Chemical and microbial additives improve stability and fermentation of corn grain silage rehydrated with wet orange pulp

ABSTRACT:

The combined use of wet pulp orange (WPO) with rehydrated ground corn grain (GCG) in silages presents as a feasible approach; however, the low aerobic stability of WPO could compromise silage quality. This study aimed to assess the effects of two ensiling durations and the application of either chemical or microbial additives on the chemical composition, in vitro digestibility, fermentation losses, microbial and fermentative profiles, and aerobic stability of GCG silages rehydrated with WPO. A mixture comprised of 37.5 % WPO and 62.5 % GCG was prepared to reach 600 g kg⁻¹ dry matter (DM) and was either subjected to a 24-h delay in sealing or sealed immediately. The mixture was treated with sodium benzoate (SB, 0.2 % DM), formic acid (FA, 4.5 L t⁻¹), or a microbial inoculant (referred to as LB) containing Enterococcus faecium, Lactiplantibacillus plantarum, and Lentilactobacillus buchneri at two concentrations (LB1: 1.5 × 10⁵ colony-forming units (CFU) g⁻¹ and LB2: 3.0 × 10⁵ CFU g⁻¹), or with water (control). Each treatment was ensiled in four buckets, each with an 8 kg capacity. Additive-treated silages exhibited 44.5 % lower yeast counts and significantly higher aerobic stability (196 h vs. 66 h) compared to the control, with no significant differences among the various additives. This reduction in yeast counts is likely a key factor contributing to the enhanced aerobic stability. Additionally, the microbial inoculants boosted acetic acid concentration by 2.03 and 2.68 times with and without delay sealing, respectively, underscoring their antifungal properties. While a 24-h delay in sealing slightly reduced silage quality, it did not adversely affect aerobic stability. All tested additives enhanced aerobic stability, enabling selection based on cost, availability, and ease of application.

Keywords:
Enterococcus faecium; Lactiplantibacillus plantarum; Lentilactobacillus buchneri; formic acid; sodium benzoate

Introduction

Wet pulp orange (WPO) is an industrial co-product typically processed for pelleting to produce citrus pulp in large orange juice enterprises. However, smaller businesses may dispose of this material, which can lead to environmental concerns (Zema et al., 2018). One practical use for WPO is in animal nutrition. Nevertheless, due to its low dry matter and high water-soluble carbohydrate (WSC) concentrations, this material tends to degrade rapidly when exposed to air (Bampidis and Robinson, 2006). To facilitate large-scale utilization on commercial farms, it is essential to develop effective storage methods for WPO. One potential approach is to ensile WPO in combination with higher dry matter content materials, such as ground corn grain (GCG).

The incorporation of up to 42 % WPO into corn silage resulted in the production of high-quality silage with a favorable fermentative profile. However, it also led to a reduction in aerobic stability by 141 h (213 h vs. 71.6 h) compared to corn silage rehydrated with water (Menezes et al., 2022a). In this context, the use of chemical or microbial additives can enhance silage preservation after silo opening (Muck et al., 2018; Menezes et al., 2022b). Acetobacter sp. bacteria have been shown to induce aerobic spoilage, irrespective of yeast counts (Ávila and Carvalho, 2020; Menezes et al., 2022a), and the application of formic acid can inhibit the growth of these undesirable bacteria. Additionally, other spoilage microorganisms, such as some yeast strains, unaffected by formic acid, may contaminate the WPO, thereby diminishing its aerobic stability. This underscores the need to evaluate the effectiveness of sodium benzoate and Lentilactobacillus buchneri (Muck et al., 2018).

Another important yet unexplored aspect of ensiling WPO silage with GCG is its response to delayed sealing, which may occur when WPO needs to be transported or cannot be ensiled on the same day it is produced. While transporting WPO can be feasible despite its high moisture content and associated costs, it is often acquired at a relatively low cost. Due to operational constraints, farms may not have the capacity to ensile WPO immediately after it is production. Therefore, evaluating the effects of delayed sealing is essential for understanding its impact on fermentation and for informing practical ensiling strategies that utilize GCG. Consequently, this study aimed to assess the effects of two different ensiling times, along with the incorporation of chemical or microbial additives, on various factors, including the chemical composition, in vitro digestibility, fermentation losses, microbial and fermentative profiles, as well as the aerobic stability of GCG silages rehydrated with WPO.

Materials and Methods

Corn and wet pulp orange purchase and silage production

The experiment was conducted in Belo Horizonte, Minas Gerais State, Brazil (19°54’58" S, 43°56’7" W, at an altitude of 852 m). Corn grain was sourced from 40 commercial farms, representing a range of real variations in corn origin and allowing for real repetitions across multiple locations of Minas Gerais. The corn was finely ground in a hammer mill to achieve an average particle size of 2 mm. Each of the 40 corn samples was ground and ensiled separately, resulting in 40 true replicates. Consequently, corn from each farm was kept distinct, with no mixing occurring between samples from different farms. The WPO utilized was produced by an orange juice enterprise located in Contagem, Minas Gerais State, Brazil (19°50’43.79" S, 44°6’4.78" W, at an altitude of 890 m). A separate sample, distinct from the one used for ensiling, was obtained from the enterprise four days prior to ensiling. The dry matter concentration of this sample was analyzed to facilitate the formulation of the mixtures before ensiling. This step was taken due to the high residual sugar levels in the WPO, which can encourage fungal growth.

Prior to ensiling, samples of corn and WPO were collected to evaluate their chemical composition (Table 1). The WPO was produced on 08 Apr 2022, with half utilized on the same day for silage production (immediate sealing), while the other half was stored in a covered shed and left exposed to air. This portion was then used for silage production the following day (with a 24-hour delay before sealing). On each ensiling day, the WPO was mixed with 20 different corn samples, to achieve a fresh matter (FM) ratio of 600 g kg⁻¹ (37.5 % WPO and 62.5 % GCG). On the ensiling day, after mixing the WPO and GCG from the 20 samples separately, the mixture was randomly divided into five groups (each containing four different samples). Each group was treated with the following additives mixed with distilled water: sodium benzoate (SB; 0.2 % DM); formic acid (FA; 4.5 L t⁻¹); a microbial additive at 1.5 × 10⁵ CFU g⁻¹ (LB1; 4.5 × 10⁴ CFU g⁻¹ of Enterococcus faecium (Orla-Jensen 1919) M74 + 3.0 × 10⁴ CFU g⁻¹ of Lactiplantibacillus plantarum (Orla-Jensen 1919) CH6072 + 7.5 × 10⁴ CFU g⁻¹ of Lentilactobacillus. buchneri (Henneberg 1903) LB1819; dosage recommended by the manufacturer); a higher dosage of the same microbial additive at 3.0 × 10⁵ CFU g⁻¹ (LB2; 9.0 × 10⁴ CFU g⁻¹ of E. faecium M74 + 6.0 × 10⁴ CFU g⁻¹ of L. plantarum CH6072 + 1.5 × 10⁵ CFU g⁻¹ of L. buchneri LB1819 (double the dosage recommended by the manufacturer) or simply with water (control). The increased dosage was employed to assess whether a higher additive concentration is necessary for this type of silage, which presents a greater fermentation challenge due to elevated water-soluble carbohydrate (WSC) concentrations and microbial contamination.

Table 1
Chemical composition of ground corn grain and wet pulp orange used in ensiling.

Variables (g kg−1 DM, unless noted)	Corn ground grain*	Wet pulp orange
Dry matter (g kg−1 FM)	865 (11.2)	149
Crude protein	90.3 (3.98)	105
Neutral detergent fiber	147 (7.80)	170
Acid detergent fiber	36.3 (1.40)	123
Acid detergent lignin	6.02 (0.28)	23.8
Ether extract	38.5 (1.92)	27.5
Ash	30.1 (1.12)	24.6
Nonfibrous carbohydrates	694 (6.66)	673

• ^*
  Values within parentheses represent the standard deviation among the 40 corn samples. DM = dry matter; FM = fresh matter.

The additives were mixed in 200 mL of distilled water and homogeneously distributed over the material. After blending and ensuring a uniform treatment mixture, 8 kg of the material was measured and ensiled in experimental silos (buckets) to achieve an average density of 1000 kg FM m⁻³. A total of four experimental silos were constructed for each treatment, resulting in a total of 40 silos stored in a covered shed. At 126 days post-ensiling, the silos were opened, their contents were thoroughly mixed, and samples were collected for further analysis.

