Planting Density and Geometry Effect on Canopy Development, Forage Yield and Nutritive Value of Sorghum and Annual Legumes Intercropping

Abstract

Forage sorghum (FS) (Sorghum bicolor L.) is a major forage in crop–livestock production systems. It has low crude protein (CP) and is rich in fiber. Its forage quality can be improved by mixing with legumes rich in CP. Achieving a greater legume contribution to achieve higher CP over monocrops is a challenge for intercropping. Field trials were conducted with lablab (Lablab purpureus L.), pigeon pea (Cajanus cajan L.) and cowpea (Vigna unguiculata), grown with sorghum as mixed 1:1 (sorghum: legumes) and 2:2 (sorghum: legumes) rows. The sorghum plant density was 250,000 and 190,000 plants ha⁻¹. Periodic crop biomass, leaf area index (LAI) and light interception (LI) were measured, together with forage mixture quality at final harvest. The LAI and LI values were greater for lablab and cowpea intercrops in the 1:1 configuration. Legume forage yield contributed up to 10–12% of the total dry matter (DM). A significant improvement in legume biomass was observed for a sorghum population at 190,000 plants ha⁻¹ with paired rows (2:2) of either lablab or cowpea. Sorghum and total forage yields were higher in both 1:1 and 2:2 configurations with 250,000 plants ha⁻¹. The CP concentration of the forage mixture was 62–75 g kg⁻¹ DM. The acid and neutral detergent fibers and the digestibility of the forage mixture were not improved to any great extent. The results suggested that sorghum planted at normal populations with paired rows of lablab or cowpea can improve sorghum–legume intercrop productivity.

1. Introduction

Sorghum and maize are the major irrigated silage crops for dairy feed in the southern Great Plains, United States. Maize needs a higher amount of water to produce one unit of dry matter (DM) compared to sorghum [1]. However, both are poor in protein content, are rich in fiber and have low digestibility. Low crude protein (CP) concentration in dairy feed with cereals is a major limitation [2]. Recently, efforts were made to identify forage crops and cropping systems with higher water-use efficiency (WUE) and better nutritional quality. Sorghum silage quality can be improved by the addition of protein-rich legumes in significant proportions [3]. The morphological features of legumes are different from those of cereals, particularly in the root system [4,5,6]. Legumes are known to improve the soil N by biological fixation, which improves soil fertility [7] and the short stature and early fast-growing nature of the crop [8]. The advantages of legumes under intercropping include the improvement of low-quality forages through the complementary effects of two or more crops grown simultaneously on the same area of land [8]. Intercropping of maize and sorghum with legumes has been explored as a means to increase CP concentration in silage [9,10,11,12,13]. Sorghum is very water-efficient and has the ability to compensate considerable yield with respect to growing conditions and planting rates. The differential spatial and temporal use of resources, viz., solar radiation, water and nutrients was relatively improved when grown in combination [4,14]. In an intercrop system, row configurations alter the amount of light transmission to the lower crop canopy and affect the competition of species for light, water and nutrients.

In forage cereal–legume mixtures, it is important that the population of the cereals should be as close as possible to that of pure stands. The proportion of legumes should not be so high as to substantially decrease the yield of the main crop. In general, a lower yield of individual crops is observed in intercropping compared to monocropping. The decreased biomass has been attributed to competition for light, moisture and nutrients. More efficient use of light can be attained by careful spatial arrangements of multi-story cropping with tall and short crops, provided the short crops are adapted to low light intensities. The choice of crop combination is key for successful intercropping. Incompatibility factors such as planting densities, and root system and nutrient competition must be considered [15].

Yu et al. [16] repoted 50:50 mixture data from a replacement series experiment in a sorghum–bean intercrop, where the cereal component contributed a greater proportion of the mixture yield. In sorghum and pigeon pea, when sorghum densities were varied from 55,000 to 22,000 plants ha⁻¹ and combined with a constant pigeon-pea density (37,000 plants ha⁻¹), the intercrop sorghum yield response was linear. When the companions are present in approximately equal densities, productivity and efficiency appear to be determined by the more aggressive crop, usually the cereal [17]. In contrast, the intercrop pigeon pea yield decreased with rising sorghum density. Contreras-Govea et al. [18] reported that the lablab forage yield (2.82 Mg ha⁻¹) was much greater than the cowpea yield (1.53 Mg ha⁻¹) but the CP concentration was higher in cowpea (286 g kg⁻¹ DM) than in lablab.

