 72 to 96 h after which the samples were ground using a mill (Wiley mill) to pass a 1 mm sieve and stored in plastic bottles pending total nitrogen analysis in the plant tissues.Plant tissue nitrogen was determined using the selenium acid digestion procedure (Anderson and Ingram, 1989) using spectrophotometer at a wavelength of 256 nm.Data collected at harvest included; stand count, whole biomass, stovers biomass, grain yield and yield components (cob weight and 100 seed weight).
Maize yield and yield components were determined from the net plot.Sub-samples of 100 seeds were taken to be used for determining the moisture content and 100 seed weight.Maize grain yield was adjusted to standard storage moisture content of 12.5%.The harvest indices were calculated by dividing grain yield by total biomass (Kemanian et al., 2007).

Analysis of variance
The data were analysed using Genstat, 15 th edition.Variables analysed in Genstat included soil and compost parameters, plant biomass, grain yield and 100 seed weight.Specific procedures performed on the data collected included descriptive statistics and tests for normality before subjecting the data to analysis of variance (ANOVA).
In the ANOVA, maize grain yield was treated as response variable whereas the farms were the random variables and crop residues (treatments) incorporated in 2011/2012 season were regarded as fixed factor.Any differences in the means were separated using the Least Significant Difference (LSD) at p<0.05 (Williams and Hervé, 2010).
Adaptability analysis (AA) was done on maize grain yield response to legumes or compost manure in order to identify technologies that resulted into maize grain yield that was above the minimum maize grain consumption requirement for a household (Hildebrand and Russell, 1996).A comparative risk of alternative technologies was done by calculating a lower confidence limit (LCM) for the mean of each technology after which the distribution of each confidence limit was graphed.Lower confidence limit may be calculated using the equation (Hildebrand and Russell, 1996).


Where: y = treatment mean of observations within the tentative recommended domain.Sd = sample standard deviation associated with the mean.n = number of observations that went into the calculation of the mean.p = probability level (from a one-tailed t table because interest was only in values lower than the means).

Effects of residues from legume cropping systems on soil chemical properties
Legume cropping systems, inorganic fertiliser and compost manure did not significantly affect various soil chemical characteristics after 2011/2012 growing season.There was less variability in soil pH which ranged between 4.9 to 5.1 for the top soil and 4.7 to 5.1 for the sub soil across all treatments.No significant differences in organic matter accumulation across the treatments.Similarly, organic nitrogen (NO 3 -N) levels were not significant across treatments in the top soil.Despite this, inorganic nitrogen was on the higher side in the top soil Significant differences (p= 0.034) were also observed on the soil organic matter (top soil) following various cropping systems in Manjawira EPA (Tables 1 and 2).

Maize response to inorganic fertiliser, compost manure and residues from legume cropping systems
Analysis of variance showed that there was significant difference (p<0.001) in maize grain yield due to different fertiliser rates (0N, 24 kg N ha -1 ) and compost manure (12.5 Mt ha -1 ) in Kandeu EPA as shown in Table 1.There were no significant differences (p = 0.431) between continuous maize supplemented with 24 kg N ha -1 and maize plus compost, supplemented with 24 kg N ha -1 . An analysis of total crop residues incorporated in the soil, plant tissue nitrogen (%N), total nitrogen contributed by residues to the soil and number of rows per ear as covariates indicated there were no significant contribution of these covariates to the differences (p<0.05) in maize grain yield following various maize based systems.
In Manjawira EPA, there were no significant differences in maize grain yield attributed to fertiliser rates as well as nitrogen source with grain yields averages shown in Table 3.A comparison between continuous maize supplemented with 24 kg N ha -1 and maize with compost manure supplemented with 24 kg N ha -1 did not show any significant difference (p = 0.893) in Manjawira EPA as well.The grain yield of sole cropped maize fertilised with 92 kg N ha -1 was higher than maize grain yield following maize plus compost in Manjawira EPA.The difference (133 kg ) between maize grain yields following maize plus compost and maize plus 92 kg N ha -1 was not significant.The narrow difference between maize grain yields following inorganic fertiliser (92 kg N ha -1 ) and compostmanure in Manjawira EPA could be attributed to quality and appropriate use of compost manure that farmers make since they get trainings from LOMADEF, a non-governmental organisation, on how to make and use compost manure.This quality compost manure may affect certain soil properties positively, leading to good maize grain yields.This difference in average maize grain yield following maize plus compost manure and maize plus 92 kg N ha -1 suggests that it is possible to use compost manure as basal fertiliser, supplemented with 24 kg N ha -1 , and harvest maize grain yield that would almost be closer to what one would harvest if they used full rate of nitrogen in Malawi-92 kg N ha -1 .Also, the average grain yield of maize following compost manure was higher than maize grain yield following maize plus 0 kg N ha -1 supplemented with 24 kg N ha -1 , implying that if the low income households can invest their time in making the compost manure, they should be able to harvest some maize grain that are much higher than if they used no fertiliser (Table 3).

