Productive and Qualitative Traits of Sorghum Genotypes Used for Silage under Tropical Conditions

: The aim of this study was to evaluate the productive potential, chemical composition, and fermentation profile of 24 genotypes of forage sorghum after ensiling. For this agronomic evaluation, a completely randomized block design with six replicates and twenty-four treatments was employed. Genotype 5 had the highest dry matter (DM) yield of 22.24 t/ha. The plant DM content ranged ( p < 0.001) from 271.8 g/kg of fresh matter (FM) in genotype 9 to 330.0 g/kg FM in genotype 3, averaging 302.9 g/kg FM. The crude protein and non-fibrous carbohydrates differed significantly ( p < 0.001) in the evaluated silages, with mean values of 61.0 and 275.3 g/kg DM, respectively. The pH of the sorghum silages also differed ( p < 0.001), despite having an average of 3.94, with values ranging from 3.68 to 4.27. No changes were observed ( p > 0.05) for the dry matter recovery of the sorghum genotype silages, which averaged at 853.9 g/kg FM. In the present study, genotype 5 resulted in being the most recommended genotype because it demonstrated a higher yield and higher nutritional value in both its fresh form and its ensiled form; had the lowest losses; and showed an ideal dry matter recovery after ensiling of 880 g/kg.


Introduction
Since 1970, there has been a considerable increase in the temperature on the Earth's surface of about 0.15-0.20 • C per decade, with predictions forecasting a 0.25 • C increase for the period of 2020-2050 [1].This increase in temperature is a reflection of climate change, which is promoted by environmental pollution from anthropogenic activities that result in land degradation and desertification in arid and semi-arid climates [2].At the same time, in the regions of Brazil with these climates, the amount of rainfall per year is approximately 800 mm [3].
Variations in the climate and in soil resulting from these climatic conditions promote decreases in the availability of feed sources or feeds with nutritional quality that can meet the nutritional requirements of ruminants [4].Therefore, in these regions, one of the main options is to produce high-quality food in the rainy season and preserve it, with silage being the main form of conservation and food source in these climates considering the characteristics of the plants [5].
Sorghum or Sorghum bicolor [(L.)Moench] is a crop of great relevance worldwide that belongs to the Gramineae family and the C4 metabolism group, with the important Crops 2024, 4 257 characteristic of being able to adapt to different climatic conditions and types of soil [6].There is a large number of sorghum hybrids, which are distributed into four categories: sweet sorghum, grain sorghum, biomass sorghum, and forage sorghum [7].
The choice of which type of sorghum to cultivate was once based on its ability to generate a high yield.The low cost of producing higher-yielding sorghum types was considered their main advantage, besides their nutritional value [8].The development of sorghum varieties has enabled the production of hybrids better suited for ensilage, aiming not only at good dry matter yields but also at improved nutritional value [9].Further, Pinho et al. [10] consider dry matter losses during ensilage a fundamental parameter in the recommendation of sorghum genotypes.However, before making the decision to produce silage, it is necessary to evaluate the characteristics of the ensiling process that promote the best ideal nutritional profile for feeding ruminate animals that results in greater productivity [11].
Many researchers [10,12] have investigated the performance of sorghum in recent years and indeed observed variations in yield, nutritional value, and losses during the ensiling process.Such studies are important, since the choice of the cultivar is the first step towards achieving high-quality ensilage.Information about the productive behavior of sorghum genotypes is thus essential when these are intended for silage production in order to ensure an efficient and highly technological production system.
Based on previous information, our hypothesis is that there is an ideal genotype for silage production.This ideal genotype should enhance the animal's productive performance and increase the income for farmers.In this sense, our study aims to assess the productive potential, chemical composition, and fermentation process of 24 different sorghum genotypes used for silage production.

