Yield and nutritional value of silage of different sorghum hybrids inoculated with Azospirillum brasilense

ABSTRACT

The objective of this study was to evaluate sorghum hybrids associated or not associated with Azospirillum brasilense and nitrogen fertilization (N) during planting on the yield, fermentative profile, and nutritional value of the respective silages. Five sorghum hybrids (Volumax, 201813B, 201814B, 201709B, BRS716) were evaluated with three nitrogen fertilization strategies using urea (100 kg ha⁻¹ of N) and Azospirillum brasilense, and urea (100 kg ha⁻¹ of N)/A. brasilense in association. A randomized block design was used in a 5 × 3 factorial scheme, with five hybrids, three fertilization strategies and three replications (blocks). The useful area of each experimental unit was 3 m x 3 m. The biomass sorghum hybrids showed a dry matter (DM) production (P = 0.01) 48.31% higher than the DM production of the Volumax forage (mean of 17.49 t ha⁻¹ of dry matter). There was no difference between the sorghum hybrid silages in the pH values (mean of 4.11; P = 0.68), gas losses (mean of 3.74% of DM; P = 0.19). The sorghum hybrids biomass 201709B and BRS 716 showed better digestible and DM productivity. Azospirillum brasilense can be used as a nitrogen fertilization strategy in partial or total replacement of urea.

KEYWORDS:

• Diazotropic bacteria
• nitrogen fertilization
• ruminal kinetics
• semi-arid
• sorghum biomass

Highlights

• The biomass sorghum hybrids showed a dry matter production

• Azospirillum brasilense can be used as a nitrogen fertilization strategy

• Sorghum inoculated during planting with Azospirilum brasilense does not modify the fermentation profile of the silage

Introduction

Changes in climatic conditions and water deficits will create major constraints in the agricultural sector, particularly in the semiarid regions of the world (Sabertanha et al. 2021). Forage plants that tolerate variations in environmental conditions and have lower water requirements are necessary to maintain agricultural sustainability and animal production in these areas (Behling Neto et al. 2017; Ribas et al. 2021; Souza et al. 2021; Queiroz et al. 2021).

There are several forages with potential in animal feed; however, sorghum (Sorghum bicolor (L.) Moench) stands out, requiring up to 53% less water (McCorkle et al. 2007) compared to corn (traditional source of energy) for animals), becoming an important crop in arid and semiarid regions. Unlike maize, sorghum grows after cutting, providing a second harvest (Yosef et al. 2009), increasing productivity and reducing dry matter costs per hectare.

According to the Food and Agriculture Organization - FAO (2019), 40.07 million hectares are cultivated with different types of sorghum worldwide. The American (38.3%) and African (39%) continents together account for 77.3% of world production (57.9 million tons). In Brazil, approximately 817.9 thousand hectares have been cultivated with sorghum annually, most of which are used for grain and forage production for animal feed (Conab 2020).

In Brazil, to reduce animal feed costs, especially in intensive production systems, the search for forages with high potential for mass production per unit of area and good nutritional value has been the objective of study in several institutions (Monção et al. 2019; Monção et al. 2020; Cordeiro et al. 2023). However, studies on the productivity and nutritional quality of sorghum hybrids for silage production under semiarid conditions are still scarce.

Urea has been widely used as a source of nitrogen (N) in agriculture. However, the acquisition cost of this input has increased production costs. Alternative sources of N aimed at reducing fertilization costs have been evaluated, highlighting the use of bacteria of the species Azospirillum brasilense that fix atmospheric N in the soil, thus favouring plant growth (Andrade et al. 2019). The ensiling process is considered metabiosis, with competition and succession of groups of microorganisms during the fermentation process. The increase in the Azospirillum population modifies the epiphytic flora and, consequently, can influence the fermentation process of the respective silages. We hypothesized that the use of Azospirillum brasilense can partially or totally replace urea as a nitrogen source for sorghum hybrids grown in a semiarid region.

Based on the above, the objective was to evaluate sorghum hybrids associated or not associated with Azospirillum brasilense and nitrogen fertilization during planting on the yield, fermentative profile, and nutritional value of the silage produced.

Materials and methods

Experiment location

The experiment was carried out at the UNIMONTES Experimental Farm in the municipality of Janaúba (geographical coordinates: 15° 52’38 ‘S, 43° 20’05’ W) for two years (2018/2019; 2019/2020). According to Köppen (1948), the climate in the region is of the BSh type with summer rains and well-defined winter dry periods. The average annual rainfall is up to 800 mm, with an average annual temperature of 27 °C. The climate is tropical mesothermal, almost megathermic, due to the altitude, subhumid and semiarid, with irregular rainfall, causing long periods of drought (Antunes, 1994). Climatic data during the experimental period are shown in Figure 1.

Figure 1. Climatic data during the experimental period in the city of Janaúba. Source: National Institute of Meteorology [INMET] (2020).

Management of the experimental area

The experiment was carried out in a flat area (50 m x 100 m) with different sorghum hybrids (Sorghum bicolor (L.) Moench), established in eutrophic red-yellow clayey soil, 54% clay, 17% silt, 29% sand, with the following chemical characteristics: pH in CaCl₂, 6.4; P (Mehlich), 21.2 mg dm⁻³; K (Mehlich), 110 mg dm⁻³; Na (Mehlich), 0.3 cmolc dm⁻³; Ca ²⁺ 3.9 cmolc dm ⁻³; Mg ²⁺ 1.1 cmolc dm ⁻³; Al ³⁺ 0.0 cmolc dm⁻³; H + A1 (calcium acetate 0.5 mol L⁻¹), 1.2 cmolc dm⁻³; sum of bases 5.5 cmolc dm⁻³; cation exchange capacity, 6.7 cmolc dm⁻³; and base saturation (V), 80%. Soil samples were collected for analysis 70 days before planting.