Maize grain yield following legume cropping systems in 2012/2013 growing season
In Kandeu EPA, there were no significant differences (p=0.137) in maize grain yield between continuous maize supplemented with 24 kg N ha -1 and maize plus crop residues from both sole and doubled-up legume cropping systems, supplemented with 24 kg N ha -1 .Significant difference (p = 0.0080) was observed between continuous maize supplemented with 24 kg N ha -1 and maize plus crop residues from sole cropped soyabean, supplemented with 24 kg N ha -1 , but no significant differences (0.801) were noted between continuous maize supplemented with 24 kg N ha -1 and maize plus crop residues from doubled-up legume cropping systems, supplemented with 24 kg N ha .The highest maize grain yield was that of maize following sole groundnut while the lowest maize grain yield was from maize following the intercrop of pigeonpea and groundnut.
A difference of 920 kg ha -1 between the highest and the lowest maize grain yield was noted.Although there was such a difference in maize grain yield following sole crop of groundnut and its intercrop with pigeonpea, crop residues N contributed by the same cropping systems were equally high with the intercrop of groundnut with pigeonpea resulting into more nitrogen returned into the soil through incorporation of crop residues than the sole crop of groundnut.
One probable reason why maize grain yield was lower in the intercrop of groundnut and pigeonpea could be because the woody stems of pigeonpea were not incorporated in the soil and as such, some nitrogen may have been exported out of the field through the stems.A study by Njira et al. (2012) in Kasungu reported higher maize grain yield following the intercrop of groundnut and pigeonpea.Maize grain yield was slightly higher in the maize following intercrop of soyabean and pigeonpea than maize following the sole crop of soyabean.
In this case, the nitrogen contribution through the crop residues from soyabean was equally higher in the intercrop than in the sole crop of soyabean.Nitrogen returned into the soil through pigeonpea in both sole crop and intercrop with groundnut and soyabean was very low compared to the rest of the legume cropping systems.Despite this, maize grain yield following sole pigeonpea was not the lowest of all and that the differences in mean grain yield of maize following all the legume cropping systems were small.These results therefore seem to suggest that there were minimal differences in average maize grain yields following the sole crops of groundnut, soyabean, pigeonpea and the intercrops of groundnut with pigeonpea and soyabean with pigeonpea.This is also clear when one looks at the average nitrogen input returned into the soil following all legume cropping systems which were not very different from each other although pigeonpea systems fell on the lowest side.
Furthermore, average maize grain yields following legume cropping systems were comparatively higher than maize grain yields from maize without nitrogen addition from inorganic fertiliser (0 kg N ha -1 ).A number of onfarm research studies done in Malawi, involving assessment of maize grain yield after incorporation of legume crop residues in the soil have reported maize grain yields following legumes like soyabean, groundnut, pigeonpea and the intercrop of soyabean and groundnut with pigeonpea to have been equal to or sometimes just slightly lower than maize grain yield following application of 92 kg N ha -1 , which is currently the recommended N rate application in Malawi (Njira et al., 2012;Ngwira et al., 2012b).In effect, what it means is that incorporation of legume crop residues in the soil has the capacity of making farmers harvest maize grain that are equal to or slightly lower than the yield that they would harvest if they applied full rates of nitrogen from inorganic fertiliser.
In Manjawira EPA, maize grain yield between continuous maize supplemented with 24 kg N ha -1 and maize plus crop residues from legume cropping systems, supplemented with 24 kg N ha -1 was not significantly different (p = 0.981) from each other.No significant differences (p = 0.982 and p = 0.741) were observed between maize grain yield following continuous maize supplemented with 24 kg N ha -1 and maize plus crop residues from sole and doubled-up legume cropping systems, supplemented with 24 kg N ha -1 .Also, no significant differences (p = 0.981) were observed in maize grain yield following the sole crops of cowpea, groundnut and pigeonpea, just as no significant differences in maize grain yield following the incorporation of crop residues from the intercrops of cowpea with pigeonpea and groundnut with pigeonpea were observed.In the 2011/2012 growing season, analyses showed that highest contribution of nitrogen into the soil through the incorporation of crop residues was from the intercrop of groundnut with pigeonpea, followed by the intercrop of cowpea with pigeonpea.
Maize grain yield following sole cowpea was the highest following legume cropping systems despite being the lowest contributor of nitrogen through incorporation of cowpea crop residues.The low nitrogen contribution by sole cowpea through incorporation of residues could be attributed to the attack of cowpea by aphids which may have led to reduced cowpea biomass.Maize grain yield following maize plus 92 kg N ha -1 was 2344 kg ha -1 which was slightly lower than maize grain yields following sole cowpea and sole pigeonpea.As such, if crop residues from legume cropping systems are incorporated in the soil, there is possibility of harvesting maize grain yield almost equal to or higher than maize grain yield harvested from maize applied with 92 kg N ha -1 . Maize grain yield following compost manure was 2211 kg ha -1 .This yield was slightly higher than or equal to maize grain yield following some legume cropping systems, for example, sole groundnut.With this, it is possible to suggest that maize grain yield following compost manure may be equal to maize grain yield following legume cropping systems.It also means that both legume cropping systems and compost manure with the supplementation of as low as 24 kg N ha -1 offer an opportunity of improving maize grain yields and enable farmers harvest maize grain yields almost equal to what they could harvest if they used 92 kg N ha -1 .
In both Kandeu and Manjawira EPAs, it was evident that maize grain yield following legume cropping systems, supplemented with 24 kg N ha -1 resulted into harvesting of maize grain yield that was in most cases equal to or higher than maize grain yield harvested from maize applied with 92 kg N ha -1 . It was also found out that compost manure supplemented with 24 kg N ha -1 made farmers to harvest maize grain that was just slightly lower than maize grain harvested from maize applied with 92 kg N ha -1 . Maize grain yield following both legume cropping systems and compost manure were higher than maize grain yield following maize plus 0 kg N ha -1 .Maize grain yields following incorporation of legume crop residues were significantly different (p<0.001) from continuous maize supplemented with 24 kg N ha -1 across farms in both Kandeu and Manjawira EPAs.But there were no significant differences (p<0.05) in maize grain yield following sole crops of cowpea, groundnut, soya bean, pigeonpea and maize grain yield following the intercrops of cowpea and pigeonpea, groundnut and pigeonpea and soyabean and pigeonpea in the two EPAs.Covariance analysis for factors like total amount of crop residues incorporated in the soil, plant tissues nitrogen, total crop residues nitrogen returned into soil through the incorporation of the residues, maize cob length and number of rows per cob were done was also done and these factors did not show any significant effects on maize grain yield following legume cropping systems.It was only the ANOVA of different farms, otherwise referred to as farm types, which showed significant contribution to maize grain yield following legume cropping systems.There were some farms which seemed to respond to any intervention while other farms seemed not to respond much to any interventions.As such, farms in the two EPAs were classified as responsive and non-responsive farms (Figures 1 and 2).