The row arrangement, in contrast to arrangements of component crops within rows, improves the amount of light transmitted to the lower legume canopy. Such arrangements can enhance legume yields and the efficiency in cereal–legume intercropping. Greater LER and yields of groundnut were obtained for 0.9 m spaced sorghum compared to 0.6 m spaced plants. Furthermore, the LI was enhanced by 20 percent in the widely spaced crops (0.9 m). Light is the most important factor in interspecies competition. Leaf area index and LI are directly related to the canopy structure and population density. The spatial and temporal distribution of radiation transmission has been reported by several researchers in row crop canopies under pure stands [16,19]. Planting geometry and row orientation has no significant effect on LI in monocropping.

The proportion of component crops and their arrangement also decide the competition and yield levels. Brown et al. [20] found 5% higher net radiation (Rn) in wide-row sorghum than in narrow-row sorghum and found that a higher plant population decreased the LI at the soil surface and increased the energy absorbed by the plants. Pal et al. [21] reported that intercropping of sorghum (100%) + cowpea (50% seed rate) may be grown for higher green fodder yield protein yield and net returns. Gonzalez and Graterol [22] realized that the intercropped sorghum yield was significantly affected by row spacing. Mohta and De [23] reported that a 31% higher soybean yield was obtained with maize and 26% with sorghum when they were arranged in a 2:2 configuration compared to 1:1 intercropping. Ibrahim et al. [24] observed improved sorghum growth with lablab intercropping at 1:1.

Quality of the ensiled material is a primary concern in any forage production system. High-quality forage must have high intake, digestibility and efficiency in utilization [17]. Compositionally, legumes are known for having a higher protein content and lower fiber content. Neutral detergent fiber (NDF) is a measure of the total cell wall fraction, and acid detergent fiber (ADF) accounts for the indigestible fraction of the forage mixture. In the sorghum–soybean system, CP and in vitro true dry matter digestibility (IVTDMD) were increased over the respective monocrop [25]. Increased CP concentration in maize + lablab (13%) and maize +velvet bean (16%) further increased under a low maize density of 550,000 plants ha⁻¹ compared to 825,000 plants ha⁻¹ [11].

Musa et al. [26] showed that barley + peas increased the DM production and yield compared with monocrops, irrespective of the planting arrangement. Improved pea growth arose from the support offered by the barley, which was greatest in the mixed and cross rows and least in the pairs of alternate rows. Belel et al. [27] reviewed the beneficial effects of crop mixtures and row proportions in cereal–legumes intercropping. The objectives of the study were to identify potential legume species in a modified row configuration and the sorghum population density in sorghum–legumes intercropping. Legumes were evaluated based on the canopy development, biomass production, legume contribution and forage quality.

2. Materials and Methods

2.1. Experimental Site

A field study was conducted in 2009 and 2010 at the Agricultural Science Center, Clovis (34°26′ N, 103°12’ W, 1332 m elevation), of New Mexico State University, NM, USA. The experimental site has Olton silty clay loam with pH 7.7 (1:1). The initially tested available soil nitrate was 21.6 kg ha⁻¹, while P and K were 13 and 504 ppm, respectively. Weather parameters were collected from a national cooperative weather service station located at the experimental site. The 25-year average annual rainfall was 450 mm, and the temperature was 22 °C. The previous crop was wheat, and after harvesting the site was kept fallow in both years. The growing season was defined as May to early October. The environmental variables collected during the growing season included maximum and minimum air temperature, rainfall and cumulative growing degree days at 10 °C base temperature. The data are presented in Figure 1. Before planting, fertilizers were incorporated at 221-51-0-19-3.5 and 210-40-0-16-6.8 N-P₂O₅-K₂O-S-Zn kg ha⁻¹ in 2009 and 2010, respectively. The herbicide Dual II Magnum was incorporated as preplant at 1.19 l ha⁻¹.

2.2. Experimental Design and Crop Management

The experimental design was a factorial split plot arrangement replicated four times. Sorghum populations (250,000 and 190,000 plants ha⁻¹) were in the main plots and cowpea, lablab and pigeon pea intercrops in sub plots. These were arranged in a split plot manner, and the treatment combinations and row configurations were arranged in a factorial design. Three row configurations were mixed, 1:1 and 2:2. For mixed intercropping, sorghum and legume seeds were mixed before planting at the normal recommended rate (150,000 plants ha⁻¹) and both were planted in a single row. For the 1:1 intercropping treatment, a single row of legume was planted between sorghum rows at 0.75 m apart. In the 2:2 row combination, sorghum was planted in paired rows at 37.5-112.5-37.5 cm, and two rows of legumes were planted between the sorghum rows (Figure 2). Legumes were selected based on shade tolerance, maturity period and morphological characteristics. Further details of the legumes are presented in

Table 1. Planting details of forage sorghum and legume crops used for intercropping in both the years of study (2009 and 2010) at Clovis, NM.