Treatments and experimental design

Five sorghum hybrids (Volumax, 201813B, 201814B, 201709B, BRS716) associated with three fertilization strategies using urea (100 kg ha⁻¹ of N) and Azospirillum brasilense and urea (100 kg ha⁻¹ of N) /A. brasilense association were evaluated (Table 1). Among the sorghum genotypes evaluated, Volumax is of forage aptitude, and the other hybrids are classified as Biomass. A randomized block design was used in a 5 × 3 factorial scheme, with five hybrids, three fertilization strategies and three replications. The variation between the different plots of soil was the blocking factor. The useful area of each experimental unit was 3 m x 3 m.

Table 1. Combinations of treatments under study considering sorghum genotypes and types of fertilization.

Sorghum hybrids	Types of fertilization
Volumax	Urea¹	Azospirillum brasilense²	Urea/A. brasilense¹^,²
201813B	Urea	Azospirillum brasilense	Urea/A. brasilense
201814B	Urea	Azospirillum brasilense	Urea/A. brasilense
201709B	Urea	Azospirillum brasilense	Urea/A. brasilense
BRS716	Urea	Azospirillum brasilense	Urea/A. brasilense

¹100 kg ha⁻¹ of N

²A. brasilense (Strains Ab-V5 and Ab-V6) were used at a concentration of 2 × 10⁸ CFU mL⁻¹.

Sorghum planting and management

Sorghum was planted with seeds donated by Embrapa Corn and Sorghum. Ploughing and two harrowings were carried out as soil preparation before planting to standardize the area. During the planting phase, nitrogen/phosphorus/potassium (NPK) fertilizer (04-30-10) was applied to meet the requirements of the sorghum crop (250 kg ha⁻¹). Aerial sprinkler irrigation was used (flow rate 0.6 m³/hour; 11.93 mm·hour⁻¹; reach of 4 metres (radius) for two hours daily). Irrigation frequency was based on soil moisture based on the reference daily evapotranspiration (7 mm). Weeds and insects were controlled by applying herbicides and insecticides based on atrazine and deltamethrin, respectively, using a sprayer coupled to the tractor.

The hybrids were sown manually at a depth of 0.03 m, observing a row spacing of 0.70 m and a seed rate of 6.5 kg ha⁻¹ (140.000 plant ha⁻¹), according to the manufacturer's recommendation. Nitrogen fertilization via urea was carried out 30 days after sowing, with a single application of 100 kg ha⁻¹ of nitrogen carried out by broadcast. After fertilizer application, the culture received a water slide for 2 h to mitigate the losses by volatilization.

Azospirillum brasilense was sprayed four days after nitrogen fertilization with the aim of not interfering with the nitrogen fertilization. Foliar spraying was carried out until the plant drained. A. brasilense (Strains Ab-V5 and Ab-V6) were used at a concentration of 2 × 10⁸ CFU mL⁻¹. A specific culture medium and commercially acquired bacteria were used. The culture medium was added to the bioreactor together with a dose of A. brasilense with 0.1% of the bacterial volume concentration of the commercial product under ideal agitation and aeration conditions.

Measurements of structural characteristics and productivity

Measurements and harvesting of the different sorghum hybrids were performed when the grains were in the pasty (Volumax) and milky (Biomass) stages. Plant height was measured at five points per plot using a measuring tape graduated in centimetres at the time of harvesting each hybrid, measuring from the soil to the insertion of the leaf blade of the last leaf of the plants. A 1-m² metal frame was used to manually collect forage samples 20 cm above the ground and estimate the green matter production (GMP) per area. GMP was estimated based on the number and spacing between lines. The samples were pre-dried in an oven with forced air ventilation at 55 °C for 72 h. Dry matter production was estimated based on GMP multiplied by DM content.

Silage production

To ensile the different sorghum hybrids, experimental silos of known weight polyvinyl chloride (PVC) were used, measuring 50 cm in length and 10 cm in diameter. The bottom of the silos contained 10 cm of dry sand (400 g), which was separated from the forage by foam to quantify the effluent produced. The material resulting from each treatment was deposited in the silos and compacted with a wooden plunger. For each treatment, silage density was quantified (550 kg of natural material m⁻³), and approximately 4 kg of chopped material was deposited from each fresh forage, as recommended by Sucu et al. (2016). After filling, the silos were closed with PVC lids equipped with Bunsen valves, sealed with adhesive tape and weighed. The silos were stored at room temperature and opened 120 days after ensiling.

Fermentative losses

Dry matter losses in silage in the form of gases and effluents were quantified by weight difference according to Jobim et al. (2007). For the effluent loss, Equation 1 was used. (1) $E=\left(\mathrm{Pab}-\mathrm{Pen}\right)/\left(\mathrm{GMfe}\right)\times 1000$(1) where:

E: effluent production (kg t⁻¹ of green mass-GM); Pab: weight of the set (silo + cover + wet sand + foam) at opening (kg); Pen: weight of the set (silo + cover + dry sand + foam) in silage (kg); GMfe: green mass of silage forage (kg).

The loss of dry matter in the form of gases was calculated by the difference between the initial and final gross weight of the ensiled dry matter in relation to the amount of DM ensiled, discounting the weight of the silo and dry sand set, according to Equation (2): (2) $G=\left[\left(\mathrm{PCen}\text{–}\mathrm{Pen}\right)\ast \mathrm{DMen}\right]-\left[\left(\mathrm{PCab}-\mathrm{Pen}\right)\ast \mathrm{DMab}\right]\times 100\left[\left(\mathrm{PCen}\text{–}\mathrm{Pen}\right)\ast \mathrm{DMen}\right]$(2) where:

G: gas losses (% DM); PCen: weight of the silo filled with silage (kg); Pen: weight of the set (silo + cover + dry sand + foam) in silage (kg); DMen: dry matter content of forage in silage; PCab: weight of the full silo at opening (kg); DMab: dry matter content of forage at opening. The DM recovery for each silo was calculated based on the initial and final DM weight of the forages and silage according to Jobim et al. (2007).