Risk assessment of harvesting maize grain below a recommended requirement
The main criterion for doing the risk assessment was maize grain requirement per year, 2500 kg per year in this case (UNICEF/Government of Malawi, 1996).A risk assessment was performed (Hildebrand and Russell, 1996) to see if each technology fell below the minimum requirement of 2500 kg grain per year required for each family especially in the low performing environments or farm type (EI<2500) which was the main domain.Figure 3 shows that none of the legume technologies performed above the minimum maize grain yield requirement of 2500 kg ha -1 per year to meet the minimum grain requirement for a household in Kandeu EPA.Despite this, the figure shows that in Kandeu EPA, maize grain yield following sole cropped groundnut surpassed continuous maize grain yield with full rate inorganic fertiliser at 15% risk and approached 2500 kg ha -1 at 25%.Maize grain yield following maize with compost manure, sole pigeonpea, sole soyabean and groundnut intercropped with pigeonpea performed below 2000 kg ha -1 at 20% risk while maize grain following sole cropped groundnut and the intercrop of pigeonpea with soyabean resulted into yield above 2000 kg ha -1 at 20% risk.This indicates that if adopted, maize grain yield following legume cropping systems pauses much less risk of harvesting maize grain below the household requirement in the low yielding domain in Kandeu EPA, particularly if the criterion was grain yield in kg ha -1 per year.The implication is that resource poor farmers should be able to harvest maize grain which could be enough to meet their grain requirement (Kamanga, 2011;Mwale et al., 2011;Ngwira et al., 2012b).
Similarly, Figure 4 shows that maize grain yields following both maize-based cropping systems and legume cropping systems were not capable of averting the risk of maize grain shortage by farmers in Manjawira EPA.This was the case because maize grain yield following each cropping system fell below the required minimum yield of 2500 kg ha -1 per year.Maize grain yield following most legume cropping systems was below 1800 kg ha -1 per year but some reached 1600 kg ha -1 per year at 20% risk.Maize grain yield following compost manure was outstanding up to 5% risk.At this point, maize grain yield following groundnut intercropped with pigeonpea became superior and remained the most profitable technology, yielding as high as 1700 kg ha -1 per year at   25% risk.Compost manure has the capacity of improving soil fertility and boosting maize yield (Negesa et al., 2001).Maize grain yield following intercrop of cowpea and pigeonpea, sole groundnut and sole pigeonpea also performed fairly well, reached a grain yield of 1500 kg ha -1 per year at 25% risk.Of important notice was the grain yield of maize following full rate of inorganic fertiliser (92 kg N ha -1 ) which was excessively and consistently low, starting with almost 0 kg ha -1 per year at 0.05% risk and finishing with less than 1500 kg ha -1 per year at 25% risk.On the overall, it can be concluded that maize grain yield following legume cropping systems and compost manure was better off compared to maize grain yield following maize with full rate of inorganic fertiliser and maize without inorganic fertiliser in both Kandeu and Manjawira EPAs and that the technologies (maize following legumes and compost manure) were worthy adopting because the risk of harvesting maize grain below the minimum requirement in the low yield domain (<2500 kg ha -1 ) was much lower compared to maize grain yield following recommended full rate of inorganic fertiliser (Figures 3 and 4).