Crop	Species	Variety	1000-Seed Weight (g)	Population Density (Plants ha−1)	Seeding Rate by Weight (kg ha−1)
Sorghum	Sorghum bicolor (L.) Moench	FS-5	32–33	250,000	8.0–8.25
Sorghum	Sorghum bicolor (L.) Moench	FS-5	32–33	190,000	6.08–6.27
Cowpea	Vigna unguiculate (L.) Walp	Iron clay	100–110	150,000	19.0–20.9
Lablab	Lablab purpureus (L.) Sweet	Rio Verde	140–160	150,000	26.6–30.4
Pigeon pea	Cajanus cajan (L.) Milsp.	GA-1	120–150	150,000	22.8–28.5

. The sorghum test variety was FS-5, planted 0.75 m apart. Each plot had six rows of sorghum and legumes. Individual plot size was 40.5 m² (4.5 m × 9 m). Planting dates were 7 June 2009 and 24 May 2010. Planting was achieved using a John Deere MaxEmerge planter with separate cones for each row. Orthene insecticide at 2.5 L ha⁻¹ was applied with a backpack sprayer at 10 DAS to control thrips. Interprid insecticide at 2.5 L ha⁻¹ was applied with irrigation water at 60 DAS. Cereals recovered better than legumes, leading to poor performance by legumes.

2.3. Experimental Measurements

The photosynthetically active radiation (PAR) was measured above and beneath the plant canopies with a SunScan canopy analysis system (Delta-T Devices, Cambridge, UK). The system has a single quantum sensor and a linear quantum sensor (SunScan probe 1 m long with 64 photodiodes equally spaced along its length) for measuring the PAR above and beneath the plant canopies, respectively. The single quantum sensor was always kept at a height of 1.8 m to record the incident light at the top of crop canopies. The linear quantum sensor was set perpendicularly to the crop row at the soil surface. All measurements were recorded on clear sunny days between 10:00 and 14:00 MST. Two sets of observations were recorded per plot. The fraction of the light transmitted (I_f) was calculated as I_f = Ig/Io, where Ig is the light at the soil surface and Io is the light at the top of the crop canopy (incident light). The fraction of light intercepted is (1 − I_f) and is expressed as a percentage.

Sorghum and legumes were harvested at 2.5 cm above the ground surface for biomass production throughout the cropping season. Final biomass harvest was done at soft dough stage of sorghum on 7 October 2009 and 15 September 2010. At each sampling, both sorghum and legumes were harvested separately from a 0.75 m² except at final harvest. Sorghum samples from each treatment were chopped, subsampled (≅500 g) and dried at 60 °C in a forced air circulation oven until a constant weight was attained (48 to 72 h), to determine the DM. The entire quantity of legumes was dried due to its small quantity. The legume contribution was calculated considering the legume DM yield and the total yield.

The soil moisture in the different intercropping systems was measured using a CPN 503 DR Hydroprobe neutron moisture probe (Campbell Pacific Nuclear International Inc., Martinez, CA, USA). Neutron probe tubes were installed in each treatment to a depth of 1.5 m. Soil moisture was recorded at 0.2 m depth intervals. Before the actual field observations, the neutron probe was calibrated, and a regression equation was developed for the experimental site. Neutron tubes were installed in intercrop rows between two sorghum rows. After installation, care was taken to prevent water stagnation around the tube during the whole season. Neutron probe readings (count ratios) were converted into % volumetric soil moisture by regression. Soil moisture was estimated at regular intervals from 54 DAS up to harvesting.

2.4. Quality Analysis

Prior to chemical analysis, dried leaves and stem samples of sorghum and legumes were ground separately using a Wiley mill (Thomas Manufacturing Philadelphia, PA, USA) to pass through a one mm screen. Ground legume samples were mixed with sorghum samples using the percentage legume contribution to the total biomass on a DW basis. The mixed samples were submitted to a National Forage Testing Association (NFTA) certified laboratory (Ward Laboratories, Kearney, NE, USA) for forage nutritive value analysis using near-infrared reflectance spectroscopy (NIRS). The major quality components included CP, ADF, NDF, NDF digestibility (NDFD), total digestible nutrients (TDN) and IVTDMD. In addition, the mineral elements Ca, K and Mg were also estimated. Relative forage quality (RFQ) is an estimate of how much available energy a non-lactating animal will obtain daily from a forage/silage when it is fed only that particular forage.