Assessment of pH and organic acids

The determination of pH and organic acids (Pryce 1969) was carried out in the furrows of the silage obtained using a 16 t mechanical press. The pH was measured with a potentiometer (DM-22, Digimed, São Paulo, SP, Brazil). Volatile fatty acids were determined by liquid chromatography (Shimadzu® Prominence System Model 20A, Kyoto, Japan) equipped with an ultraviolet–visible (UV-Vis) detector set at 210 nm, automatic injector calibrated for 5-μL sample volume and 300 x RezexTM ROA-Acid Column Organic + 7.8 mm (Phenomenex, Torrance, CA, USA) kept at 60 °C in an oven. The analytes were diluted with 2.5 mM H₂SO₄ at a flow rate of 0.6 mL min⁻¹. External standards were used for quantitative calibration purposes.

Chemical composition and ruminal kinetics

A part of each silage was pre-dried in a forced ventilation oven at 55 °C. Subsequently, the samples were ground in a knife mill with a mesh sieve with perforations of 1 mm in diameter for laboratory analysis and in a sieve with sieves of 2 mm in diameter for in situ incubation. The samples were analysed for dry matter content (INCT-CA G-001/2 and G-003/2), crude protein (CP) and soluble fraction (INCT-CA N-001/2), ether extract (INCT-CA G −004/2), and ash (INCT-CA M-001/2), neutral detergent fibre (NDF; INCT-CA F-001/2) and acid detergent fibre (ADF; INCT-CA F-003/2), with the necessary corrections for ash (INCT-CA M-002/2) and proteins (INCT-CA N-004/2), indigestible neutral detergent fibre (NDFi) (INCT-CA F-008/2), contents of insoluble nitrogen compounds in neutral detergent and in acid detergent, lignin (INCT-CA F-005/2) and nonfibrous carbohydrates, following the recommendations described in Detmann et al. (2021). The content of total digestible nutrients (TDN) was estimated according to NRC (2001). For the ruminal kinetics assay, the methodology (Method G-009/1) described by Detmann et al. (2021) was used. Two rumen-cannulated crossbred steers were used, with an average body weight of 500 ± 30 kg and an average age of 8 years. The animals were adapted for 14 days to a diet containing 4 kg of concentrate (25% CP and 65% TDN), divided into two meals, in the morning and in the afternoon, in addition to the supply of forage based in sorghum silages. The roughage:concentrate ratio diet was 80:20 based on DM. Water and mineral salt were also provided ad libitum. The DM intake of the animals was estimated at 2.3% of body weight, and the average pH and ruminal ammoniacal nitrogen at the time of incubation were 6.98 and 13.04 mg d L⁻¹, respectively. The in situ degradability technique was performed using 7.5 × 15 cm nonwoven fabric bags (weight 100) with an approximate porosity of 60 μm, according to Casali et al. (2009); the number of samples was determined from the ratio of 20 mg DM·cm^-² bag surface area (Nocek 1988). The samples were deposited in the region of the ventral sac of the rumen for 0, 3, 6, 12, 24, 48, 72, 96, 120, and 144 h, with the end of the nylon thread tied to the cannula. The bags referring to time zero were not incubated in the rumen but were washed in running water, similar to the incubated bags. All samples were removed and washed in cold water to stop ruminal fermentation. Subsequently, the samples were placed in an oven at 55 °C for 120 h and then cooled in a desiccator and weighed. The residues remaining in the nonwoven fabric bags collected in the rumen were analysed for MS and NDF contents according to the aforementioned methodology. The percentage of degradation was calculated by the proportion of feed remaining in the bags after ruminal incubation. The data obtained were adjusted for nonlinear regression using the Gauss–Newton method and SAS 9.0 software (SAS Institute Inc., Cary, NC) according to the equation proposed by Detmann et al. (2021): Dt = A + B x (1-e^{-c x t}), where Y = accumulated degradation of the analysed nutrient component after time t; a = intercept of the degradation curve when t = 0, which corresponds to the water-soluble fraction of the analysed nutritional component; b = potential for the degradation of the water-insoluble fraction of the analysed nutritional component; a + b = potential degradation of the analysed nutrient component when time is not a limiting factor; c = fractional rate of degradation (h⁻¹); and t = incubation time. After being calculated, the coefficients a, b and c were applied to the equation proposed by Detmann et al. (2021): ED = a + (b x c/c + k), where ED = effective ruminal degradation of the analysed nutrient component and k = rate of passage of particles in the rumen estimated at 5% h⁻¹ (AFRC 1993).

Principal component analyses

A principal component analysis (PCA) was applied to better understand the nature of the relationship between the variables studied and the independent variables. For this analysis, 31 studied characteristics were considered. From the correlation matrix between the characteristics, the data were submitted to PCA, in which the variables were standardized for mean equal to zero and variance equal to one. A correlation matrix was used instead of a covariance matrix (Johnson and Wichern 2007). The method proposed by Kaiser (1960) was used to select which main components best simplified the variability present in the dataset and to compose the other analyses and interpretations. In this method, eigenvalues equal to or greater than one (1) were kept since the original variables also have a variance equal to one after being standardized. The first six CPs presented eigenvectors above 1 that explained 97.45% of the total variance of the results (Figure 2).

Figure 2. Screen plot of eigenvalues corresponding to each of the 7 principal components with variance greater than 0.8%.