Determination of chemical composition

Following silo opening, the samples were immediately dried in a forced ventilation oven at 55 °C for 72 h and subsequently ground to a particle size of 1 mm using a knife mill, specifically the Thomas Wiley model 4 from Thomas Scientific. The concentrations of acid detergent insoluble lignin (ADL), neutral detergent insoluble fiber (NDF), and acid detergent insoluble fiber (ADF) were determined using the sequential method outlined by Van Soest et al. (1991). NDF concentration was established by adding 2 mL of heat-stable amylase, while ADL was quantified through cellulose solubilization using sulfuric acid. DM concentration was assessed in an oven set at 105 °C (AOAC, 1990; method ID 934.01), the crude protein (CP) content was evaluated via the Kjeldahl method (AOAC, 1990; method ID 954.01), and ether extract (EE) was determined using the Soxhlet method (AOAC, 1990; method ID 963.15). Ash concentration was measured in a muffle furnace at 600 °C for 4 h (AOAC, 1990; method ID 942.05). The concentrations of acid detergent insoluble protein (ADIP) and neutral detergent insoluble protein (NDIP) were calculated by measuring the nitrogen residue in the ADF and NDF samples and multiplying by 6.25. The concentration of non-fibrous carbohydrates (NFC) was calculated according to the NRC (2001), as shown in Eq. (1):

(1) $$\text{NFC}=100-(\text{\%NDF}+\text{\%CP}+\text{\%EE}+\text{ash})$$

Determination of in vitro dry matter digestibility

In vitro dry matter digestibility (IVDMD) was assessed using the method outlined by Tilley and Terry (1963), as adapted by Holden (1999). For each sample, two replicates of 0.25 g each were prepared in F57 sacs. These sacs were then incubated in a Dayse II Ankom incubator for a total of 72 hours: 48 h with buffer solutions A and B and bovine rumen juice, followed by an additional 24 h after the addition of hydrochloric acid and pepsin. All procedures were conducted in accordance with the guidelines established by the Ethics in Animal Use Committee (CEUA) of the Pontifícia Universidade Católica de Minas Gerais, as specified in protocol 07/2022. The ruminal juice was obtained from a fistulated steer (crossbred - Holstein × Gyr; body weight = 600 kg) that was raised on a two-hectare pasture. The steer's diet consisted of 1.5 kg of concentrate (18 % CP) and whole-plant corn silages per day, supplemented with mineral salt and water fed ad libitum.

Determination of fermentative variables

The concentrations of aldehydes, ethanol, esters, and organic acids in silages were assessed using an aqueous extract, followed by filtration and centrifugation for 15 min at 1.05 × 10³ rad s⁻¹. Analysis was conducted using a GC-MS QP 2010 plus gas chromatograph with mass detector (Shimadzu®), equipped with a Stabilwax capillary column from Restek® measuring 60 m length, with a diameter of 0.25 mm, and a 0.25 µm crossbond carbowax polyethylene glycol. The concentration of lactic acid was determined using the colorimetric method developed by Pryce (1969), with measurements obtained on a spectrophotometer (Maroconi Janway 6305) at a wavelength (λ) of 565 nm.

Determination of losses

The experimental silos were equipped with Bunsen valves and a specially designed lid to facilitate the release of gas. At the bottom of each bucket, a cotton bag containing approximately 2.0 kg of dried sand was placed to allow for effluent measurement, following the methods outlined by Pedroso et al. (2008). Prior to ensiling, the empty experimental silos – along with their lids and dry sandbags – were weighed. After filling the silos with mixed corn grain and compacting the contents, they were sealed with adhesive tape and weighed again. When the silos were opened, they were weighed while still fully loaded to determine gas production, as calculated according to Eq. (2):

(2) $$\begin{array}{cc}GL=[(WFBen-PSW)\times DMen]-[(WFBab-SW)& \\ \times DMab]\times 100\}/[(WFBen-PSW)\times DMen]& \end{array}$$

where GL represents gas losses (%DM); WFBen denotes the weight of the full bucket at ensiling (kg); PSW signifies the predetermined weight (empty bucket + lid + dry sand + bag) at ensiling (kg); DMen indicates the dry matter (DM) concentration of the rehydrated corn at ensiling (%); WFBab represents the weight of the full bucket at silo opening (kg); and DMab denotes the DM concentration of the corn at silo opening (%).

The silos were opened, and the silage was extracted, homogenized, and sampled before being weighed as a set. This procedure allowed for the quantification of effluent production using Eq. (3):

(3) $$EL=[(WEf\times 1000)/EMCR]$$

where EL represents effluent losses (g kg⁻¹ FM); WEf denotes the weight of the effluent (empty set weight after silo opening minus empty set weight before filling); and EMCR indicates the mass of corn grain ensiled (kg).

Total loss of dry matter (TLDM) was determined by calculating the difference between the initial and final gross dry mass weight of the experimental silos, relative to the amount of dry mass ensiled. This calculation considered the designated weight at both the time of ensiling and silo opening. Consequently, this procedure enabled the quantification of TLDM using Eq. (4):

(4) $$\begin{array}{cc}& \text{TLDM}=\{[(WFBen-PSW)\times DMen]-[(WFBab-\\ & SEab)\times DMab]\times 100\}/[(WFBen-PSW)\times DMen]\end{array}$$

where TLDM represents total loss of dry matter (%DM); SEab indicates the set weight (bucket + lid + wet sand + bag) at silo opening (kg).

Determination of aerobic stability

For each sample, two plastic buckets were utilized: one with 2 kg of silage and the other with 3 kg. The buckets, which were covered with aluminum foil to permit airflow, were placed in a controlled environment maintained at a specific temperature of 25.0 ± 1 °C. The temperature of the silage in the 2 kg bucket was monitored every 2 h for a period of ten days. Monitoring involved inserting a mercury thermometer 10 cm into the center of the silage mass. The determination of silage aerobic stability was performed by observing the time required for the silage temperature to rise by 2 °C above the ambient room temperature, following the method described by Ranjit and Kung Jr (2000).

The silage in the 3 kg bucket was thoroughly mixed, and 100 g samples were collected at 0, 24, 48, 96, and 240 h to assess pH and the concentration of NH₃ in relation to total nitrogen (NH₃-N/TN). Immediately after collection, the samples were frozen and stored in a cold chamber for at least seven days to ensure consistency across all samples, as outlined in the protocol by Ferraretto et al. (2018).

A 20 g sample, left undried and unground, was diluted tenfold by mass in double-distilled water for the determination of pH and NH₃-N/TN, in accordance with the methods described by Ferraretto et al. (2016). The mixture was stirred for 30 s using a high-speed mixer, and then filtered through four layers of gauze. The pH measurements were taken with a digital potentiometer (HI 221; Hanna Instruments). The determination of NH3-N/TN was carried out via Kjeldahl distillation, excluding the digestion step, as specified in the AOAC method (AOAC, 2012; method ID 984.13).

Determination of microbiological count

To enumerate aerobic microorganisms, including bacteria, molds, and yeasts, the spread plate method was employed using a dilution of 10⁻¹. Initially, a 25 g sample was weighed and transferred to flasks containing 225 mL of a 0.1 % peptone solution. These flasks were then sterilized in an autoclave at 121 °C for 15 min and manually stirred for 2 min. After a 1-min resting period, decimal dilutions (ranging from 10⁻² to 10⁻⁵) were prepared from the resulting extract in test vessels containing 9 mL of sterile 0.1 % peptone solution. Subsequently, 0.1 mL from each dilution was spread onto disposable Petri dishes containing agar using a Drigalski spatula.

The total yeast count was conducted on agar plates enriched with Tryptone Glucose Yeast (TGY) extract after incubation under aerobic conditions for four days at 30 ± 1 °C. The total mold count was assessed on plates containing Dichloran Rose-Bengal Chloramphenicol (DRBC) after a 5-day incubation period at 25 ± 1 °C. The total bacterial count was determined on Plate Count Agar (PCA) after 48 h of aerobic incubation at 35 °C, as described by Pitt and Hocking (2009). Microbial counts were performed both at silo opening and after aerobic stability assessment. The resulting data were reported as colony-forming units per gram of silage (CFU g⁻¹).