Treatments, Soil, and Climate Conditions
Our experiment was developed on a dystrophic cohesive yellow latosol on the Experimental Farm at the Center for Agricultural, Environmental, and Biological Sciences of the Federal University of Recôncavo da Bahia (UFRB), located in Cruz das Almas-BA, Brazil (12 • 40 ′ 19 ′′ S and 39 • 06 ′ 23 ′′ W; 220 m.a.s.l.).The climate in this municipality is a warm tropical climate, varying from a tropical monsoon-type climate to a humid tropical climate, according to the classification proposed by Köppen.The average annual precipitation in the region is 1170 mm, with variations between 800 and 1400 and a greater incidence of rainfall in the period from March to August.The average air humidity is approximately 80%, and the average annual temperature is 24.5 • C. Soil samples were collected from the experimental area to evaluate the chemical attributes at a depth of 0-20 cm.
Soil preparation, fertilizer application at planting, and furrowing were mechanized following the results of the evaluation of soil chemical properties (Table 1).The soil was plowed at a depth of 30 cm and disked twice, followed by the application of 20 kg/ha N, 19.57kg/ha P, and 39.83 kg/ha K. Thirty days after plant emergence, 41.4 kg/ha N was administered.= aluminum; Ca +2 = calcium; Mg +2 = magnesium; SB = base saturation; CEC = cation-exchange capacity; SB = sum of bases; OM = organic matter.The area was irrigated (except on rainy days) using manual 20 L sprinklers to apply a water depth of 2.55 mm until the day before the harvest of each genotype, aiming to reach a minimum volume of 240 mm in the first three months after sowing.

---mg/dm 3 --------------------------cmol c /dm 3 -------------------
A randomized-block experimental design was adopted for the agronomic evaluation, with 3 blocks (replicates) and 24 treatments that corresponded to each sorghum genotype assessed (Table 2).Each block consisted of twenty-four 2.10 × 3.00 m plots, and each plot had three planting rows spaced 0.70 m apart.The seeds were sown manually in September 2014, using 33 lots of seeds in triplicate originating from the Preliminary Forage-Sorghum Trial of the Agronomic Institute of Pernambuco (IPA).The sorghum genotypes used were accessible and these were the genotypes that possessed the attributes necessary for forage production at the time of the investigation.The plants were thinned upon reaching 10 cm in height, resulting in 12 plants per linear meter.Maximum-minimum environmental temperature and precipitation data were obtained from the meteorological station located in the headquarters of EMBRAPA-CNPMF, in Cruz das Almas-BA, Brazil.These data were recorded monthly and during the period of cultivating the sorghum genotypes (Figure 1).
Observations (measurements) were made in the usable plot, which corresponded to 2 m of the center row of each plot, disregarding 0.5 m from each extremity of that row.The sorghum was harvested at the dough-to-dent stage, approximately 80 days after planting.The material collected from each plot was divided into panicles, leaves, and stems and weighed separately.A subsample of each fraction was pre-dried at 65 • C until reaching a constant weight to estimate the moisture content.Afterwards, the material was dried in a forced-air oven at 105 • C for 24 h to determine the dry matter content, following the AOAC method [13].
dried in a forced-air oven at 105 °C for 24 h to determine the dry matter content, following the AOAC method [13].

Yield and Nutritional Composition of Sorghum Genotypes
The following traits were evaluated upon the harvest of the plants: plant height (m)obtained as the average of five competitive plants in the usable area, measured with a tape measure from the soil level to the top of the panicle; stem diameter (cm)-measured using a caliper rule on five competitive plants in the usable area at 10 cm above the soil; number of plants per hectare-determined by multiplying the number of plants in a linear meter of cultivated ground by the total number of linear meters per hectare, which is calculated by dividing 10,000.00by the spacing between rows (0.70 m); number of leaves per plantcounted directly, determining the average of five competitive plants in the usable area considering definitive and fully expanded leaves; average leaf length (cm)-determined using a tape measure, considering definitive and fully expanded leaves of five competitive plants in the usable area; average panicle length (cm)-determined using a tape measure on five competitive plants in the usable area.
Dry matter yield was estimated as the product of fresh matter yield and dry matter content, which was then converted to dry matter yield per hectare.Immediately after the harvest, the forages were chopped to particles of approximately 2 cm in a mill (TRP 40, Trapp©, São Paulo, Brazil), placed in sealed plastic bags, and frozen for later laboratory determinations of their chemical compositions.

Nutritional Composition and Fermentative Parameters of Silages
A portion of the chopped fresh forage was used to make silage.The forage was placed in PVC experimental mini-silos (10 cm width × 40 cm length) with the capacity for 2.5 kg of silage (600 kg/m 3 ).Then, 1.5 kg of sand dried in a forced-air oven at 55 °C for 72 h was placed at the bottom of each silo.The sand was separated from the forage with a plastic screen so that the retained effluent production could be calculated.
The material was compacted with a wooden dowel and the silos were closed with PVC lids equipped with a Bunsen valve.The lids were sealed with adhesive tape, and the 144 experimental silos were stored in a covered room during the experimental period.