Statistical analysis

Data were analysed in SISVAR® with a model containing the fixed effects of sorghum hybrids and fertilization strategy (treatments). Means were pooled using the Scott-Knott test at 5% probability. The UNIVARIATE procedure was used to detect outliers or influential values and to examine the normality of residuals.

The variables related to the fermentation profile and chemical composition were analysed according to the model: $\mathrm{Yijk}=\mu +\mathrm{Ti}+\mathrm{ADj}+\mathrm{Ti}\times \mathrm{ADj}+\mathrm{eijk},$where:

Yijk = observed value for variable ‘i’ in relation to the hybrid and fertilization ‘j’ in repetition k;

μ = mean of all experimental units for the variable under study;

Ti = effect of hybrid ‘i’ with i = 1,2,3,4 and 5;

ADj = effect of fertilization strategies ‘j’ with j = 1,2 and 3;

Ti x ADj = interaction effect;

eijk = error associated with independent Yijk observation, which by hypothesis has a normal distribution with mean zero and variance δ².

The DM and NDF ruminal degradability assays were carried out in a randomized block design in split plots, with fifteen treatments (5 hybrids x 3 fertilization strategies; plots) and 10 incubation times (subplots). The variation in body weight of each animal was the blocking factor. Data were analysed in SISVAR® with a model containing the fixed effects of treatments. Means were compared using the Scott-Knott test at 5% probability.

The following statistical model was used: $\mathrm{Yijk}=\mu +\mathrm{Ti}+\mathrm{Bj}+\mathrm{eij}+\mathrm{Pk}+T\times \mathrm{Pik}+\mathrm{eijk}$

where:

Yijk = The observation regarding time (P) in subplot k of treatment (T) i in block j;

μ = constant associated with all observations;

Ti = Effect of treatment ‘i’, with i = 1, 2, 3, 4 … , 15; Bj = Effect of block j, with j = 1 and 2;

eij = experimental error associated with the plots that, by hypothesis, have a normal distribution with zero mean and variance δ2;

P = Effect of incubation time k, with k = 1,2,3,4,5,6,7,8,9 and 10;

TPik = Effect of the interaction of Treatment level i with the Incubation Time level k;

Eijk = experimental error associated with all observations that by hypothesis have a normal distribution with zero mean and variance δ².

Results

There was no interaction of sorghum hybrid factors and nitrogen fertilization strategies (P = 0.99) on the structural, productive and chemical characteristics of the in natura sorghum plant (Table 2). The fertilization strategies with A. brasilense and/or urea did not change any of the aforementioned characteristics (P = 0.96), except for the ash content (P = 0.04), which was 7.21% higher in the fertilization with A. brasilense and A. brasilense with urea compared to fertilization with urea alone (average of 64.3 g kg DM⁻¹). Hybrids 201814B and BRS 716 showed plant heights 59.21% and 60.20% higher than Volumax, respectively. Dry matter production and dry matter content were higher in biomass hybrids than in Volumax. There was no difference between sorghum hybrids in crude protein (mean 64.09 g kg DM⁻¹) and neutral detergent fibre (687.5 g kg DM⁻¹) content.

Table 2. Structural, productive and chemical characteristics of different sorghum hybrids fertilized with Azospirillum brasilense with or without urea at the time of ensiling.

Item (g kg DM^−1)	Fertilization ^1			SEM^2	Sorghum hybrids					SEM	P value^3
	Az	U	Az + U		Volumax	201813B	201814B	201709B	BRS716		Hib	Ad	Hib x Ad
Height, m	3.39	3.61	3.45	0.08	1.66c	3.79b	4.07a	3.71b	4.17 a	0.11	<0.01	0.22	0.85
Dry matter yield, t ha^−1	29.58	32.56	29.78	2.08	17.49 b	33.96 a	30.05 a	36.28a	35.08 a	2.68	0.01	0.54	0.88
Crude protein yield, t ha^−1	1.89	1.94	1.90	0.12	1.35 b	2.17 a	1.74 b	2.21 a	2.05 a	0.16	<0.01	0.95	0.29
TDN yield, t ha^−1	14.94	16.73	15.48	1.05	10.04 b	17.62 a	15.15 a	17.97 a	17.60 a	1.36	<0.01	0.48	0.95
Dry matter	246.30	255.70	250.40	4.30	232.30b	251.60a	261.8a	261.90a	254.30 a	5.50	<0.01	0.31	0.49
Ash	69.3a	64.3 b	69.3 a	1.50	81.80 a	63.60b	65.6b	61.30b	66.30b	1.90	<0.01	0.04	0.91
Crude protein	62.8	58.3	70.8	4.10	74.00	63.60	63.07	63.70	56.10	5.10	0.24	0.11	0.99
Neutral detergent fibre	686.6	689.8	686.2	9.80	692.60	670.20	700.40	673.50	700.80	12.70	0.27	0.96	0.30

¹ Nitrogen fertilization strategies (Az- Azospirillum brasilense, Ab-V5 and Ab-V6 concentration of 2 × 108 CFU mL⁻¹; 200 litres ha⁻¹ of solution; U – Urea 46%, 100 kg ha⁻¹ of nitrogen; Az + U – Association of A. brasilense and Urea 46%)

² SEM – Standard error of the mean

³ P- Probability (Hib - effects of hybrids; Ad - effect of fertilization; Hib x Ad - effect of the interaction of factors)

Means followed by the same letter on the line do not differ from each other by the Scott-Knott test (P > 0.05).