Conclusions
Maize grown in rotation with groundnut compared to both legume based cropping systems and maize based cropping systems gave the highest grain yields.Incorporating crop residues from sole groundnuts helped to increase maize grain yield by 2046 kg ha -1 over continuous maize without nitrogen addition from inorganic fertiliser and 383 kg ha -1 over maize with 92 kg N ha -1 . Organic matter addition from sole crops of soya bean and pigeonpea led to higher maize grain yield compared to continuous maize without inorganic fertiliser.
Organic matter addition from sole crop of sole cowpea increased maize grain yield in Manjawira EPA compared to other legume cropping systems and maize based systems except continuous maize supplemented with 24 kg N ha -1 . It also resulted into maize grain yield which was 449 kg ha -1 higher than maize grain yield from maize without inorganic fertiliser.Compost manure has the potential to boost maize yield and it is comparable to maize grain yield in rotation with legumes.
In both EPAs, ISFM has shown to be a good option to improving maize grain yield.There can be a substantial increase in maize grain yield from maize following the incorporation of crop residues from both legume cropping systems and maize based cropping systems, supplemented with 24 kg N ha -1 . In Kandeu EPA, farmers may increase the maize grain yield by 1442 kg ha -1 while in Manjawira EPA, an increase of 678 kg ha -1 is possible if 24 kg N ha -1 is applied where maize crop residues have been incorporated in the soil compared to no supplementation of inorganic fertiliser (0 N ha -1 ). Incorporation of biomass from legume cropping systems supplemented with 24 kg N ha -1 can help farmers both in Kandeu and Manjawira EPA to harvest high maize grain yield which is comparable to maize grain yield from maize applied with 92 kg N ha -1 .