2.5. Statistical Analysis

Analysis of variance (ANOVA) for the factorial split plot design was performed to determine the effect of population density, row configuration and legume intercrop species on biomass yield, LI and quality. Interaction effects of the main and subsidiary plots were analyzed for each parameter. PROC GLM was used to analyze combined multi-year data. If the ANOVA for the multi-year combined data showed a significant interaction effect between treatments and years, a separate ANOVA was conducted for each individual year. SAS 9.3 (SAS Inst., Cary, NC, USA 2008) was used for all analyses.

3. Results and Discussion

3.1. Environmental Conditions

Seasonal precipitation, along with minimum- and maximum-temperature daily GDDs are presented in Figure 1. The distribution of rainfall during the experimental period varied between the two growing seasons. The total precipitation received from planting up to the final harvest was 328.7 mm and 313.2 mm in 2009 and 2010, respectively. Trials were irrigated under a center pivot in both years to prevent water stress. The total amount of irrigation water provided in 2009 and 2010 was 127.0 and 280.7 mm, respectively. Research site received less than normal precipitation in June (32 and 75 mm in 2009 and 2010, respectively) and September (22 and 35 mm in 2009 and 2010, respectively), and therefore the research site was irrigated to supplement the rainfall to near-normal levels. These conditions were suitable for better early development of both sorghum and legumes. Both the growing seasons were characterized by only small differences in seasonal maximum (32.4 and 34.6 °C, respectively) and minimum temperatures (16.8 and 17.1 °C, respectively). The accumulated GDDs during the cropping season were 2060 and 2242 °C days in 2009 and 2010, respectively. The region generally has higher temperatures between May and August. The seasonal mean incident radiation was 24.5 and 22.5 MJ m⁻² d⁻¹ for the 2009 and 2010 experimental seasons. Early crop growth in both the years was uniform.

3.2. Leaf Area Index and Light Interception (%)

The LAI of legume-based sorghum intercropping in 2009 and 2010 was either similar or slightly greater among the treatment combinations. It peaked at between 80 to 90 DAS (Figure 3). At the middle stage of the crop period (62 DAS), greater LAI values were recorded in the 1:1 (6.87) and mixed (6.86) row combinations than in the 2:2 (5.65) row combination. At the final observation, no significant difference was observed. The LAI of sorghum + legumes at 2:2 (0.79 to 7.8) was significantly lower than at 1:1 (0.92 to 8.57) and mixed proportions (0.7 to 8.68). There was no difference in LAI between sorghum plant densities, except during the grand growth phase (45–75 DAS).

Sorghum + pigeonpea had a lower LAI throughout the growing season. At 62 DAS, the LAI of sorghum with cowpea (6.7) and lablab (6.6) was greater than with pigeon pea (6.2). Significant differences were observed up to 45 days, with no differences thereafter. There was a strong relationship (r² = 0.99) between the growing period and the LI (Figure 3 and

Table 2. Parameters of regression analysis between leaf area index (LAI) and light interception (LI) in various forage sorghum–legume intercrop species combinations tested under various row proportions and sorghum population densities in Clovis, in 2009 and 2010. The nonlinear regression was fitted in a linear manner with equation y = y₀ + a ∗ ln (x − x₀) where y = light interception, x = leaf area index, x₀ = intercept on x-axis, a = slope and y₀ = intercept on y-axis.

Intercropping System	x0	y0	a	SE	Adj R2	p Value at 0.05	Predicted LAI at 90% LI
Row proportions
Sorghum + legumes (1:1)	0.865	80.39	9.96	1.468	0.99	<0.0001	3.4
Sorghum + legumes (2:2)	0.731	78.84	11.03	1.069	0.99	<0.0001	3.51
Sorghum + legumes mixed	0.669	78.29	10.81	1.505	0.99	<0.0001	3.63
Sorghum population density (plants ha−1)
250,000	0.806	79.77	10.26	1.185	0.99	<0.0001	3.52
190,000	0.689	77.49	11.46	1.358	0.99	<0.0001	3.68
Intercrop species
Cowpea	0.808	79.84	10.23	1.089	0.99	<0.0001	3.51
Lablab	0.744	79.11	10.44	1.303	0.99	<0.0001	3.59
Pigeon pea	0.707	77.82	11.42	1.196	0.99	<0.0001	3.62

). In all the treatments, rapid canopy development was observed up to 60 DAS. Biomass accumulation during the vegetative phase may be responsible for the sharp increase in LAI. At 30 DAS, the LAI ranged from 0.71 to 0.92, reaching 8.18 to 8.66 at final harvesting. Generally, an LAI in the range of 3 to 5 was considered as optimum for intercepting all solar radiation. The LAI of all treatments consistently increased as growth advanced. The results in this study are also consistent with the results reported by Arshad and Ranamukhaarachchi [28] and by Darapuneni et al. [19], who reported that monocropping results in a lower LAI than legume-based intercropping.