Regarding the characteristics related to the fermentative profile of the silages produced with the different sorghum hybrids and fertilization strategies (Table 3), there was no interaction between the factors (P = 0.80) and no isolated effect of fertilization strategies with A. brasilense and urea (P = 0.97). There was no difference between sorghum hybrid silages in pH values (mean of 4.11; P = 0.68), gas losses (mean of 3.74% of DM; P = 0.19), dry matter recovery (936.66 g kg of DM⁻¹; P = 0.29), concentration of acetic acid (average of 2.44 g kg of DM⁻¹), propionic acid (mean of 2.48 g kg of DM⁻¹) and ethanol (mean 3.46 g kg DM⁻¹). The silages of Volumax, 201709B and BRS 716 had higher temperature values (mean of 24.80 °C) than the other silages. In the silages of Volumax and 201709B, higher DM losses were observed in the form of effluents. Sorghum silage 201814B presented a lactic acid concentration and lactic:acetic acid ratio 22.26% and 29.45% higher than the other silages (average of 12.08 g kg of DM⁻¹ and 5.00), respectively. A higher concentration of butyric acid was found in sorghum silage 201813B.

Table 3. Fermentation profile of sorghum hybrid silage fertilized with Azospirillum brasilense with or without urea.

Item	Fertilization ^1			SEM^2	Sorghum hybrids					SEN	P value³
	Az	U	Az + U		Volumax	201813B	201814B	201709B	BRS716		Hib	Ad	Hib x Ad
pH	4.10	4.10	4.13	0.03	4.08	4.06	4.15	4.13	4.13	0.04	0.68	0.84	0.51
Temperature, °C	24.68	24.65	24.68	0.09	24.82a	24.35b	24.58b	24.76a	24.84a	0.12	0.04	0.97	0.36
Gas losses, % DM	3.77	3.73	3.74	0.08	3.90	3.73	3.88	3.61	3.62	0.10	0.19	0.95	0.33
effluent losses, kg of GM/t	34.93	36.88	38.92	2.48	39.85a	36.11b	28.42b	45.00a	33.34b	3.19	0.01	0.37	0.28
DM recovery, g kg of DM^−1	952.50	911.42	943.33	3.22	931.11	984.44	970.00	861.11	936.66	4.16	0.29	0.65	0.59
Lactic acid, g kg of DM^−1	12.83	12.93	12.56	0.79	10.75b	12.31b	15.54a	12.7b	12.57b	1.02	0.04	0.94	0.12
Acetic Acid, g kg of DM^−1	2.31	2.64	2.35	0.19	2.65	2.32	2.20	2.50	2.53	0.25	0.73	0.43	0.80
Lactic:acetic acid ratio	5.19	5.84	5.28	0.39	3.91b	5.32b	7.09a	5.37b	5.42b	0.50	<0.01	0.46	0.74
Propionic acid, g kg of DM^−1	2.45	2.58	2.43	0.15	2.13	2.78	2.40	2.59	2.51	0.20	0.25	0.77	0.19
Butyric Acid, g kg of DM^−1	0.31	0.35	0.47	0.10	0.37 b	0.70a	0.13b	0.44b	0.20b	0.13	0.04	0.52	0.19
Ethanol, g kg of DM^−1	3.71	3.18	3.46	0.35	3.17	3.51	3.19	3.64	3.82	0.46	0.82	0.58	0.70

DM – Dry matter; GM- Green matter

¹ Nitrogen fertilization strategies (Az- Azospirillum brasilense, Ab-V5 and Ab-V6 concentration of 2 × 108 CFU mL⁻¹; 200 litres ha⁻¹ of solution; U – Urea 46%, 100 kg ha⁻¹ of nitrogen; Az + U – Association of A. brasilense and Urea 46%)

² SEM – Standard error of the mean

³ P- Probability (Hib – effects of hybrids; Ad – effect of fertilization; Hib x Ad – effect of the interaction of factors)

Means followed by the same letter on the line do not differ from each other by the Scott-Knott test (P > 0.05).

There was no interaction between sorghum hybrid silages and fertilization strategies (P = 0.99) and no isolated effect of fertilization strategies (P = 0.77) for any of the characteristics of the chemical composition of the silages (Table 4). Sorghum silages 201813B, 201814B and BRS716 had higher dry matter content than the others (average of 243.9 g kg DM⁻¹). The Volumax sorghum silage presented ash and crude protein content 21.83% and 21.00% higher than the ash and crude protein content of the other silages, respectively. There was no difference between the silages in the readily soluble fraction of crude protein (mean of 34.32% of total nitrogen; P = 0.29), ether extract (mean 21.4 g kg of DM⁻¹; P = 0.14) and nonfibrous carbohydrates (mean 193.4 g kg DM⁻¹; P = 0.11). The Volumax sorghum silage presented a total digestible nutrient content 11.38% higher than the total digestible nutrient content of the biomass sorghum silages (mean of 506.27 g kg DM⁻¹; P < 0.01). The silages of sorghum biomass, 201813B, 201814B, 201709B, and BRS716, showed higher levels of neutral detergent fibre (P < 0.01) and acid detergent fibre (P < 0.01) compared to forage sorghum silage (Volumax). The in vitro digestibility of dry matter, neutral detergent fibre and acid detergent fibre was higher in the sorghum silages Volumax and 201814B. The content of indigestible dry matter, indigestible neutral detergent fibre and indigestible acid detergent fibre were higher in sorghum silages 201813B, 201709B and BRS 716.

Table 4. Chemical composition and digestibility of sorghum hybrid silage fertilized with Azospirillum brasilense with or without urea.