Statistical analyses

The normality of data and homoscedasticity of variances were assessed using the Shapiro Wilk and Bartlett tests, respectively. The experimental design employed was a completely randomized in a factorial arrangement with two factors: sealing time (immediate or delayed) × additives (control, SB, FA, LB1, or LB2). The data were analyzed using a two-way analysis of variance (ANOVA) according to the following model:

(5) $$Y_{ijk}=\text{μ}+S_i+A_j+{(S\times A)}_{ij}+{\text{ε}}_{ijk}$$

where Y_ijk = observed value of the response variable of repetition k, at sealing time i and additive j; μ = overall mean; S_i = effect of sealing time (immediate or delayed); A_j = effect of additive (control, SB, FA, LB1, or LB2); (S × A)_ij = interaction effect between sealing time and additive; ε_ijk = random error associated with each observation, assumed to be normally distributed with mean μ and variance σ².

Means were compared using Tukey's test (p < 0.05) when there was interaction between delayed sealing and additive, or when there was only a main effect of additive. In cases where only the main effect of delayed sealing was present, Fisher's test (p < 0.05) was applied. Statistical analyses were performed using R software, version 4.2.3.

Results

The interaction between sealing time and additives showed no significant impact on the concentrations of DM, CP, EE, NDIP, ADIP, and NFC, as well as IVDMD (p > 0.05) (Table 2). Silages treated with sodium benzoate and formic acid had higher DM concentrations. In contrast, those treated with LB1 and the control showed intermediate DM concentrations, whereas silages treated with LB2 had lower DM concentrations (p < 0.001). Additionally, the NDIP concentration was 8.9 % higher in silages with delayed sealing compared to those with immediate sealing (p = 0.013). IVDMD was 5.86 % higher in silages treated with formic acid and sodium benzoate than in those treated with LB1 and control (813 vs. 768 g kg⁻¹ DM) (p = 0.004). No significant effect was observed on the fibrous fraction (p > 0.05) (Table 3).

Table 2
Chemical composition parameters and in vitro dry matter digestibility of corn grain silage rehydrated with wet pulp orange treated with different additives and subjected or not to a 24-hour delay in sealing.

Variables		Additive					Mean	SEM	p-value
		Control	SB	FA	LB1	LB2			A	D	A × D
Dry matter (g kg−1 NM)
	With delay	584	585	595	577	560	580	3.14	< 0.001	0.545	0.209
	Without delay	582	616	599	581	569	589
	Mean	583 ABC	600 A	597 AB	579 BC	564 C	-
Crude protein (g kg−1 DM)
	With delay	8.74	9.14	9.12	9.00	8.46	8.89	0.064	0.661	0.492	0.060
	Without delay	8.98	8.80	8.73	8.55	8.98	8.81
	Mean	8.86	8.97	8.93	8.78	8.72	-
Ether extract (g kg−1 DM)
	With delay	32.8	31.9	34.0	34.3	32.1	33.1	0.608	0.956	0.228	0.953
	Without delay	34.1	34.6	34.3	35.1	35.2	34.7
	Mean	33.5	33.3	34.2	34.7	33.7	-
Neutral detergent insoluble protein (g kg−1 DM)
	With delay	1.75	1.47	1.58	1.43	1.66	1.58 a	0.028	0.085	0.013	0.352
	Without delay	1.54	1.46	1.46	1.41	1.38	1.45 b
	Mean	1.65	1.47	1.52	1.42	1.52	-
Acid detergent insoluble protein (g kg−1 DM)
	With delay	0.310	0.240	0.240	0.280	0.270	0.268	0.013	0.864	0.235	0.835
	Without delay	0.320	0.310	0.310	0.280	0.280	0.300
	Mean	0.315	0.275	0.275	0.280	0.275	-
Nonfibrous carbohydrates (g kg−1 DM)
	With delay	707	729	723	729	734	724	2.98	0.140	0.705	0.380
	Without delay	717	735	726	704	729	722
	Mean	712	732	725	717	732	-
In vitro dry matter digestibility (g kg−1 DM)
	With delay	785	819	824	789	781	800	5.62	0.004	0.058	0.255
	Without delay	751	809	798	748	802	782
	Mean	768 B	814 A	811 A	768 B	792 AB	-

• Means followed by distinct uppercase letters within a row (comparing additive) and lowercase in the column (comparing delayed sealing) differ by Tukey's test (p < 0.05). SB = sodium benzoate (0.2 %); FA = Formic acid (4.5 L t⁻¹); LB1 = microbial additive in dosage recommended by the manufacturer; LB2 = microbial additive in double of the dosage recommended by the manufacturer; SEM = standard error of mean; A = additive effect; D = delay sealing effect; A × D = interaction between additive and delay sealing effect; NM = natural matter; DM = Dry matter.

Table 3
Fibrous fractions of corn grain silage rehydrated with wet pulp orange treated with different additives and subjected or not to a 24-hour delay in sealing.

Variables		Additive					Mean	SEM	p-value
		Control	SB	FA	LB1	LB2			A	D	A × D
Neutral detergent fiber (g kg−1 DM)
	With delay	139	112	116	117	115	120	2.59	0.060	0.865	0.380
	Without delay	126	111	119	135	112	121
	Mean	133	112	118	126	114	-
Acid detergent fiber (g kg−1 DM)
	With delay	25.4	25.5	23.1	30.0	27.2	26.2	2.59	0.119	0.511	0.732
	Without delay	27.1	24.2	25.2	38.1	25.1	27.9
	Mean	26.3	24.9	24.2	34.1	26.2	-
Acid detergent lignin (g kg−1 DM)
	With delay	35.3	32.9	32.8	36.1	35.4	34.5	0.431	0.249	0.860	0.195
	Without delay	32.5	35.5	34.8	37.0	33.5	34.7
	Mean	33.9	34.2	33.8	36.6	34.5	-
Ash (g kg−1 DM)
	With delay	33.8	36.0	35.1	30.4	34.6	34.0	0.899	0.984	0.751	0.170
	Without delay	33.6	31.4	33.1	40.0	34.5	34.5
	Mean	33.7	33.7	34.1	35.2	34.6	-

• SB = sodium benzoate (0.2 %); FA = Formic acid (4.5 L t⁻¹); LB1 = microbial additive in dosage recommended by the manufacturer; LB2 = microbial additive in double of the dosage recommended by the manufacturer; SEM = standard error of mean; DM = Dry matter; A = additive effect; D = delay sealing effect; A × D = interaction between additive and delay sealing effect.

The interaction between sealing time and additives had a significant impact on total DM loss (p = 0.017) (Table 4). In silages subjected to delayed sealing, no significant differences in total DM loss were observed among the various additives. However, with immediate sealing, sodium benzoate produced the lowest total DM loss at 2.61 %, significantly differing from the other additives. Only the silages treated with sodium benzoate exhibited a higher total DM loss under delayed sealing (7.23 %) compared to immediate sealing (2.61 %). No significant effects were observed for gas or effluent losses (p > 0.05).

Table 4
Losses of corn grain silage rehydrated with wet pulp orange treated with different additives and subjected or not to a 24-hour delay in sealing.

Variables		Additive					Mean	SEM	p-value
		Control	SB	FA	LB1	LB2			A	D	A × D
Gas loss (% DM)
	With Delay	1.65	1.69	1.80	1.66	1.79	1.72	0.073	0.870	0.719	0.626
	Without Delay	2.10	2.13	1.64	1.62	1.63	1.82
	Mean	1.88	1.91	1.72	1.64	1.71	-
Effluent loss (g kg−1 FM)
	With Delay	0.982	0.786	0.750	0.768	1.00	0.857	0.047	0.179	0.285	0.481
	Without Delay	0.839	0.804	0.821	1.13	1.21	0.961
	Mean	0.911	0.795	0.786	0.949	1.11	-
Total DM loss (% DM)
	With Delay	4.20 Aa	7.23 Aa	4.05 Aa	5.26 Aa	8.38 Aa	5.82	0.501	0.008	0.235	0.017
	Without Delay	5.19 ABCa	2.61 Cb	4.93 BCa	8.23 ABa	10.2 Aa	6.23
	Mean	4.70	4.92	4.49	6.75	9.29	-

• Means followed by distinct letters, uppercase in the line (comparing additive) and lowercase in the column (comparing delayed sealing), differ by Tukey's test (p < 0.05). SB = sodium benzoate (0.2 %); FA = Formic acid (4.5 L t⁻¹); LB1 = microbial additive in dosage recommended by the manufacturer; LB2 = microbial additive in double of the dosage recommended by the manufacturer; SEM = standard error of mean; DM = Dry matter; FM = fresh matter; A = additive effect; D = delayed sealing effect; A × D = interaction between additive and delayed sealing effect.