Light interception by sorghum + legumes was significantly influenced by the row arrangement, plant density and intercrop species (Figure 3). The two rows of legumes planted between two rows sorghum (2:2) showed less light interception than the 1:1 and mixed arrangements. At 28 DAS, the 1:1 arrangement had intercepted 51.5% of the PAR, reaching a maximum (99%) much earlier than the 2:2 and mixed arrangements. At 46 DAS, the 1:1, 2:2 and mixed arrangements had intercepted 93.2, 85.9 and 90.4%, respectively. This means that longer period was necessary to reach maximal LI levels in the 2:2 and mixed arrangements. However, all crop arrangements reached 99% LI at 62 DAS. Rapid expansion of the LAI may be responsible for LI saturation at an early crop stage. Models accurately predicted the LAI based on the LI values of spatially arranged sorghum and legumes (Table 2). A nonlinear logarithmic relation was observed between the LAI and LI in each row combination. This indicated that the LAI was a perfect predictor of the LI. Although the LAI range required to intercept 90% of light is narrow, the 1:1 combination reached the maximum LI (3.4) much earlier than the 2:2 and mixed intercropping combinations. A sorghum plant density of 250,000 ha⁻¹ resulted in a higher LI at an early stage than a density of 190,000 ha⁻¹. After 42 DAS, the LI did not differ significantly between plant densities. The predicted 90% LI value based on LAI showed that it was reached much earlier for a density of 250,000 ha⁻¹ (3.52) than for 190,000 (3.68).

At 35 DAS, sorghum with cowpea (72.2%) and lablab (72.5%) attained higher LI values than with pigeon pea (65.1%). Rapid increase in LAI and LI occurred at early stages, and the optimum LAI was achieved much earlier in cowpea (3.51) and lablab (3.59) than in pigeon pea (3.62). The time period for canopy coverage and LI was prolonged for the sorghum–pigeon pea combination. After 60 DAS, the LI for all treatment combinations reached a plateau (>98%). The sorghum–legume interaction in the 1:1 arrangement with 250,000 plants ha⁻¹ and either cowpea or lablab as intercrops was found to be effective with regard to LI. It was indicated that legume in association with sorghum increased the LI much earlier in the season. Fuchs and Stanhill [29] opined that row structure and foliage geometry are determinants for greater light interception in sorghum canopy. Malik et al. [30] also reported that significant influence of seed rate and row spacing on sorghum fodder, it was highest when planted at 75 kg ha⁻¹ and narrow spacing. Furthermore, Gao et al. [31] and Tsubo et al. [32] reported that intercropped beans had greater specific leaf area than bean monocrops. The difference found was remarkable in beans. It may be responsible for the higher LI in legume intercropping. In spite of the improvement in canopy development under intercropping, the use of solar radiation occurred much earlier in the season. The results of our study are also in agreement with the results of Wang et al. [33] and Darapuneni et al. [19], who opined that synchrony of the maturity period of component crops plays a major role in the success of intercropping systems.

3.3. Soil Moisture

Soil water content under intercropping was greatly influenced by both sorghum plant populations and row proportions (Figure 4). However, in sorghum–legumes row proportions, the soil water variation was observed to have a deeper profile. For all observations, the least soil moisture beyond a 0.5 m depth was recorded in sorghum–legumes mixed intercropping. Across the different treatment combinations, during the period from 56 to 117 DAS, the average soil water content decreased from 24 to 22%. A marginal difference (<2.0%) was observed from 54 DAS up to the end of the season. A lower soil water content was observed for mixed intercropping (23.9 to 21.9%) than for 1:1 (24.4 to 22.2%) and 2:2 (24.4 to 22.6%) row proportions. It was estimated that it was 1.3 to 2.3% lower than for 1:1 and 2:2 row proportions. No great difference in soil water content was observed between densities of 250,000 and 190,000 plants ha⁻¹.

3.4. Forage Yield

Forage DM yields were primarily influenced by the effects of crop geometry, density and intercrop species combinations (

Table 3. Sorghum, legumes and total forage dry matter (Mg ha⁻¹) of various forage sorghum–legume intercrop combinations with different row proportions and sorghum population densities at Clovis, NM, in 2009.