Yield and Nutritional Composition of Sorghum Genotypes
The following traits were evaluated upon the harvest of the plants: plant height (m)obtained as the average of five competitive plants in the usable area, measured with a tape measure from the soil level to the top of the panicle; stem diameter (cm)-measured using a caliper rule on five competitive plants in the usable area at 10 cm above the soil; number of plants per hectare-determined by multiplying the number of plants in a linear meter of cultivated ground by the total number of linear meters per hectare, which is calculated by dividing 10,000.00by the spacing between rows (0.70 m); number of leaves per plantcounted directly, determining the average of five competitive plants in the usable area considering definitive and fully expanded leaves; average leaf length (cm)-determined using a tape measure, considering definitive and fully expanded leaves of five competitive plants in the usable area; average panicle length (cm)-determined using a tape measure on five competitive plants in the usable area.
Dry matter yield was estimated as the product of fresh matter yield and dry matter content, which was then converted to dry matter yield per hectare.Immediately after the harvest, the forages were chopped to particles of approximately 2 cm in a mill (TRP 40, Trapp©, São Paulo, Brazil), placed in sealed plastic bags, and frozen for later laboratory determinations of their chemical compositions.

Nutritional Composition and Fermentative Parameters of Silages
A portion of the chopped fresh forage was used to make silage.The forage was placed in PVC experimental mini-silos (10 cm width × 40 cm length) with the capacity for 2.5 kg of silage (600 kg/m 3 ).Then, 1.5 kg of sand dried in a forced-air oven at 55 • C for 72 h was placed at the bottom of each silo.The sand was separated from the forage with a plastic screen so that the retained effluent production could be calculated.
The material was compacted with a wooden dowel and the silos were closed with PVC lids equipped with a Bunsen valve.The lids were sealed with adhesive tape, and the 144 experimental silos were stored in a covered room during the experimental period.
For the evaluation of the genotypes after ensiling, a completely randomized design was adopted with 6 replicates and 24 treatments corresponding to each silage.
The mini-silos were opened 150 days after ensiling, on which occasion we aerated them for 30 min for the gases to be released; weighed them with and without the lid to measure the total gas losses; and measured their pH with a digital pH meter [14].Liquid effluent (LEL) and gas (GL) losses, as well as the recovery of dry matter (RDM), were then quantified [15,16].
Gas losses were calculated using the equation below: where GL = gas losses (%DM); WFSC = weight of full silo at the time of its closure (kg); and WFSO = weight of full silo at the time of its opening (kg).
The following equation was used to calculate the liquid effluent losses: where LEL = liquid effluent losses [% fresh matter (FM)]; WESSO = weight of empty silo with sand at the time of its opening (kg; after ensiling); and WESSC = weight of empty silo with sand at the time of its closure (kg; before ensiling).The dry matter recovery was calculated with the formula below: where DMR = dry matter recovery (%DM); FFME = forage fresh matter at the time of ensiling (kg); FDME = forage dry matter at the time of ensiling (%FM); SFMO = silage fresh matter at the time of silo opening (kg); and SDMO = silage dry matter at the time of silo opening (%FM).After this procedure, the silage was homogenized and two samples were collected; one was packed in a plastic bag and frozen to determine the ammoniacal nitrogen content (N-NH 3 ) [17].The second sample was pre-dried at 65 • C until reaching a constant weight to estimate the moisture content [13].Subsequently, the dried sample was ground through a Wiley mill with a 1 mm sieve and placed in polyethylene containers for further analyses.
The whole plants and silage were analyzed for their chemical compositions.Concentrations of DM, organic matter (OM), ether extract (EE), ash, crude protein (CP), and lignin were estimated as described by AOAC [13].Neutral (NDF) and acid (ADF) detergent fiber with corrections for the ash and CP contents were analyzed according to Mertens [18] and Licitra et al. [19].The resulting data were used to estimate the neutral (NDIP) and acid (ADIP) detergent insoluble protein levels.Total digestible nutrients (TDN) were estimated using the equation proposed by Cappelle et al. [20] through the following formula: TDN = 74.49− 0.5635 FDA.