The interaction between sealing time and additives had no significant impact on aerobic stability or on counts of aerobic bacteria, molds, and yeasts at silo opening and during the assessment of aerobic stability (p > 0.05) (Table 5). During silo opening, the counts of aerobic bacteria (p = 0.009), molds (p = 0.001), and yeasts (p < 0.001) were 13.3 %, 64.8 %, and 20 % higher, respectively, in silages subjected to delayed sealing compared to those sealed immediately. Additionally, the yeast count was, on average, 44.5 % lower (p = 0.006; 2.21 vs. 3.98 log₁₀ CFU g⁻¹) in silages treated with additives compared to untreated silages, with no observed differences among the additives. At the time of assessing aerobic stability loss, counts of aerobic bacteria and molds were 16.7 and 54.8 % lower, respectively (p < 0.001), in silages sealed immediately compared to those subjected to delayed sealing. The counts of aerobic bacteria and molds were also, on average, 20.9 % (5.02 vs. 3.97 log₁₀ CFU g⁻¹) and 50.5 % (2.91 vs. 1.44 log₁₀ CFU g⁻¹) lower (p < 0.001) in the silages treated with additives compared to the untreated silages, with no observed differences among the additives. Yeast counts were similar among silages treated with formic acid, sodium benzoate, and the control, as well as among those treated with all additives. However, the yeast count was, on average, 52.8 % lower (4.81 vs. 2.27 log₁₀ CFU g⁻¹) (p < 0.001) in silages treated with the microbial inoculant compared to untreated silages, with no observed differences between the doses. The aerobic stability of the silages treated with the additives was 2.97 times greater (p < 0.001) than that of untreated silages (196 vs. 66 h), with no significant differences noted among the additives.

Table 5
Microbial profile at silo opening and at aerobic stability loss of corn grain silage rehydrated with wet pulp orange treated with different additives and subjected or not to a 24-hour delay in sealing.

Variables		Additive					Mean	SEM	p-value
		Control	SB	FA	LB1	LB2			A	D	A × D
Microbial counts at silo opening (log10 CFU g−1)
Aerobic bacteria
	With delay	5.25	5.00	4.92	5.45	5.82	5.29 a	0.126	0.149	0.009	0.218
	Without delay	4.08	4.97	4.92	4.16	5.25	4.67 b
	Mean	4.67	4.99	4.92	4.81	5.54
Molds
	With delay	3.64	1.98	2.17	1.96	2.24	2.39 a	0.172	0.390	0.001	0.058
	Without delay	1.11	2.07	1.66	1.10	1.30	1.45 b
	Mean	2.38	2.03	1.92	1.53	1.77
Yeast
	With delay	4.07	2.70	2.52	3.31	2.9	3.23 a	0.339	0.006	< 0.001	0.695
	Without delay	3.26	1.65	1.90	1.18	1.49	1.89 b
	Mean	3.98 A	2.18 B	2.21 B	2.25 B	2.19 B
Microbial counts at aerobic stability loss (log10 CFU g−1)
Aerobic bacteria
	With delay	5.66	4.25	4.57	4.14	4.20	4.56 a	0.138	< 0.001	< 0.001	0.274
	Without delay	4.38	3.99	3.83	3.29	3.51	3.80 b
	Mean	5.02 A	4.12 B	4.20 B	3.71 B	3.86 B
Molds
	With delay	3.66	2.11	2.88	1.83	1.48	2.39 a	0.144	< 0.001	< 0.001	0.209
	Without delay	2.15	1.02	0.870	0.583	0.798	1.08 b
	Mean	2.91 A	1.56 B	1.87 B	1.20 B	1.14 B
Yeast
	With delay	4.91	3.58	3.48	2.89	2.77	3.53	0.437	< 0.001	0.385	0.452
	Without delay	4.70	4.27	3.62	1.70	1.74	3.21
	Mean	4.81 A	3.92 AB	3.55 AB	2.29 B	2.25 B
Aerobic stability, hours
	With delay	66.0	240	196	153	197	170	13.3	< 0.001	0.946	0.718
	Without delay	66.0	211	217	197	153	169
	Mean	66.0 B	226 A	206 A	175 A	175 A

• Means followed by distinct letters, uppercase in the line (comparing additive) and lowercase in the column (comparing delay sealing), differ by Tukey's test (p < 0.05). SB = sodium benzoate (0.2 %); FA = Formic acid (4.5 L t⁻¹); LB1 = microbial additive in dosage recommended by the manufacturer; LB2 = microbial additive in double of the dosage recommended by the manufacturer; SEM = standard error of mean; CFU = colony-forming units; A = additive effect; D = delayed sealing effect; A × D = interaction between additive and delayed sealing effect.

The interaction between sealing time and additives showed no significant impact on either pH or the concentration of NH₃-N/TN (p > 0.05) (Table 6). However, the application of additives significantly affected the results (p < 0.001), leading to a reduction in pH and NH₃-N/TN by 1.6 % (3.80 vs. 3.86) and 49.1 % (6.35 vs. 9.47 g kg⁻¹), respectively, in the treated silages compared to the control, with no notable differences among the additives. Additionally, the time after silo opening resulted in a linear increase (p < 0.001) in both pH and NH₃-N/TN levels in the silage.

Table 6
Evaluations of pH and NH₃-N/TN (g kg⁻¹) at different times after silo opening of corn grain silage rehydrated with wet pulp orange treated with different additives and subjected or not to a 24-hour delay in sealing.

Variables		Additive						Mean	SEM	p-value
		Control	SB	FA	LB1	LB2	A			D	T	A × D	A × T	D × T	A × D × T
pH
	T. 0 h	3.85	3.77	3.74	3.76	3.75	3.77	0.007	< 0.001	0.497	< 0.001	0.126	0.657	0.151	0.896
	T. 24 h	3.85	3.76	3.79	3.78	3.76	3.79
	T. 48 h	3.87	3.78	3.80	3.78	3.77	3.80
	T. 96 h	3.88	3.83	3.83	3.81	3.80	3.83
	T. 240 h	3.87	3.86	3.93	3.79	3.85	3.86
	Mean	3.86 A	3.80 B	3.82 AB	3.78 B	3.78 B
Amoniacal nitrogen/total nitrogen (g kg−1)
	T. 0 h	7.19	4.72	4.02	4.80	4.68	5.08	0.014	< 0.001	0.575	< 0.001	0.801	0.269	0.624	0.986
	T. 24 h	8.83	5.59	5.46	5.42	5.18	6.10
	T. 48 h	9.95	6.42	6.71	6.63	6.34	7.21
	T. 96 h	9.42	6.55	6.79	6.79	6.42	7.19
	T. 240 h	12.0	8.63	8.79	8.26	8.79	9.29
	Mean	9.47 A	6.38 B	6.36 B	6.38 B	6.28 B

• Means followed by distinct uppercase letter in the line (comparing additive) differ by Tukey's test (p < 0.05). SB = sodium benzoate (0.2 %); FA = Formic acid (4.5 L t⁻¹); LB1 = microbial additive in dosage recommended by the manufacturer; LB2 = microbial additive in double of the dosage recommended by the manufacturer; SEM = standard error of mean; A = additive effect; D = delayed sealing effect; T = time after silo opening; A × D = interaction between additive and delayed sealing effect; A × T = interaction between additive and time after silo opening; D × T = interaction between delayed sealing and time after silo opening; A × D × T = interaction between additive, delayed sealing and time after silo opening effect.

The interaction between sealing time and additives had a significant impact on lactic acid concentration (p = 0.002) (Table 7). In silages subjected to delayed sealing, lactic acid concentrations were 44.5 %, 74.8 %, and 28.6 % higher (p = 0.002) in untreated silages compared to those treated with sodium benzoate, formic acid, and LB1, respectively. In addition, lactic acid concentrations were lower in silages treated with sodium benzoate or formic acid than in those treated with LB2, with no significant observed difference between LB1 and LB2. Overall, untreated silages exhibited the highest lactic acid concentrations, followed by silages treated with sodium benzoate, LB1 and LB2 at intermediate levels, with the lowest concentrations found in those treated with formic acid. In untreated silages and those treated with sodium benzoate, lactic acid concentrations were reduced by 23.2 % and 32.4 %, respectively, when delayed sealing was employed compared to silages sealed immediately. In contrast, no significant differences were observed with the other additives.