Item (g kg DM^−1)	Fertilization ^1			SEM^2	Sorghum hybrids					SEM	P value^3
	Az	U	Az + U		Volumax	201813B	201814B	201709B	BRS716		Hib	Ad	Hib x Ad
Dry matter	258.0	256.2	251.9	3.2	261.8a	256.4b	270.8a	248.3b	261.9a	4.2	<0.01	0.23	0.60
Ash	80.3	75.6	78.3	3.3	94.7a	73.9b	79.4b	70.6b	72.2b	4.3	<0.01	0.61	0.49
Crude protein	65.6	62.6	65.0	2.9	77.5a	64.6b	58.6b	62.9b	58.8b	3.8	0.01	0.77	0.50
Fraction A, % TN	31.71	36.14	35.41	2.49	30.87	33.46	39.39	31.17	36.72	3.21	0.29	0.40	0.45
Ether extract	23.2	17.0	23.0	2.4	28.4	17.3	18.2	21.7	20.4	3.2	0.14	0.16	0.99
Neutral detergent fibre	649.1	650.9	628.5	11.1	576.8b	642.0a	651.9a	674.2a	669.2a	14.3	<0.01	0.30	0.18
Acid detergent fibre	434.3	424.1	406.1	12.4	349.1b	418.5a	441.6a	453.2a	446.3a	15.9	<0.01	0.27	0.26
Nonfibrous carbohydrates	181.6	193.7	204.9	11.0	222.4	202.0	191.7	170.3	179.3	14.2	0.11	0.32	0.12
Total digestible nutrients	511.0	516.4	531.9	8.5	571.3a	520.4 b	505.6 b	497.3b	501.8b	11.0	<0.01	0.19	0.21
In vitro digestibility of DM	602.2	617.7	608.3	18.2	610.6a	541.3b	583.4b	679.2a	630.7 a	23.4	0.04	0.83	0.83
DMY digestible, t ha^−1	17.69	20.07	18.24	1.41	10.69c	18.38b	17.54 b	24.57 a	21.89 a	1.82	<0.01	0.47	0.89
In vitro digestibility of NDF	522.4	552.6	535.2	20.2	541.7 a	444.8 b	512.2 b	584.3a	597.0 a	26.0	<0.01	0.57	0.98
In vitro digestibility of ADF	432.2	457.9	497.8	21.7	487.3a	361.3b	428.7b	508.9 a	524.2 a	42.6	<0.01	0.11	0.95
iDM	358.3	370.4	347.5	7.3	321.7b	364.9a	345.9b	375.4a	384.5a	9.5	<0.01	0.11	0.31
iNDF	282.1	294.0	275.9	5.4	248.4c	275.9b	285.6b	300.8a	308.1a	7.0	<0.01	0.08	0.23
iADF	226.8	235.6	219.0	4.6	196.9c	220.8b	228.6b	242.8 a	245.6a	6.0	<0.01	0.06	0.16

TN – total nitrogen; DM – dry matter; iDM – indigestible dry matter; NDF – neutral detergent fibre; iNDF – indigestible neutral detergent fibre; ADF – acid detergent fibre; iADF – indigestible acid detergent fibre.

¹ Nitrogen fertilization strategies (Az- Azospirillum brasilense, Ab-V5 and Ab-V6 concentration of 2 × 108 CFU mL-1; 200 litres ha⁻¹ of solution; U – Urea 46%, 100 kg ha⁻¹ of nitrogen; Az + U – Association of A. brasilense and Urea 46%)

² SEM – Standard error of the mean

³ P- Probability (Hib - effects of hybrids; Ad - effect of fertilization; Hib x Ad - effect of the interaction of factors)

Means followed by the same letter on the line do not differ from each other by the Scott-Knott test (P > 0.05).

There was no interaction between the silages of different sorghum hybrids and fertilization strategies on the fraction readily soluble in water (Fraction ‘a’), rate of degradation of the insoluble fraction (fraction ‘b’) ‘c’ and effective degradability of dry matter (Table 5). The smallest fraction ‘a’ of dry matter was observed in Volumax sorghum silage (mean of 22.56%), and the other silages did not differ from each other (mean of 26.70%). There was an interaction of factors on the insoluble, potentially degradable fraction (fraction ‘b’) and potential degradability of dry matter. BRS 716 sorghum silage associated with fertilization using Azospirilum brasilense or Azospirilum brasilense with urea showed lower fraction ‘b’ contents and potential degradability of dry matter. Among the Volumax silages, the one with fertilization with only urea was verified to have the lowest average potential degradability of dry matter. Sorghum silages 201813B and 201814B were observed to show higher effective degradability of dry matter (mean of 41.49%) compared to other silages (average of 38.59%).

Table 5. Ruminal kinetics of silage dry matter of sorghum hybrids fertilized with Azospirillum brasilense with or without urea.

Item (%)	Fertilization ¹	Sorghum hybrids					SEM²	P value³
		Volumax	201813B	201814B	201709B	BRS716		Hib	Ad	Hib x Ad
Fraction a	Az	24.63	28.45	29.72	27.03	24.49	1.63	<0.01	0.18	0.69
	U	21.01	27.63	27.72	25.41	27.49
	Az + U	22.04	27.86	25.93	24.15	24.57
Fraction b	Az	55.99Aa	54.44Aa	53.2Aa	52.61Aa	44.63Bb	1.78	<0.01	0.11	0.01
	U	50.43Aa	51.47Aa	54.24Aa	50.23Aa	50.59Aa
	Az + U	54.5Aa	52.08Aa	54.39Aa	43.23Bb	44.87Bb
Degradation rate c, % h^−1	Az	1.75	1.75	1.75	1.50	2.25	<0.01	0.32	0.28	0.19
	U	2.00	1.75	1.25	1.75	1.50
	Az + U	1.50	1.50	1.75	3.50	2.75
Potential Degradability	Az	80.62Aa	82.89Aa	82.92Aa	79.64Aa	69.12Bb	2.26	<0.01	0.01	<0.01
	U	71.44Bb	79.10Aa	81.97Aa	75.64Ab	78.08Aa
	Az + U	76.55Aa	79.95Aa	80.32Aa	67.38Bb	69.45Bb
Effective degradability, k = 5% h^−1	Az	39.56	41.75	44.14	39.02	38.96	1.44	0.01	0.16	0.45
	U	36.567	40.43	40.11	38.04	39.4
	Az + U	35.58	41.07	41.49	41.55	38.69