Statistical Analyses
The data obtained were subjected to an analysis of variance, and when they were significant, the Scott-Knott test was performed with a probability of 5% (p < 0.05) for type-I errors.The statistical analyses were carried out using the computational resources of the "R" program (version 3.5.1).

Yield and Nutritional Composition of Sorghum Genotypes
The fresh matter yield differed (p < 0.001) between the genotypes, averaging 31.12 t/ha, with respective minimum and maximum values of 14.29 and 68.88 t/ha found in genotypes 4 and 5, respectively.There were differences (p < 0.001) between the treatments for the dry matter yield: genotype 5 had the highest value (22.24 t/ha), while genotype 4 was the least productive, generating 4.45 t/ha, which contributed to the generation of an overall mean of all genotypes of 9.52 t/ha (Table 3).Means followed by the same letter in the column do not differ significantly according to the Scott-Knott test at 5%.There were differences (p < 0.001) in the plant height variable, which had an overall mean of 2.15 m.The greatest plant height was found in genotype 4 (2.61 m), while the lowest value was observed in genotype 21 (1.24 m).Differences (p < 0.001) were also detected for the number of plants per hectare, with stands ranging from 119,047.62 (genotype 4) to 233,333.33 (genotype 24), averaging 183,333.35plants/ha.The stem diameter trait averaged 1.56 cm, ranging from 1.18 cm in genotype 14 to 1.43 cm in genotype 1, indicating significant differences (p < 0.001) (Table 3).
The number of leaves per plant differed (p < 0.001) across the hybrids, averaging 11.65.Genotype 11 had the highest value for this variable (14.67 leaves), while genotype 21 had the lowest number of leaves (8.00).The average leaf length differed (p < 0.001) across the cultivars, with respective maximum, mean, and minimum values of 77.27 (genotype 4), 60.66, and 50.03 cm (genotype 19).There were differences between the genotypes (p < 0.001) in the average panicle length, which averaged 26.38 cm and had maximum and minimum values of 35.43 and 15.47 cm, observed in genotypes 13 and 2, respectively (Table 3).
The total digestible nutrient content differed significantly (p < 0.001), with a minimum value of 669.8 g/kg DM, found in genotype 14, and a maximum value of 708.6 g/kg DM, observed in genotype 5, averaging 686.6 g/kg DM.The concentration of neutral detergent fiber corrected for ash and protein (NDFap) in the whole plant ranged (p < 0.001) from 525.0 g/kg DM in genotype 2 to 649.2 g/kg DM in genotype 18, averaging 594.4 g/kg DM.The non-fibrous carbohydrate content differed significantly (p < 0.001), with respective mean, minimum, and maximum values of 283.7, 222.8, and 363.7 g/kg DM.The acid detergent fiber content differed (p < 0.001), ranging from 280.6 (genotype 2) to 377.6 g/kg DM (genotype 8), averaging 329.1 g/kg on a DM basis.The hemicellulose contents differed as a function of the genotypes (p < 0.001) and ranged from a minimum of 260 g/kg DM in genotype 24 to a maximum of 339.2 g/kg DM in genotype 9. Cellulose also differed (p < 0.001), with respective maximum, minimum, and mean values of 326.4 g/kg DM (genotype 6), 251.6 g/kg DM (genotype 4), and 291.2 g/kg DM.The lignin content differed across the genotypes (p < 0.001) and ranged from a minimum of 36.9 g/kg DM (genotype 22) to a maximum of 51.7 g/kg DM (genotype 10) (Table 4).