Table 7
Acids produced in the fermentation of corn grain silage rehydrated with wet pulp orange treated with different additives and subjected or not to a 24-hour delay in sealing.

Variables		Additive					Mean	SEM	p-value
		Control	SB	FA	LB1	LB2			A	D	A × D
Lactic acid (g kg−1 DM)
	With delay	29.2 Ab	20.2 Cb	16.7 Ca	22.7 BCa	27.0 ABa	23.2	1.06	< 0.001	< 0.001	0.002
	Without delay	38.0 Aa	29.9 Ba	17.5 Da	23.6 Ca	26.6 BCa	27.1
	Mean	33.6	25.1	17.1	23.2	26.8
Acetic acid (g kg−1 DM)
	With delay	17.9 Ba	19.2 Ba	19.6 Ba	38.3 Aa	38.6 Aa	26.7	1.70	< 0.001	< 0.001	0.005
	Without delay	12.5 Bb	13.8 Bb	14.2 Bb	36.4 Ab	35.8 Ab	22.5
	Mean	15.2	16.5	16.9	37.4	37.2
Butyric acid (m g kg−1 DM)
	With delay	36.5	21.5	27.00	13.5	27.0	25.1 a	2.47	0.002	0.025	0.235
	Without delay	37.8	17.3	7.00	8.00	4.75	15.0 b
	Mean	37.1 A	19.4 B	17.0 B	10.8 B	15.9 B	-
Propionic acid (mg kg−1 DM)
	With delay	667	586	657	613	711	647	20.3	0.760	0.855	0.865
	Without delay	670	618	682	591	612	635
	Mean	669	602	670	602	662
Valeric acid (mg kg−1 DM)
	With delay	4.75	4.25	4.00	4.25	4.00	4.25	0.145	0.836	0.235	0.836
	Without delay	3.23	5.75	5.00	4.50	6.25	4.95
	Mean	3.99	5.00	4.50	4.38	5.13
Iso-valeric acid (mg kg−1 DM)
	With delay	16.3	16.8	16.3	16.3	17.3	16.6	0.366	0.768	0.659	0.756
	Without delay	17.3	14.5	16.5	15.8	17.0	16.2
	Mean	16.8	15.7	16.4	16.1	17.2
Iso-butyric acid (mg kg−1 DM)
	With delay	9.00	7.50	9.50	9.00	9.50	8.90	0.421	0.350	0.398	0.943
	Without delay	9.25	6.25	9.50	7.25	8.50	8.15
	Mean	9.13	6.88	9.50	8.13	9.00

• Means followed by distinct letters, uppercase in the line (comparing additive) and lowercase in the column (comparing delay sealing), differ by Tukey's test (p < 0.05). SB = sodium benzoate (0.2 %); FA = Formic acid (4.5 L t⁻¹); LB1 = microbial additive in dosage recommended by the manufacturer; LB2 = microbial additive in double of the dosage recommended by the manufacturer; SEM = standard error of mean; DM = Dry matter; A = additive effect; D = delayed sealing effect; A × D = interaction between additive and delayed sealing effect.

The interaction between sealing time and additives significantly influenced acetic acid concentration (p = 0.005) (Table 7). In silages that were sealed immediately and those subjected to delayed sealing, acetic acid concentrations were 2.03 and 2.68 times higher (p = 0.005) when using LB1 or LB2, respectively, compared to the other silages. On average, delaying the sealing process resulted in a 26.8 % increase in acetic acid concentration across all silages. However, the interaction between sealing time and additives had no significant impact on the concentrations of butyric acid, propionic acid, valeric acid, iso-valeric acid, and iso-butyric acid (p > 0.05). Specifically, the butyric acid concentration was 67.3 % lower (p = 0.025) in silages sealed immediately compared to those subjected to delayed sealing and 57.5 % lower (p = 0.002) in treated silages versus untreated ones, with no discernible differences among the additives.

The interaction between sealing time and additives did not significantly influence ethanol concentration (p = 0.123) (Table 8). Silages sealed immediately exhibited a 64.3 % reduction in ethanol concentration (p < 0.001) compared to those sealed under delayed conditions. Furthermore, untreated silages presented a higher ethanol concentration (p < 0.001) than those treated with sodium benzoate and formic acid. When comparing treatments, silages treated with formic acid demonstrated a lower ethanol concentration (p < 0.001) than those treated with LB1. In terms of 1,2-propanediol concentration, the interaction between sealing time and additives was significant (p = 0.013). In silages subjected to delayed sealing, the concentration of 1,2-propanediol was 86.3 % higher (423 vs. 788 mg kg⁻¹ DM) (p = 0.013) in the silages treated with LB1 or LB2 compared with untreated silages or those treated with sodium benzoate. In contrast, the concentrations in silages treated with formic acid were consistent across all treatments. In silages sealed immediately, the concentrations were similar between additives. Notably, the concentrations of 1,2-propanediol were 2.40, 2.76, and 3.01 times higher in silages treated with formic acid, LB1, and LB2, respectively, when delayed sealing was applied compared to immediate sealing. The interaction between sealing time and additives did not significantly affect the concentrations of 2,3-butanediol or ethyl lactate (p > 0.05). However, the 2,3-butanediol concentration was substantially higher at 97.8 % (p < 0.001) in the silages under delayed sealing compared to those sealed immediately. Additionally, the ethyl lactate concentration was 22.2 % higher (77.6 vs. 63.5 mg kg⁻¹ DM) (p < 0.001) in untreated silages and those treated with LB1 and LB2 compared to those treated with sodium benzoate and formic acid.

Table 8
Fermentative compounds of corn grain silage rehydrated with wet pulp orange treated with different additives and subjected or not to a 24-hour delay in sealing.

Variables		Additive					Mean	SEM	p-value
		Control	SB	FA	LB1	LB2			A	D	A × D
Ethanol (g kg−1 DM)
	With delay	4.40	3.80	3.40	4.80	4.33	4.15 a	0.230	< 0.001	< 0.001	0.123
	Without delay	2.15	1.38	1.08	1.58	1.20	1.48 b
	Mean	3.28 A	2.59 BC	2.24 C	3.19 AB	2.76 ABC
1,2-propanediol (mg kg−1 DM)
	With delay	429 Ba	416 Ba	713 ABa	800 Aa	775 Aa	627	37.9	0.102	< 0.001	0.013
	Without delay	425 Aa	255 Aa	294 Ab	290 Ab	257 Ab	304
	Mean	427	336	504	545	516
2,3-butanediol (mg kg−1 DM)
	With delay	2308	2700	2525	2636	2523	2538 a	103	0.098	< 0.001	0.236
	Without delay	1197	1331	1275	1313	1302	1283 b
	Mean	1752	2016	1900	1974	1912
Ethyl lactate (mg kg−1 DM)
	With delay	77.0	67.8	60.8	79.8	74.8	72.0	1.42	< 0.001	0.957	0.599
	Without delay	76.3	62.0	63.3	81.0	77.0	71.9
	Mean	76.6 A	64.9 B	62.0 B	80.4 A	75.9 A
1-propanol (mg kg−1 DM)
	With delay	1204 Aa	1208 Aa	1017 Aa	1066 Ab	1089 Ab	1117	68.8	0.008	0.002	0.030
	Without delay	1223 Ba	1137 Ba	1136 Ba	2065 Aa	1775 ABa	1467
	Mean	1214	1173	1077	1566	1432
2-butanol (mg kg−1 DM)
	With delay	19.0	21.5	22.5	22.5	24.3	22.0	0.581	0.179	0.236	0.415
	Without delay	22.5	25.8	21.5	21.3	25.5	23.3
	Mean	20.8	23.7	22.0	21.9	24.9
Propyl acetate (mg kg−1 DM)
	With delay	39.0	38.8	37.8	37.5	38.8	38.4	0.510	0.646	0.053	0.949
	Without delay	42.0	39.5	39.3	39.8	41.8	40.5
	Mean	40.5	39.2	38.6	38.7	40.3
Isopropyl alcohol (mg kg−1 DM)
	With delay	1.00	1.75	1.75	0.25	3.00	1.55	0.207	0.310	0.328	0.266
	Without delay	1.25	1.00	1.00	1.25	1.25	1.15
	Mean	1.13	1.38	1.38	0.75	2.13
Methanol (mg kg−1 DM)
	With delay	815	609	527	669	560	636	62.9	0.788	0.408	0.780
	Without delay	936	708	832	1031	1010	903
	Mean	876	659	680	850	785
Ethyl acetate (mg kg−1 DM)
	With delay	54.5	41.8	45.5	47.3	57.3	49.3	2.81	0.776	0.923	0.901
	Without delay	47.3	44.8	44.5	56.5	50.3	48.7
	Mean	50.9	43.3	45.0	51.9	53.8
Acetone (mg kg−1 DM)
	With delay	17.3	20.3	17.0	18.0	18.8	18.3	0.474	0.709	0.265	0.427
	Without delay	17.8	18.0	20.5	20.0	20.5	19.4
	Mean	17.6	19.2	18.8	19.0	19.7