¹ Nitrogen fertilization strategies (Az- Azospirillum brasilense, Ab-V5 and Ab-V6 concentration of 2 × 108 CFU mL⁻¹; 200 litres ha⁻¹ of solution; U – Urea 46%, 100 kg ha⁻¹ of nitrogen; Az + U – Association of A. brasilense and Urea 46%)

² SEM – Standard error of the mean

³ P- Probability (Hib - effects of hybrids; Ad - effect of fertilization; Hib x Ad - effect of the interaction of factors)

Means followed by the same letter on the line do not differ from each other by the Scott-Knott test (P > 0.05)

There was an interaction between sorghum hybrid silages and nitrogen fertilization strategies on the standardized insoluble fraction, potentially degradable (fraction Bp; P < 0.01) and degradation rate ‘c’ (P < 0.01) of detergent fibre neutral (Table 6). Regardless of the fertilization strategy used, there was no difference between sorghum silages 201813B and 201814B on fraction Bp. Among the Volumax sorghum silages, the highest averages of fraction Bp were verified in the fertilization strategies with Azospirilum brasilense and Azospirilum brasilense associated with urea. Fertilization with Azospirilum brasilense combined with urea improved the rate of degradation of the fibrous fraction of sorghum 201709B and BRS 716 silages. There was greater effective degradability of neutral detergent in sorghum silages fertilized with Azospirilum brasilense (mean of 32.81%) in relation to fertilization with urea or Azospirilum brasilense with urea (mean of 30.82%). Sorghum silages 201813B and 201814B showed higher effective fibre degradability (mean of 33.93%) than sorghum silage BRS 716 (mean of 31.62%). The lowest effective degradability of neutral detergent fibre was verified in Volumax sorghum silage.

Table 6. Ruminal kinetics of neutral detergent fibre in the silage of sorghum hybrids fertilized with Azospirillum brasilense with or without urea.

Item (%)	Fertilization ¹	Sorghum hybrids					SEM²	P value³
		Volumax	201813B	201814B	201709B	BRS716		Hib	Ad	Hib x Ad
Fraction Bp	A	71.47Ab	77.27Aa	75.13Aa	71.22Ab	61.40Bc	1.68	<0.01	0.01	<0.01
	U	65.12Bb	71.56Aa	73.41Aa	67.49Ab	69.89Aa
	A + U	70.84Aa	75.02Aa	73.35Aa	60.00Bb	60.40Bb
Degradation rate c, % h^−1	A	1.50Ab	1.50Ab	2.00Ab	1.50Bb	3.00Aa	<0.01	<0.01	0.11	<0.01
	U	1.75Aa	1.75Aa	1.50Aa	2.00Ba	1.75Ba
	A + U	1.50Ab	1.50Ab	2.00Ab	3.25Aa	2.50Aa
Effective degradability, k = 5%	A	29.17	33.71	37.22	30.78	33.19	1.27	<0.01	<0.01	0.24
	U	25.78	32.8	31.15	29.24	32.00
	A + U	27.46	33.79	34.93	31.53	29.61

¹ Nitrogen fertilization strategies (Az- Azospirillum brasilense, Ab-V5 and Ab-V6 concentration of 2 × 108 CFU mL⁻¹; 200 litres ha⁻¹ of solution; U – Urea 46%, 100 kg ha⁻¹ of nitrogen; Az + U – Association of A. brasilense and Urea 46%)

² SEM – Standard error of the mean

³ P- Probability (Hib - effects of hybrids; Ad - effect of fertilization; Hib x Ad - effect of the interaction of factors)

Means followed by the same letter on the line do not differ from each other by the Scott-Knott test (P > 0.05).

In the analysis of the principal components of the different sorghum hybrids (Figure 3), components CP1 (57.78%) and CP2 (16.48%) were found to present eigenvalues that explained 74.26% of the data variance.

Figure 3. Schematic representation of the first (CP1) and second (CP2) principal components of the dependent variables analysed in the different sorghum hybrids.

Within CP1, the digestibility of dry matter (0.59), neutral detergent fibre (0.57) and acid detergent fibre (0.51) was found to present higher weighting coefficients that explained 57.78% of the variance in the results. In CP2, the production of dry matter, losses by effluents and total digestible nutrients presented coefficients of 0.80, 0.27 and 0.17, respectively. Based on Figure 3, the hybrids that showed higher in vitro dry matter digestibility and dry mass productivity stood out in the exploratory data analysis.

Discussion

Due to edaphoclimatic variations in the semiarid region of Brazil, related mainly to irregular rainfall and high temperatures during the summer season, forage sorghum Volumax has been widely cultivated for forage production. However, other sorghum varieties or hybrids with higher mass yields for silage, such as biomass sorghum, have been sought.

In this research, the forage potential of the sorghum hybrids studied was verified. However, due to genetic and natural selection characteristics, all biomass hybrids (201813B, 201814B, 201709B and BRS716) were found to present higher heights and dry matter yields than Volumax sorghum at the time of ensiling. The difference in dry matter productivity was 17.87 t ha⁻¹ or 49.57% higher than the average observed in Volumax sorghum. The advantage of Volumax sorghum is in the early harvest compared to the hybrids studied, allowing the harvest of the second cycle. Another result verified in this research is that the nitrogen fertilization strategies using Azospirullum brasilense and/or urea did not change the height (average of 3.48 m), dry matter productivity (average of 32.71 t ha⁻¹), dry matter content (average of 250.8 g kg of DM⁻¹), crude protein or neutral detergent fibre of the evaluated hybrids. This information is relevant because Azospirullum brasilense, in sorghum, can contribute to the reduction of production costs compared to chemical fertilization. This result is an important novelty for tropical agriculture, given that countries such as Brazil are still not high enough in the production of chemical nitrogen fertilizers, which have high production costs. Thus, the Azospirillum inoculation could serve as a replacement for urea in sorghum cultivation.