Nutritional Composition and Fermentative Parameters of Silages
There was a significant effect (p < 0.001) for the dry matter (DM) contents of the silages of the 24 sorghum genotypes tested in this experiment.The DM content ranged from 281.9 g/kg FM in the silage from genotype 7 to 326.9 g/kg FM in that from genotype 3. The organic matter, ash, and total digestible nutrient contents exhibited significant differences (p < 0.001), with mean values of 958.9, 41.0, and 570.5 g/kg DM.There was a difference (p < 0.001) in the ether extract content in the silages evaluated in the present study, which ranged from 30.6 (genotype 14) to 49.6 g/kg DM (genotype 8).The crude protein and non-fibrous carbohydrates differed (p < 0.001), and the average contents of these nutritional components in the silages evaluated in this experiment were 61.0 and 275.3 g/kg DM, respectively (Table 5).Dry matter (DM; g/kg of fresh matter), organic matter (OM), crude protein (CP), ether extract (EE), total digestible nutrients (TDN), neutral detergent fiber (NDF), non-fibrous carbohydrates (NFC), acid detergent fiber (ADF), hemicellulose (HEM).Means followed by the same letter in the column do not differ significantly according to the Scott-Knott test at 5%.
There was a difference (p < 0.001) between the silages for neutral detergent fiber, which averaged 581.8 g/kg DM, with minimum and maximum values of 488.7 and 647.1 g/kg DM, respectively.Acid detergent fiber also differed (p < 0.001) between the tested silages, averaging 315.1 g/kg DM, with respective maximum and minimum values of 346.1 and 276.8 g/kg DM.Cellulose, hemicellulose, and lignin averaged 279.0, 265.6, and 37.0 g/kg DM, respectively, also varying (p < 0.001) across the studied silages (Table 5).
The fermentative and productive parameters of the silages are shown in Table 6.There was divergence (p = 0.003) between the pH of the sorghum silages despite the average pH being 3.94, with values ranging from 3.68 to 4.27.The ammoniacal nitrogen contents in the silages ranged from 4.08 to 13.76% of the total nitrogen, averaging 10%.Gas losses differed significantly (p < 0.001), with values that ranged from 53.0 to 119.6 g/kg DM.Liquid effluent losses were significantly different (p < 0.001) between the ensiled genotypes.Despite the significant effects on gas and effluent losses, no divergences (p = 0.280) were observed in the dry matter recovery of the silages of different sorghum genotypes, which averaged 853.9 g/kg.Table 6.Mean values for the potential of hydrogen (pH), ammoniacal nitrogen content (N-NH 3 , % total N), gas losses (GL, g/kg DM), liquid effluent losses (LEL, kg/ton), and dry matter recovery (DMR, g/kg DM) of the silages of 24 forage sorghum genotypes.