Treatment	52 DAP			95 DAP			128 DAP
	Sorghum	Legumes	Total	Sorghum	Legumes	Total	Sorghum	Legumes	Total
Row proportion
Sorghum + legume (1:1)	10.451 a	0.406 b	10.82 ab	13.41	0.505 b	13.95 ab	15.76 a	0.57 b	16.33 a
Sorghum + legume (2:2)	10.867 a	0.758 a	11.60 a	14.21	1.253 a	15.52 a	14.63 a	1.32 a	15.95 ab
Sorghum + legume (Mixed)	9.267 b	0.314 b	9.60 b	13.01	0.415 b	13.49 a	15.11 b	0.435 c	15.55 b
LSD (p = 0.05)	1.44	0.209	1.34	NS	0.224	2.03	NS	0.18	0.57
Sorghum population density (Plants ha−1)
250,000	10.19	0.48	10.67	13.71	0.76	14.44	15.68	0.679	16.36
190,000	9.091	0.51	10.68	13.38	0.69	14.20	13.05	0.83	13.88
LSD (p = 0.05)	NS	NS	NS	NS	NS	NS	NS	NS	NS
Intercrop legumes
Cowpea	10.42	0.649 a	11.07	13.90	0.760	14.66	15.23 c	0.894 ab	16.12
Lablab	10.02	0.578 a	10.61	13.29	0.687	13.98	15.04 a	0.930 a	15.97
Pigeonpea	10.14	0.249 b	10.39	-	-	-	15.23 b	0.501 b	15.73
LSD (p = 0.05)	NS	0.089	NS	NS	NS	NS	0.63	0.13	NS

Same alphabet against the mean values in each column are non-significant at p = 0.05. NS—non-significant.

and

Table 4. Sorghum, legumes and total forage dry matter (Mg ha⁻¹) of various forage sorghum–legume intercrop combinations with different row proportions and sorghum population densities at Clovis, NM, in 2010.

	48 DAS			65 DAS			114 DAS
	Sorghum	Legumes	Total	Sorghum	Legumes	Total	Sorghum	Legumes	Total
Row Proportions
Sorghum + legumes (1:1)	4.58 a	0.218 b	4.80 a	9.62 a	0.321 b	9.95	19.70	0.308 b	20.01
Sorghum + legumes (2:2)	4.09 b	0.338 a	4.43 c	8.82 b	0.537 a	9.36	18.85	0.644 a	19.50
Sorghum + legumes (Mixed)	4.50 a	0.108 c	4.61 b	9.70 a	0.181 c	9.88	20.21	0.133 c	20.34
LSD (p < 0.05)	0.17	0.038	0.14	0.45	0.03	0.42	NS	0.05	0.65
Sorghum population density (Plants ha−1)
250,000	4.53	0.209 b	4.74	9.51	0.330 b	9.84	19.54	0.367	19.91
190,000	4.25	0.234 a	4.48	9.26	0.362 a	9.62	19.63	0.356	19.99
LSD (p < 0.05)	NS	0.022	NS	NS	0.015	NS	NS	NS	NS
Intercrops
Cowpea	4.48	0.231 b	4.71	9.26 b	0.334 b	9.59 b	18.90 b	0.317 b	19.22 b
Lablab	4.31	0.345 a	4.66	8.96 b	0.527 a	9.49 b	19.29 b	0.602 a	19.89 b
Pigeonpea	4.38	0.089 c	4.47	9.93 a	0.178 c	10.11 a	20.57 a	0.167 c	20.74 a
LSD (p < 0.05)	NS	0.022	NS	0.36	0.034	0.37	0.78	0.035	0.78

Same alphabet against the mean values in each column are non-significant at p = 0.05. NS—non-significant.

). Sorghum DM yield was greater in 2010 compared to 2009. At final sampling, neither crop arrangement nor population density influenced sorghum DM. It was much lower in the 2:2 combination than in the 1:1 and mixed arrangements. Among intercrops, pigeon pea affected the sorghum DM (15.23 and 20.57 in 2009 and 2010, respectively) much more than the other crops. Pigeon pea DM contribution was much lower than the cowpea contribution, and lablab did not affect the sorghum DM. An increased legume contribution in the mixture reduced the sorghum DM, which was greater in the 1:1 and mixed row proportions. As legume density increased, sorghum DM decreased. Results from our study also indicated that differences in sorghum and legumes DM were due to plant density. Armstrong and Albrecht [11] also reported that beans DM was reduced by increased maize density. Furthermore, increasing the bean density from 0 to 120,000 plants ha⁻¹ reduced the maize DM by 2.5 Mg ha⁻¹. This implies that the plant density of both species is crucial in achieving a greater DM mixture yield. Results of this study also indicated that a difference in sorghum and legume DM yield was due to plant density. Reduced sorghum plant density (25%) had no significant effect on both sorghum and total DM yield. At final harvest, the total forage yield of the mixture ranged from 15 to 21 Mg ha⁻¹. However, there was a greater contribution from the sorghum DM yield (>88%) than the legumes DM yield (<12%). Dharapuneni et al. [19] and Lepcha [34] also reported forage yield differences in sorghum–legume intercropping, primarily originating from the sorghum DM rather than legumes. Maasdorp and Titterton [35] found that in a maize–velvet bean mixture that contained 30% bean, the maize production was 4.1 Mg ha⁻¹ compared to 8.0 Mg ha⁻¹ in a maize pure stand.