Despite the early harvest for silage of Volumax sorghum considering the pasty grain stage, there was a lower dry matter content compared to biomass sorghum. This lower average can be explained by the low proportion of senescent material in the young Volumax plant compared to biomass sorghum (Souza et al. 2021). The low dry matter content, not only in the Volumax sorghum silage but also in the 201709B sorghum silage, justifies the greater dry matter losses in the form of effluents and allows the postponement of the aerobic stability break since they presented high temperature values after opening the silo. Sorghum silage 201814B had a higher concentration of lactic acid and lactic:acetic acid ratio than the other silages. The production of lactic acid is important during the fermentation process for the rapid reduction of the pH of the ensiled mass, especially after closing the silo, where pH of the ensiled mass presents values close to neutrality. This reduction may have inhibited the growth of undesirable microorganisms (i.e. enterobacteria, filamentous fungi and clostridia), favouring the conservation process in an anaerobic medium. After opening the silo, acetic acid has better control over the growth of these undesirable microorganisms, which increases the time for breaking the aerobic stability.

Regarding the chemical composition of the sorghum hybrids, the better panicle structure and fillers of the Volumax sorghum grains may have contributed to the higher ash concentration in the whole plant in natura and in the silage. In relation to nitrogen fertilization strategies using urea, the lower ash content may be associated with the dilution effect on the plant, which justifies the lower values.

Due to the higher plant growth in sorghum biomass hybrids, there are changes in the cell wall proportion to the detriment of cell content. Therefore, there was higher dry matter content in these sorghum varieties (except for hybrid 201709B), neutral detergent fibre and acid detergent fibre in relation to Volumax sorghum silage. The higher fibre content in the sorghum silage biomass of the hybrids studied favoured a reduction in the content of total digestible nutrients, even under harvesting conditions with panicle emission. The panicle of sorghum biomass verified in this study was lower than the panicle of sorghum biomass of Volumax sorghum, which contributes to a lower proportion of starch in the whole plant. Even so, sorghum silage Volumax and sorghum silage 201814B stood out with higher in vitro digestibility of dry matter and fibrous fraction and lower content of indigestible dry matter. This information is relevant because in the principal component analysis, the aforementioned digestibility associated with dry matter yield was identified as the most important dependent variable in the selection of hybrids for silage production. Based on the principal components (CPs) presented in Figure 3, CP1 (57.78%) and CP2 (16.48%) represent 74.26% of the total variation of the data, which highlights the hybrids 201814B and BRS 716 as promising for silage in the semiarid region due to the high productivity of dry matter digestible in relation to the others studied.

Regarding fertilization strategies, the digestible dry matter productivity was found to be the same (average of 18.66 t ha⁻¹). Under the conditions of this study, the use of Azospirillum brasilense seems promising for farmers because it can be produced on the farm itself. Azospirillum brasilense improves nutrient utilization by sorghum plants during nutrient cycling and incorporates atmospheric nitrogen into the soil, making it available to the plant. Nitrogen is the most important nutrient for increasing mass productivity in plants because this molecule is directly involved in mitotic cell division processes, favouring the synthesis of amino acids.

For the effective degradability of the dry matter of the silages, the highest averages were verified in the hybrids 201813B and 201814B, mainly influenced by the higher fraction ‘b’, since fraction ‘a’ was not different between treatments (mean of 25.87%). The ‘a’ fraction of dry matter represents the readily soluble nutrients (i.e. soluble nitrogen, water-soluble carbohydrates) with the highest participation of nonfibrous carbohydrates, especially water-soluble carbohydrates. In different sorghum hybrids, Behling Neto et al. (2017) found that the soluble carbohydrate content can vary from 4.54% to 12.81% (on a DM basis). The higher concentration of lactic acid in sorghum silage 201814B is indicative of a high content of water-soluble carbohydrates because these are used by homolactic bacteria to produce this acid.

Conclusions

All sorghum hybrids evaluated in this study have potential for ensiling. The sorghum hybrid with the best response in productivity and nutritional value of silage was biomass 201814B and BRS 716.

Azospirillum brasilense can be used as a nitrogen fertilization strategy in partial or total replacement of urea in sorghum culture, but it does not influence the fermentative characteristics of the respective silages.

Compliance with ethical standards

This study was conducted in the Experimental Feedlot of the State University of Montes Claros. Experimental protocol (no 173/2018) and animal-use procedures were approved and followed guidelines recommended by the Animal Care Committee of the same institution. The manuscript does not contain clinical studies or patient data.

Acknowledgements

The authors are grateful to the Minas Gerais Research Foundation (FAPEMIG), Montes Claros State University (Unimontes), National Council for Scientific and Technological Development (CNPq), EMBRAPA-Milho and Sorghum and the National Institute of Science and Technology (INCT – Animal Science) for financial support and granting scholarships and other resources. This project was partially financed by the Coordination for the Improvement of Higher Education Personnel - Brazil (CAPES) - Financial Code 001.

Data availability statement

The authors confirm that the data supporting the findings of this study are available within the article [and/or] its supplementary materials.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This study was financed in part by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - Brasil (CAPES ) - Finance Code 001. Conselho Nacional de Desenvolvimento Científico e Tecnológico