Yield and Nutritional Composition of 24 Sorghum Genotypes
The present results obtained for the fresh matter yield (31.12 t/ha) are similar to some reported in the literature.Perazzo et al. [12] found the fresh matter yield to be over 30 t/ha for groups of higher-yielding forage cultivars.Moraes et al. [21], however, in an experiment with four sorghum hybrids-one forage type and two dual-purpose types-found an average fresh matter yield of 31.67 t/ha, with values ranging from 29.04 to 35.66 t/ha.In the present study, genotypes 5 and 1 stood out for their high yields, which were greater than 60 t/ha.It is important to note that FMY is the main factor for most livestock producers around the world when choosing a cultivar [11].
Monteiro et al. [22] analyzed 51 sorghum cultivars and found the DM yield to vary from 0.53 to 12.92 t/ha.Silva et al. [23] evaluated 25 sorghum cultivars in a semi-arid region and observed values ranging from 7679.87 to 20,948.70,averaging 13,799.30t/ha.
Yield-related traits are affected by limiting factors such as water deficit, low soil fertility, soil degradation, inadequate plant population, unsuitable cultivar, weeds, pests, and diseases [24].However, the current study proves that, under similar environmental conditions, genotype is a crucial factor to consider when determining the ideal option among cultivars.
Perazzo et al. [12] reported a strong positive correlation between yield and plant height; however, for yield, the DM content of the plant must also be considered.Silva et al. [23] found an average plant height of 2.01 m in an evaluation of 25 sorghum hybrids in a Brazilian semi-arid region, which is close the value obtained in the current study (2.15 m).This identifies the foraging potential of the genotypes used.
Avelino et al. [25] suggested that the ideal population for sorghum plants depends on the genetics, soil fertility, water availability, and time of sowing.Once these factors are satisfactory, the yield can increase with a larger plant production.In the present study, genotype 5 showed a higher FMY, a higher DMY, and also a higher NPH, reinforcing the above-mentioned authors' statement.As the climatic conditions and planting density were the same, it can be stated that the plant's genotype was the factor that influenced the viability observed in germination, thus increasing the NPH and, consequently, productivity.In this sense, considering that the sowing density was the same, we can affirm that genotype 5 has the highest germination value.Germination denotes a biochemical process in which various enzymes (such as proteases, amylases, β-glucanases, and phytases) are activated, leading to the breakdown of starch to meet the energy requirements for the biochemical alterations necessary for germination [26].
Perazzo et al. [12] found a variation of 2.24 to 1.64 cm in stem diameter, with an average of 1.92 cm.These authors explained that agronomic traits such as plant height, stem diameter, and the number of plants per hectare are variables that can directly affect biomass production.The number of plants is linked to the tiller population as a function of the area, and the stem diameter and plant height are related to volume and/or tiller weight values [27].Associations among these agronomic traits have important effects such as on the density of stands and on the tiller size, influencing the biomass production per area.
The average number of leaves per plant in the present study was higher than the 7.28 leaves per plant found by Perazzo et al. [12].However, a comparison of the average leaf length and average panicle length revealed that they are near those observed by the aforementioned authors (57.94 and 26 cm, respectively).The forage quality can be highly undervalued when the participation of leaves and panicles is low, since these are the most nutritious part of forage plants.These morphometric values contribute to the explanation of the relative participation of the plant components and its nutritional quality.
The current research reveals that genotype 5 exhibits enhanced traits in dry matter yield while showing a decrease in certain nutritional components like EE, ash, NDF, NFC, and lignin.Nevertheless, it maintains a higher protein content.Therefore, we can affirm that this genotype, with its superior growth characteristics, produces a greater quantity of agronomic parts of high nutritional value (mainly leaves), which indicates its importance and necessary management characteristics in silage production.The DM content of the studied sorghum genotypes is above 250 g/kg, which is considered adequate according to Mc Donald et al. [28].These authors also explained that the plant DM content is important in the ensilage process, since it is one of the factors determining the type of fermentation along with the soluble carbohydrate content.Some genotypes that obtained a higher value for CP also displayed an elevated panicle length, as was the case for genotype 5, evidencing the great importance of this component for plant quality.The crude protein content in the dry matter of eight genotypes (1, 5, 6, 7, 17, 21, 22, and 24) of the materials analyzed in this experiment was above 70 g/kg, which is the minimum required to ensure adequate ruminal fermentation [29].For this reason, these genotypes can be used as the sole roughage source in animal feeding if there are no restrictions regarding the other nutritional components of the forage.
Genotypes 2, 4, 5, 15, and 21 had higher values for TDN, which exceeded 700 g/kg.Genotype 4, together with genotypes 2 and 8, showed superior values for NFC.Genotype 4 also had 41.2 g/kg of lignin, one of the highest values amidst the genotypes for this chemical component.This response is a consequence of the greater height of genotype 4, resulting from its longer stem.The presence of this longer stem, according to Zanine et al. [30], possibly explains the high NFC value, since soluble carbohydrates are present in this plant component, making it an important substrate for adequate lactic fermentation in ensilage.However, because genotype 4 was taller, higher lignification was observed.
Neutral detergent fiber levels should be between 500 and 600 g/kg, because higher values may compromise intake due to a greater participation of fibrous carbohydrates, which move slowly through the digestive tract of ruminants, causing the sensation of a full rumen and limiting their intake rate.Higher values of this fraction may be correlated with a longer permanence of plants in the field and their age at the time of harvest [24].In the present study, the ADF levels were within the limits recommended in the literature for a feed not to compromise the attack of microorganisms and enzymes on the fiber as a result of the presence of lignin, which works as a barrier.Forages with an ADF content around 300 g/kg DM or less are consumed at high levels, unlike those with ADF values greater than 400 g/kg [29].
The NDIP and ADIP fractions are composed of nitrogen forms associated with lignin, tannin-protein complexes, and components resulting from the Maillard reaction.The components of these fractions are highly resistant to microbial and enzymatic attacks and are thus not completely insoluble and/or indigestible in the gastrointestinal tract.The higher ADIP levels can be explained by the plant height [24], as is the case of genotype 4.