In 2009, the total DM yield did not differ significantly among row proportions (Table 3). However, in 2010, a greater forage yield was obtained from sorghum–legumes in 1:1 (20.34 Mg ha⁻¹) and mixed (20.0 Mg ha⁻¹) combinations. Lowest total forage DM yield was recorded in sorghum–legume 2:2 row proportions (19.49 Mg ha⁻¹). It was mainly due to a lower sorghum forage yield as compared to former combinations. Yu et al. [16] opined that increasing the density of the cereal species in a cereal + legume intercrop would not only increase the productivity of the cereals but also increased the productivity of the intercrop as a whole. The positive effect of increasing the relative density of the cereal on its relative yield was stronger than the negative effect on the companion legume. The results reported by Lepcha [34] revealed that lablab and sorghum proportions did not result in increased DM, due to the lower legumes DM contribution (<12%).

The effect of sorghum–legumes intercropping on legume DM was assessed in detail by harvesting biomass samples throughout the 2009 and 2010 growing seasons (Figure 5). Row proportions × plant density × intercrops interactions significantly influenced the legume DM contribution to the total forage yield. Sorghum+ legumes (2:2) had a significant influence on legumes DM: at final harvest, it was 24% and 109% greater than for 1:1 and 384% and 151% greater than for mixed intercropping in 2009 and 2010, respectively. The lowest legume DM was recorded in mixed intercropping. A sorghum plant density of at 190,000 ha⁻¹ did not have significant effect on legumes DM (9–12%) compared with a density of 250,000 ha⁻¹. However, in 2010, at early growth stages, a lower sorghum plant density had a significant influence on legumes DM. The proportion of lablab DM to the total was relatively greater (up to 7.4%) than that of cowpea (<5%) and pigeon pea (<4.0%). A density of 190,000 plants ha⁻¹ in a 2:2 arrangement resulted in a greater legumes contribution (10–12%). A strong relationship was observed between the proportion of legumes to the total DM and plant density (r² = 0.77–0.99), intercrop species (r² = 0.49–0.92) and planting geometry (r² = 0.51–0.88) (Figure 5). Another study conducted at the same site by Angadi et al. [10] reported that the legumes DM potential varied with the type of FS and the intercrop species. The yield of FS was reduced when mixed with lablab and lima bean but not when mixed with cowpea and pole bean. In the same study, although the lima bean DM contribution was relatively greater, the total DM was much lower than for FS–cowpea and FS–pole bean. Riday and Albrecht [36] stated that growing maize with annual legumes did not improve the production but did rearrange the DM constituents. Contreras-Govea et al. [37] also reported that at the same study site, the legume DM proportion was lower in forage–sorghum mixtures.

3.5. Forage Nutritive Value

A change in the forage quality of the sorghum–legume mixture was observed in both 2009 and 2010 (

Table 5. Forage quality NIR analysis of sorghum and legume mixture at the end of the growing season. Legumes were mixed with sorghum in proportion to the respective contribution. Samples were selected from a population of 250,000 plants ha⁻¹ with cowpea and lablab intercrops.

	CP	ADF	NDF	NDFD	IVTDMD	RFQ
	2009
Row proportions g kg−1 DM
Sorghum + legumes (1:1)	72.0	323.5	492.3	80.3	902.5	146.3 a
Sorghum + legumes (2:2)	74.8	331.0	500.6	76.3	880.1	137.9 b
Sorghum + legumes (Mixed)	71.0	327.6	507.4	77.6	885.9	139.9 b
LSD (p = 0.05)	NS	NS	12.9	3.5	19.8	4.9
Intercrop species
Cowpea	72.5	329.2	505.0	77.3	884.5	139.3
Lablab	72.7	325.6	495.2	78.8	894.5	143.3
LSD (p = 0.05)	NS	NS	NS	NS	NS	3.5
	2010
Row proportions
Sorghum + legumes (1:1)	61.8 b	368.4 b	563.5	525.0 a	732.1 a	102.4
Sorghum + legumes (2:2)	70.4 a	379.8 a	571.3	503.8 b	715.9 b	101.1
Sorghum + legumes (Mixed)	63.4 b	374.0 ab	567.0	530.0 a	733.4 a	103.4
LSD (p = 0.05)	2.7	6.6	NS	10.8	7.1	NS
Intercrop species
Cowpea	62.8 b	375.3	568.8	524.2	729.3	101.3
Lablab	67.5 a	372.8	565.7	515.0	725.0	103.3
LSD (p = 0.05)	3.6	NS	NS	NS	NS	NS

CP—crude protein; ADF—acid detergent fiber, NDF—neutral detergent fiber; IVTDMD—in vitro true dry matter digestibility; NDFD—% NDF digestibility; RFQ—relative feed quality. NS—non-significant. Same alphabet against the mean values in each column are non-significant at p = 0.05.