• Means followed by distinct letters, uppercase in the line (comparing additive) and lowercase in the column (comparing delay sealing), differ by Tukey's test (p < 0.05). SB = sodium benzoate (0.2 %); FA = Formic acid (4.5 L t⁻¹); LB1 = microbial additive in dosage recommended by the manufacturer; LB2 = microbial additive in double of the dosage recommended by the manufacturer; SEM = standard error of mean; DM = Dry matter; A = additive effect; D = delayed sealing effect; A × D = interaction between additive and delayed sealing effect.

The interaction between sealing time and additives significantly influenced the concentration of 1-propanol (p = 0.030) (Table 8). In silages sealed immediately, the 1-propanol concentrations were notably higher (p = 0.030) in silages treated with LB1 compared to those treated with formic acid, sodium benzoate, and the untreated silages. Additionally, silages treated with LB2 exhibited concentrations equivalent to those of all additives. In silages treated with LB1 and LB2, the 1-propanol concentration was 93.7 and 63 % higher, respectively, when sealed immediately compared to those sealed with a delay. No significant effects were observed for the concentrations of 2-butanol, propyl acetate, isopropyl alcohol, methanol, ethyl acetate, or acetone (p > 0.05).

Discussion

The elevated DM concentration observed in the silages treated with sodium benzoate and formic acid is likely due to these additives directly inhibiting the growth of spoilage microorganisms during fermentation. This inhibition reduces nutrient consumption, especially water-soluble carbohydrates (WSC), ultimately leading to an increase in silage DM concentration. Similar increases in DM and reductions in WSC were observed by Daniel et al. (2022) in corn grain silage treated with sodium benzoate, and by Menezes et al. (2023) in forage silage treated with formic acid. Conversely, silages subjected to the highest levels of lactic acid bacteria (LB) exhibited lower DM concentrations. This may be attributed to these microorganisms employing a heterofermentative metabolic pathway that converts lactic acid and WSC into acetic acid, ethanol, and carbon dioxide (McDonald et al., 1991).

The lower WSC concentration also accounts for the higher IVDMD observed in silages treated with formic acid and sodium benzoate compared to those treated with LB1 and the untreated silages. The increased NDIP concentration in silages under delayed sealing likely resulted from the Maillard reaction, which can decrease protein availability (Krizsan and Randby, 2007). In the present study, silages under delayed sealing exhibited a temperature of 28.9 °C at the time of silo opening, whereas those sealed immediately showed a temperature of 26.2 °C. The Maillard reaction typically occurs at temperatures above 40 °C in silages (Muck et al., 2003); however, higher concentrations of acid detergent-insoluble crude protein and heat-damaged protein have been observed at higher temperatures in corn silage at either 20 °C or 40 °C (Kim and Adesogan, 2006). This implies that increases in insoluble nitrogen fractions may still occur even at comparatively lower temperatures. Furthermore, the temperature during the entire storage period, which was not monitored in this study, may have surpassed the values recorded at the time of silo opening. Although this increase in less digestible protein fractions may reduce protein availability for animals, the delay in sealing did not affect the other chemical composition parameters, suggesting that delayed sealing does not influence silage composition.

The findings of the present study suggest that the evaluated additives effectively enhance the aerobic stability of corn grain silages rehydrated with WPO. This aligns with previous research highlighting the crucial role of both microbial and chemical additives in improving silage preservation (Muck et al., 2018). When selecting additives, it is important to consider not only their effectiveness but also factors such as cost, regional availability, ease of application, and employee safety. Among the additives tested, sodium benzoate and L. buchneri are often preferred due to their greater commercial availability. Nonetheless, formic acid remains a viable alternative, particularly with the advent of buffered formulations that minimize corrosiveness while maintaining adequate acidification capacity (Wilkinson and Rinne, 2018).

Another critical factor to consider when selecting the additive is the time required for the silage to be utilized. The application of sodium benzoate to corn grain silages has been shown to enhance aerobic stability (24 vs. > 500 h) after just 21 days of fermentation (Silva et al., 2015). In contrast, L. buchneri exhibits slow growth, requiring approximately 60 days to reach full maturity (Muck et al., 2018). Notably, substantial differences in acetic acid levels between untreated silages and those inoculated with L. buchneri were observed only after 60 days of storage, further underscoring the slow growth of this microorganism (Silva et al., 2019). Thus, where there is a need for the timely use of the silage, sodium benzoate should be the preferred choice. Additionally, since sodium benzoate operates through a chemical mechanism, its antifungal effect is not dependent on the growth of added microorganisms. This feature may enhance its effectiveness compared to L. buchneri in commercial farm settings, particularly when the ensiling process is not optimally managed, and issues such as the use of water containing antimicrobial residues or improper additive storage are prevalent.

The aerobic bacteria and mold counts at the point of aerobic stability loss were found to be lower in silages treated with additives compared to those that were untreated ones, demonstrating the effectiveness of all additives in controlling the growth of these microorganisms. However, only L. buchneri demonstrated a reduction in yeast counts compared to untreated silages. These findings align with those of Morais et al. (2017), who conducted a meta-analytic analysis revealing that a combination of homo- and heterofermentative bacteria consistently decreased yeast counts by increasing acetic acid concentration in corn grain silage.

Increased aerobic stability was also observed in corn grain silage treated with sodium benzoate by Kleinschmit and Kung Jr (2006), Santos et al. (2019), and Daniel et al. (2022), as well as in corn grain silage treated with L. buchneri by Torres et al. (2021) and Silva et al. (2019). The increase observed in silage treated with sodium benzoate is attributed to the additive's direct action. Conversely, in silages treated with L. buchneri, the enhancement is associated with an increase in acetic acid (Oude Elferink et al., 2001), as observed in the present study. Providing silage with low aerobic stability that has experienced aerobic spoilage can reduce intake and overall performance (Gerlach et al., 2013).

The addition of formic acid may reduce aerobic stability by failing to control yeast growth adequately and by increasing water-soluble carbohydrate (WSC) concentrations (McDonald et al., 1991). An enhancement in aerobic stability was observed when formic acid was used in combination with other additives, such as propionic acid, benzoic acid, and potassium sorbate, in a meta-analytical study by Menezes et al. (2023) examining formic-acid-treated forage silages. Conversely, formic acid effectively inhibits the growth of undesirable bacteria by rapidly acidifying the ensiled material (Muck et al., 2018), which may help manage Acetobacter bacteria associated with reduced aerobic stability in corn silages rehydrated with WPO (Menezes et al., 2022a).

The LB additive contained homofermentative microorganisms that likely enhanced lactic acid production, resulting in a lower acid dissociation constant (pKa) and a reduced pH. In the present study, the pH observed was lower than that reported by Kung Jr et al. (2018), which is advantageous, as benzoic acid has a pKa of 4.20, allowing a greater proportion of its active (undissociated) form at the pH levels in our samples. Additionally, WPO is rich in water-soluble carbohydrates (WSC) (Bampidis and Robinson, 2006; Menezes et al., 2022a), which may facilitate lactic acid production and further lower the pH levels observed in this study.