Nutritional Composition and Fermentative Parameters of Silages
The forage dry matter content at the time of ensiling is one of the main factors determining the fermentation process [11].As can be seen in our results, the average DM content of the plants remained at 300 g/kg FM in the silage, indicating no alterations that might signify effluent losses due to high moisture of the ensiled mass.Furthermore, according to Ali et al. [31], high-quality silages originate from plants with TDN values greater than 650 g/kg, which were reached by some of the evaluated genotypes.The types of sorghum evaluated in this study exhibit a forage behavior, i.e., these plants have a lower percentage of panicles, which results in lower concentrations of crude protein in the silage.This was confirmed by Pedreira et al. [32], who found 88 g/kg CP in grain sorghum (BR-700) and 65 g/kg CP in forage sorghum (hybrid 498111).Silva et al. [33] also found divergences for NDF and ADF in silages of sorghum hybrids, which averaged 57.0 and 34.0 g/kg, respectively, which are similar to the results found in the current experiment.
There are no alterations in the lignin or cellulose contents during the ensilage process, suggesting that the fermenting microorganisms in the silo do not degrade these fractions [34], and losses of these portions are associated with situations in which there is deterioration caused by fungi [29].However, there is evidence that attributes hemicellulose degradation in sorghum silages to the action of hemicellulases in the early stages of preservation and acidity in the medium [35].In the current study, there was a numerical decrease in the cellulase and hemicellulase contents of the silage in relation to the chemical composition of the plant.The degradation or biochemical transformations of nutritional components in the silage are mainly the result of the fermentation process; as a result, the main and most important factor affected is the "pH".Nascimento et al. [36] explained that 3.8 to 4.2 is the pH range in which silage shows a good quality.However, in addition to the ideal range, the rapid pH decline favors an inhibition of the activity of undesirable microorganisms from the start of fermentation.
Kung Jr. et al. [11] described that silages with ammoniacal nitrogen contents below 10% of the total N are considered high-quality silages.Ammoniacal nitrogen is considered one of the main variables determining the fermentation quality because it is an indicator of proteolysis during the fermentation process, performed mainly by bacteria of the genus Clostridium.This proteolysis is extended during fermentation when the acidic conditions are not sufficient for the undesirable microorganisms to be inhibited [11].
According to McDonald et al. [28], a significant increase in gas losses occurs when there is alcohol production (ethanol or mannitol) from secondary fermentations.The studied sorghum varieties had a high proportion of succulent stems with high soluble carbohydrate contents, and the excess of this component in some genotypes might have led to higher gas losses.
Oliveira et al. [37] evaluated several crops and observed liquid effluent losses of 96.9 and 69.7 kg/ton in sorghum silage.These values are much higher than the values of 17.02 to 38.38 kg/ton found in the present study.Such a low level of effluent losses is due to the dry matter content of the plant, as explained previously.There was no alteration for this variable in the chemical composition of the silage, meaning there was no leaching of the silo moisture.This indicates that the dry matter losses of the silages stemmed mainly from the gas losses due to the typical alcoholic fermentation occurring in sorghum silages.
Pinho et al. [10] evaluated silages of different sorghum cultivars and observed DMR values ranging from 757 to 904 g/kg.Because of this discrepancy, the authors stressed the importance of taking into account not only the DM yield of the sorghum crop but also the recovery of silage dry matter when recommending sorghum genotypes.In the present study, because there were no divergences between the DMRs, genotypes can be recommended based on their yield and nutritional values.

Conclusions
Of all hybrids evaluated in this study for ensilage, genotype 5 is the most recommended under the conditions set in this study because it demonstrated the highest yield and nutritional value in both its fresh form and its ensiled form.Additionally, its fermentation losses were lower and its dry matter recovery after ensiling was 880 g/kg.It is crucial to highlight that several other sorghum genotypes also demonstrated noteworthy values in terms of their forage and silage production capacities.Therefore, further studies under diverse climatic and management conditions are imperative to identify the optimal genotypes suited for specific climates or regions.
51.7 a 235.5 d 540.3 c 606.3 b 312.2 b 290.8 b 281.2 a 34.4 c 307.1 a 956.2 b 71.51.1 a 237.9 d 518.0 c 588.7 b 276.8 c 298.4 b 257.9 b 32.4 c 311.4 a 961.3 a 43.4 d 39.7 b 38.7 c 260.6 c 630.1 a 617.6 a 304.3 b 298.0 b 284.0 a 35.6 c 294.8 b 954.4 b 64.4 b 40.1 b 45.6 b 282.4 c 627.5 a 567.5 c 309.4 b 242.9 d 289.8 a 34.8 c

Table 1 .
Chemical properties of the soil in the experimental area.

Table 2 .
List of the 24 sorghum genotypes evaluated.

Table 4 .
Mean values of nutritional components (g/kg DM) in whole-plant samples of 24 forage sorghum genotypes.

Table 5 .
Mean values of nutritional components (g/kg DM) found in silages of 24 forage sorghum genotypes.