). These changes also depend on the legumes DM contribution, total forage yield and species composition. The CP concentration ranged from 62 to 75 g kg⁻¹ DM. Previous researchers also reported that the CP concentration depends on legume and cereal plant densities [11,37] At high maize density and lower lablab proportions, the CP was much lower and did not contribute to CP improvement. In our study, due to the lower legumes DM (<12%), the CP concentration was not greatly increased. A difference in CP, ADF, NDF and digestibility was noted for different row proportions. However, the intercrop species did not have any influence on forage quality traits in either of the years. Contreras-Govea et al. [37] compared maize and sorghum silage samples with different concentrations of lablab and reported that the CP concentration of the silage was enhanced by an increased concentration of beans in the forage mixture. In another laboratory study by Contreras-Govea et al. [38] at the same center, they reported that when cowpea was mixed in different proportions with maize and sorghum, the CP concentration was improved along with the cowpea proportion. Therefore, it was clear that a combination of sorghum or maize with lablab or cowpea increased the silage CP concentration, but the impact or contribution of legumes in the mixture depends on the sorghum or maize planting density. Angadi et al. [10] also observed that the CP concentration of sorghum + lima bean (7.9%) was greater than for sorghum with pole bean, cowpea or lablab.

The ADF and NDF concentrations were not affected by either the row proportions or the intercrops (Table 5). The ADF concentration ranged from 323 to 380 g kg⁻¹ DM, and the NDF ranged from 492 to 571 g kg⁻¹ DM. Both ADF and NDF were dependent on the legume contribution. In our study, the small differences across treatment combinations were also due to the lower legume contribution. The NDF is composed of the hemicellulose, cellulose and lignin parts of the forage, and the ADF accounts for cellulose and lignin. These were negatively correlated with IVTDMD. High-quality forage has a low concentration of both NDF and ADF and a high digestibility. Grasses are higher in hemicellulose and lower in lignin than legumes, which may have an effect on digestibility. In this study, due to the lower legume contribution, marginal improvements in NDF and ADF concentrations and a decreased IVTDMD was observed. Titterton and Maasdorp [39] reported a similar effect when ensiling maize with soybean, and Contreras-Govea et al. [37] when ensiling maize with different climbing beans.

The relative feed quality (RFQ) is used to predict the dry matter intake and available energy when forage is fed as a source of energy and protein [17]. The RFQ categories are >140, 110–139, 90–109 and <90, considered as premium, good, fair and low, respectively. In 2010, neither intercrop legume species nor sorghum–legume row proportions had any influence on RFQ. Values ranged from 101–103, considered as fairly good forage quality. However, in 2009, RFQ values were 138–146, adjudged as premium quality. Significantly higher RFQ values were recorded in sorghum–legumes in a 1:1 row proportion (146.3) and in sorghum–lablab combinations (143.3) than in the rest of the treatment combinations.

4. Conclusions

Results of study showed that the LI and LAI were increased early in the season and thereafter reached a plateau. Sorghum growth was also faster than legume growth later in the season, resulting in a lower legumes DM contribution. Sorghum plant density of 250,000 ha⁻¹ in a 1:1 arrangement, with cowpea and lablab as intercrops, achieved a higher LI (>90%) much earlier than rest of the treatments. Sorghum forage yield was greater at 250,000 plants ha⁻¹ and over 25% lower than normal in 1:1 or mixed arrangements. It was higher with pigeon pea than with cowpea or lablab mixtures. At final sampling, the legume contribution was 10–12% of the total forage yield. It was further improved in the 2:2 arrangement, with 190,000 sorghum plants ha⁻¹ and with the lablab mixture. The CP concentration of the sorghum mixture ranged from 62 to 75 g kg⁻¹ DM. Across seasons and treatment combinations, the fiber concentration and digestibility was not improved to any great extent. Sorghum and legumes DM yields and quality can be improved by altering the crop geometry, sorghum plant density and potential annual legumes for intercropping. Cowpea and lablab are potential intercrops for improving sorghum mixtures and the nutritive value of the silage.