In contrast, the meta-analytical study conducted by Menezes et al. (2023) involving forage silages treated with formic acid revealed that improvements in aerobic stability were observed only when formic acid was mixed with other additives. This finding is noteworthy, as formic acid directly impacts acidification and reduces the growth of deteriorative microorganisms, which can enhance silage stability. However, it also leads to a decrease in total acid production within the silage. This reduction in acid production may hinder the silage's ability to inhibit mold and yeast growth after the silo is opened, ultimately affecting aerobic stability. Additionally, in the present study, the silos were stored in a covered shed, which may have reduced the proliferation of deteriorative microorganisms. Therefore, further research under commercial conditions is necessary to determine whether the increased aerobic stability observed in formic-acid-treated silages is reliable.

Higher counts of aerobic bacteria, molds, and yeasts were observed during silo opening, as well as elevated levels of aerobic bacteria and mold in delayed-sealing silages, likely due to increased growth of these microorganisms during silage exposure. However, this heightened microbial growth did not compromise aerobic stability. These findings suggest that delayed sealing can be implemented when necessary without negatively impacting aerobic stability. Nevertheless, although the reduction in silage quality was minimal, it is advisable to proceed with ensiling shortly after WPO production.

The primary fermentation products exhibited values within the parameters established by Kung Jr et al. (2018), indicating that the mixture of corn grain silage with WPO yielded good fermentation quality. Furthermore, the silage demonstrated low DM loss and good aerobic stability when supplemented with additives, reinforcing the findings of Menezes et al. (2022a) that storing WPO in this manner is a technically viable strategy. Given that corn is a high-cost feed rich in nutrients, it is essential to adopt strategies to maximize its utilization. The results of the present study highlight the importance of using additives in producing high-quality, stable silage (Daniel et al., 2022). Additionally, it is noteworthy that the production of rehydrated silage enhances corn utilization by improving feed efficiency and total-tract digestibility of DM and starch, while reducing DM and metabolizable energy intakes (Jacovaci et al., 2021).

The elevated lactic acid concentration in untreated silages is likely attributed to the inhibitory effects of sodium benzoate and formic acid on the growth of lactic acid bacteria. In contrast, L. buchneri aids in the conversion of lactic acid into acetic acid, ethanol, and 1,2-propanediol (Oude Elferink et al., 2001). However, this study observed higher levels of these compounds without a significant decrease in lactic acid concentration. This may be attributed to the presence of homofermentative bacteria in the inoculant. Additionally, an increase in acetic acid and 1,2-propanediol concentrations in corn grain silage treated with L. buchneri was also observed by Silva et al. (2018) and Silva et al. (2019). The increase in 1-propanol concentrations in silages treated with LB1, reported by Silva et al. (2018), likely resulted from these bacteria increasing 1,2-propanediol concentrations. Specific bacteria within the epiphytic microbiota, such as Lactobacillus diolivorans, utilize 1,2-propanediol as a sole carbon source to synthesize propionic acid and 1-propanol under aerobic conditions (Zielińska et al., 2017). Higher concentrations of acetic acid are crucial, as it inhibits the growth of molds and yeasts, thereby lowering their populations and enhancing the aerobic stability observed in the present study.

The application of additives and immediate silo sealing effectively reduced levels of butyric acid and ethanol, likely by suppressing Clostridium and Enterobacter growth. Decreases in ethanol concentration and yeast count in corn grain silage treated with sodium benzoate have also been observed by Silva et al. (2015). The increased concentration of 2,3-butanediol in silages subjected to delayed sealing appears to be associated with increased Enterobacter growth. This suggests that silage should ideally be prepared immediately after WPO production without a delay in sealing.

The application of sodium benzoate, formic acid, or a microbial inoculant containing E. faecium M74, L. plantarum CH6072, and L. buchneri LB1819 effectively reduces yeast counts, enhances aerobic stability, and improves overall quality of rehydrated corn grain silages combined with wet pulp orange (WPO). These microbial inoculants also lead to an increase in acetic acid concentration. Additionally, delaying the sealing of the silage by up to 24 h can boost microbial counts without significantly compromising silage quality, indicating that this could be a viable management strategy when necessary. Further research should focus on determining the optimal inclusion rate of this silage in animal diets and assessing its impacts on animal performance and economic returns.

• Declaration of use of AI Technologies
  No AI technologies were used.

Data availability statement

The data are available under reasonable request.

Acknowledgments

The authors would like to thank the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES), the Escola de Veterinária of Universidade Federal de Minas Gerais (UFMG), the Fundação de Amparo à Pesquisa do Estado de Minas Gerais (FAPEMIG) and the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq).