- 1 .
Maize grain yield following both sole and doubled-up legume cropping systems, supplemented with 24 kg N ha -1 , was not significantly different (p = 0.129) from each other.Maize grain yield following legume cropping systems ranged from 2622 kg ha -1 to 3542 kg ha -1

Figure 1 .
Figure1.Maize grain yield following inorganic fertiliser, compost manure, maize and legume cropping systems in Kandeu EPA Note: 1.The error bars represent the standard error of the means; 2.Mz+0N/Mz+0N = Maize without inorganic nitrogen rotation, Mz+0N/Mz+24nkg N = Maize without inorganic nitrogen rotated with maize applied with 24 kg nitrogen per ha, Mz+Compost/Mz+24 kg N = Maize applied with compost manure rotated with maize applied with 24 kg nitrogen per ha, GN/Mz+24 kg N = Groundnut rotated with maize applied with 24 kg nitrogen per ha, SB/Mz+24 kg N = Soyabean rotated with maize applied with 24 kg nitrogen per ha, PP+SB/Mz+24 kg N = Pigeonpea intercropped with soyabean rotated with maize applied with 24 kg nitrogen per ha, PP+GN/Mz+24 kg N = Pigeonpea intercropped with groundnut rotated with maize applied with 24 kg nitrogen per ha, PP/Mz+24 kg N = Pigeonpea rotated with maize applied with 24 kg nitrogen per ha.

Figure 2 .
Figure 2. Maize grain yield following inorganic fertiliser, compost manure and legume cropping systems in Manjawira EPA.Note: 1.The error bars represent the standard error of the means; 2.Mz+0N/Mz+0N = Maize without inorganic nitrogen rotation, Mz+0N/Mz+24 kg N = Maize without inorganic nitrogen rotated with maize applied with 24 kg nitrogen per ha, Mz+Compost/Mz+24 kg N = Maize applied with compost manure rotated with maize applied with 24 kg nitrogen per ha, CP/Mz+24 kg N = Cowpea rotated with maize applied with 24 kg nitrogen per ha, GN/Mz+24 kg N = Groundnut rotated with maize applied with 24 kg nitrogen per ha, PP+CP/Mz+24 kg N = Pigeonpea intercropped with cowpea rotated with maize applied with 24 kg nitrogen per ha, PP+GN/Mz+24 kg N = Pigeonpea intercropped with groundnut rotated with maize applied with 24 kg nitrogen per ha, PP/Mz+24 kg N = Pigeonpea rotated with maize applied with 24 kg nitrogen per ha.

Figure 3 .
Figure3.Risk assessment (lower confidence limit), for kg ha -1 year -1 ; on-farm maize response to legume cropping systems and compost manure in Kandeu EPA.

Figure 4 .
Figure 4. Risk assessment (lower confidence limit), for kg ha -1 year -1 ; on-farm maize response to legume cropping systems and compost manure in Manjawira EPA.

Table 1 .
Effects of legume cropping systems, inorganic fertiliser and compost manure on soil chemical properties at 0-20 cm in 2011/2012 season, Kandeu EPA.

Table 2 .
Effects of legume cropping systems, inorganic fertiliser and compost manure on soil chemical properties, Manjawira EPA.

Table 3 .
Maize grain yield at difference N fertiliser and compost manure.