References

• Association of Official Analytical Chemists – International [AOAC]. 1990. Official Methods of Analysis. 16ed. AOAC, Rockville, MD, USA.
• Association of Official Analytical Chemists – International [AOAC]. 2012. Official Methods of Analysis. 18ed. AOAC, Rockville, MD, USA.
• Ávila CLS, Carvalho BF. 2020. Silage fermentation: updates focusing on the performance of micro-organisms. Journal of Applied Microbiology 128: 966-984. https://doi.org/10.1111/jam.14450
  » https://doi.org/10.1111/jam.14450
• Bampidis VA, Robinson PH. 2006. Citrus by-products as ruminant feeds: a review. Animal Feed Science and Technology 128: 175-217. https://doi.org/10.1016/j.anifeedsci.2005.12.002
  » https://doi.org/10.1016/j.anifeedsci.2005.12.002
• Daniel JLP, Auerbach HU, Gomes ALM, Weiss K. 2022. Applying the benzoate equivalent concept in rehydrated corn grain silage. Animal Feed Science and Technology 294: 115482. https://doi.org/10.1016/j.anifeedsci.2022.115482
  » https://doi.org/10.1016/j.anifeedsci.2022.115482
• Ferraretto LF, Fredin SM, Muck RE, Shaver RD. 2016. Case study: Microbial inoculant and ensiling time effects on fermentation profile, nitrogen fractions, and ruminal in vitro and in situ starch digestibility in corn shredlage and late-maturity corn silage. The Professional Animal Scientist 32: 861-868. https://doi.org/10.15232/pas.2016-01546
  » https://doi.org/10.15232/pas.2016-01546
• Ferraretto LF, Silva Filho WI, Fernandes T, Kim DH, Sultana H. 2018. Effect of ensiling time on fermentation profile and ruminal in vitro starch digestibility in rehydrated corn with or without varied concentrations of wet brewers grains. Journal of Dairy Science 101: 4643-4649. https://doi.org/10.3168/jds.2017-14329
  » https://doi.org/10.3168/jds.2017-14329
• Gerlach K, Roß F, Weiß K, Buscher W, Südekum KH. 2013. Changes in maize silage fermentation products during aerobic deterioration and effects on dry matter intake by goats. Agricultural and Food Science 22: 168-181. https://doi.org/10.23986/afsci.6739
  » https://doi.org/10.23986/afsci.6739
• Holden LA. 1999. Comparison of methods of in vitro dry matter digestibility for ten feeds. Journal of Dairy Science 82: 1791-1794. https://doi.org/10.3168/jds.S0022-0302(99)75409-3
  » https://doi.org/10.3168/jds.S0022-0302(99)75409-3
• Jacovaci FA, Salvo PAR, Jobim CC, Daniel JLP. 2021. Effect of ensiling on the feeding value of flint corn grain for feedlot beef cattle: a meta-analysis. Revista Brasileira de Zootecnia 50: e20200111. https://doi.org/10.37496/rbz5020200111
  » https://doi.org/10.37496/rbz5020200111
• Kim SC, Adesogan AT. 2006. Influence of ensiling temperature, simulated rainfall, and delayed sealing on fermentation characteristics and aerobic stability of corn silage. Journal of Dairy Science 89: 3122-3132. https://doi.org/10.3168/jds.S0022-0302(06)72586-3
  » https://doi.org/10.3168/jds.S0022-0302(06)72586-3
• Kleinschmit DH, Kung Jr L. 2006. A meta-analysis of the effects of Lactobacillus buchneri on the fermentation and aerobic stability of corn and grass and small-grain silages. Journal of Dairy Science 89: 4005-4013. https://doi.org/10.3168/jds.S0022-0302(06)72444-4
  » https://doi.org/10.3168/jds.S0022-0302(06)72444-4
• Krizsan SJ, Randby AT. 2007. The effect of fermentation quality on the voluntary intake of grass silage by growing cattle fed silage as the sole feed. Journal of Animal Science 85: 984-996. https://doi.org/10.2527/jas.2005-587
  » https://doi.org/10.2527/jas.2005-587
• Kung Jr L, Shaver RD, Grant RJ, Schmidt RJ. 2018. Silage review: interpretation of chemical, microbial, and organoleptic components of silages. Journal of Dairy Science 101: 4020-4033. https://doi.org/10.3168/jds.2017-13909
  » https://doi.org/10.3168/jds.2017-13909
• McDonald P, Henderson AR, Heron SJE. 1991. The Biochemistry of Silage. Chalcombe Publications, Marlow, England
• Menezes GL, Oliveira AF, Lana AMQ, Pires FPAA, Menezes RA, Sousa PG, et al. 2022a. Effects of different moist orange pulp inclusions in the corn grain rehydration for silage production on chemical composition, fermentation, aerobic stability, microbiological profile, and losses. Animal Science Journal 93: e13701. https://doi.org/10.1111/asj.13701
  » https://doi.org/10.1111/asj.13701
• Menezes GL, Oliveira AF, Gonçalves LC, Pires FPAA, Menezes RA, Sousa PG, et al. 2022b. Efficacy of formic acid, enzymes, and microbial additives in silage on the performance of sheep: Systematic review and meta-analysis. Small Ruminant Research 212: 106706. https://doi.org/10.1016/j.smallrumres.2022.106706
  » https://doi.org/10.1016/j.smallrumres.2022.106706
• Menezes GL, Oliveira AF, Pires FPAA, Gonçalves LC, Menezes RA, Sousa PG, et al. 2023. Efficacy of adding chemical and microbial additives to silage on beef cattle performance: Systematic review and meta-analysis. Grass and Forage Science 78: 1-22. https://doi.org/10.1111/gfs.12579
  » https://doi.org/10.1111/gfs.12579
• Morais G, Daniel JLP, Kleinshmitt C, Carvalho PA, Fernandes J, Nussio LG. 2017. Additives for grain silages: a review. Slovak Journal of Animal Science 50: 42-54.
• Muck RE, Moser LE, Pitt RE. 2003. Postharvest factors affecting ensiling. p. 251-304. In: Buxton DR, Muck RE, Harrison JH. eds. Silage science and technology. Soil Science Society of America, Madison, WI, USA. https://doi.org/10.2134/agronmonogr42.c6
  » https://doi.org/10.2134/agronmonogr42.c6
• Muck RE, Nadeau EMG, McAllister TA, Contreras-Govea FE, Santos MC, Kung Jr L. 2018. Silage review: recent advances and future uses of silage additives. Journal of Dairy Science 101: 3980-4000. https://doi.org/10.3168/jds.2017-13839
  » https://doi.org/10.3168/jds.2017-13839
• National Research Council [NRC]. 2001. Nutrient Requirements of Dairy Cattle. 7ed. National Academy Press, Washington, DC, USA.
• Oude Elferink SJWH, Krooneman J, Gottschal JC, Spoelstra SF, Faber F, Driehuis F. 2001. Anaerobic conversion of lactic acid to acetic acid and 1,2-propanediol by Lactobacillus buchneri. Applied and Environmental Microbiology 67: 125-132. https://doi.org/10.1128/AEM.67.1.125-132.2001
  » https://doi.org/10.1128/AEM.67.1.125-132.2001
• Pedroso AF, Nussio LG, Loures DRS, Paziani SF, Ribeiro JR, Mari LJ, et al. 2008. Fermentation, losses, and aerobic stability of sugarcane silages treated with chemical or bacterial additives. Scientia Agricola 65: 589-594. https://doi.org/10.1590/S0103-90162008000600004
  » https://doi.org/10.1590/S0103-90162008000600004
• Pitt JI, Hocking AD. 2009. Fungi and Food Spoilage. 3ed. Springer, London, New York, USA.
• Pryce JD. 1969. A modification of Barker-Summerson method for the determination of lactic acid. Analyst Journal 94: 1151-1152. https://doi.org/10.1039/an9699401151
  » https://doi.org/10.1039/an9699401151
• Ranjit NK, Kung Jr L. 2000. The effect of Lactobacillus buchneri, Lactobacillus plantarum, or a chemical preservative on the fermentation and aerobic stability of corn silage. Journal of Dairy Science 83: 526-535. https://doi.org/10.3168/jds.S0022-0302(00)74912-5
  » https://doi.org/10.3168/jds.S0022-0302(00)74912-5
• Santos WP, Salvati GGS, Arthur BAV, Daniel JLP, Nussio LG. 2019. The effect of sodium benzoate on the nutritive value of rehydrated sorghum grain silage for dairy cows. Animal Feed Science and Technology 256: 114267. https://doi.org/10.1016/j.anifeedsci.2019.114267
  » https://doi.org/10.1016/j.anifeedsci.2019.114267
• Silva TC, Smith ML, Barnard AM, Kung Jr L. 2015. The effect of a chemical additive on the fermentation and aerobic stability of high-moisture corn. Journal of Dairy Science 98: 8904-8912. https://doi.org/10.3168/jds.2015-9640
  » https://doi.org/10.3168/jds.2015-9640
• Silva NC, Nascimento CF, Nascimento FA, Resende FD, Daniel JLP, Siqueira GR. 2018. Fermentation and aerobic stability of rehydrated corn grain silage treated with different doses of Lactobacillus buchneri or a combination of Lactobacillus plantarum and Pediococcus acidilactici Journal of Dairy Science 101: 4158-4167. https://doi.org/10.3168/jds.2017-13797
  » https://doi.org/10.3168/jds.2017-13797
• Silva NC, Nascimento CF, Campos VMA, Alves MAP, Resende FD, Daniel JLP, et al. 2019. Influence of storage length and inoculation with Lactobacillus buchneri on the fermentation, aerobic stability, and ruminal degradability of high-moisture corn and rehydrated corn grain silage. Animal Feed Science and Technology 251: 124-133. https://doi.org/10.1016/j.anifeedsci.2019.03.003
  » https://doi.org/10.1016/j.anifeedsci.2019.03.003
• Tilley JMA, Terry RA. 1963. A two-stage technique for the in vitro digestion of forage crops. Grass and Forage Science 18: 104-111. https://doi.org/10.1111/j.1365-2494.1963.tb00335.x
  » https://doi.org/10.1111/j.1365-2494.1963.tb00335.x
• Torres RNS, Ghedini CP, Coelho LM, Ezequiel JMB, Almeida Júnior GA, Almeida MTC. 2021. Meta-analysis of the effects of silage additives on high-moisture grain silage quality and performance of dairy cows. Livestock Science 251: 104618. https://doi.org/10.1016/j.livsci.2021.104618
  » https://doi.org/10.1016/j.livsci.2021.104618
• Van Soest PJ, Robertson JB, Lewis BA. 1991. Methods for dietary fiber, neutral detergent fiber, and nonstarch polysaccharides in relation to animal nutrition. Journal of Dairy Science 74: 3583-3597. https://doi.org/10.3168/jds.S0022-0302(91)78551-2
  » https://doi.org/10.3168/jds.S0022-0302(91)78551-2
• Wilkinson JM, Rinne M. 2018. Highlights of progress in silage conservation and future perspectives. Grass and Forage Science 73: 40-52. https://doi.org/10.1111/gfs.12327
  » https://doi.org/10.1111/gfs.12327
• Zema DA, Calabrò PS, Folino A, Tamburino V, Zappia G, Zimbone SM. 2018. Valorisation of citrus processing waste: a review. Waste Management 80: 252-273. https://doi.org/10.1016/j.wasman.2018.09.024
  » https://doi.org/10.1016/j.wasman.2018.09.024
• Zielińska K, Fabiszewska A, Świątek M, Szymanowska-Powałowska D. 2017. Evaluation of the ability to metabolize 1,2-propanediol by heterofermentative bacteria of the genus Lactobacillus Electronic Journal of Biotechnology 26: 60-63. https://doi.org/10.1016/j.ejbt.2017.01.002
  » https://doi.org/10.1016/j.ejbt.2017.01.002

Edited by

• Edited by:
  Antonio Faciola https://orcid.org/0000-0003-0935-6233

Publication Dates

• Publication in this collection
  18 May 2026
• Date of issue
  2026

History

• Received
  13 Apr 2024
• Accepted
  13 Aug 2025