Increasing emphasis has been placed on the adoption of natural compounds as replacements for synthetic chemical additives in various agribusiness systems. In silage production, interest in essential oils as additives remains limited, and the available literature is limited to trials with forages (Kung Jr. et al., 2008; Chaves et al., 2012; Soycan-Önenç et al., 2015; Foskolos et al., 2016; Hodjatpanah-Montazeri et al., 2016; Maghsoudloo et al., 2017; Soycan Önenç et al., 2017; Turan and Önenç, 2018; Önenç and Turgud, 2019; Besharati and Niazifar, 2020; Besharati et al., 2020; Cantoia Júnior et al., 2020). Essential oils have shown favorable results on the fermentation and aerobic stability of whole plant corn, sugarcane and alfalfa silage, probably due to the antimicrobial properties of terpenes and phenolic compounds in their composition (Chao et al., 2000; Dorman and Deans, 2000), which interfere with the development and the activity of epiphytic microbiota.

Except for the work of Foskolos et al. (2016), however, the reported effects on silage have been associated with crude oils extracted from plants rather than with specific molecular compounds isolated from essential oils. Furthermore, there is a lack of information regarding the use of essential oils as additives for grain silage, particularly those based on cereals such as corn and sorghum. Corn grains, for instance, can be rehydrated and ensiled both to preserve a surplus from the harvest and as a processing technique to increase starch availability for ruminants (Santos et al., 2016). From a processing perspective, the greatest improvements in starch digestibility are achieved within the first 60 days of storage due to the accelerated degradation of zeins (Fernandes et al., 2021; Oliveira et al., 2023). In emergency storage scenarios, it is important to determine whether the silos can be used earlier without compromising conservation or nutritional value. Shorter storage lengths of 15 and 30 days have been evaluated with rehydrated corn grain silage (RCGS) (Hoffman et al., 2011; Oliveira et al., 2023; Silva Neto et al., 2023). Rehydrated corn grain silage, however, may present favorable conditions for undesirable clostridial development when epiphytic fermentation is not capable of adequately acidifying and preserving the silage.

Thus, the present study aimed to evaluate commercially purified essential oil compounds to assess molecule-specific effects when applied as additives for RCGS. These compounds were expected to modulate silage fermentation toward improved conservation quality and to enhance post-opening aerobic stability after 30 and 60 d of storage.

There was effect of strategy (P = 0.01) on DM content, although only the CA1000 silage presented a lower DM content than the CA200 silage, and none of the strategies differed from the control silage (Table 2). There was a decrease in DM content (P = 0.03) when storage length was extended from 30 to 60 d. There was an interaction effect between strategy and storage length (P<0.01) on DM losses. The SB strategy decreased DM losses, whereas the MI strategy did not differ from the control silage. At 30 d of storage, the CA600 strategy increased DM losses, whereas the CI600, CI1000, and LI200 strategies decreased DM losses compared with the control silage; however, only the CI600 and CI1000 strategies did not differ from the SB strategy (Table 2). At 60 d of storage, CA600 and CA1000 strategies presented greater DM losses than the control silage, while CI600 and CI1000 strategies decreased DM losses, although only the CI1000 strategy did not differ from the SB strategy. Dry matter losses increased in the CA1000 strategy when storage length was extended from 30 to 60 d. There was an interaction effect between strategy and storage length (P<0.01) on silage pH (Table 2). In comparison to the control silage, both SB and MI strategies decreased pH at 30 days of storage, and only SB decreased pH at 60 d. Both CA600 and CA1000 silage presented greater pH values than the control silage at both storage lengths. At 30 d of storage, all strategies for cinnamaldehyde, D-limonene and citral (except CR200) reduced pH values relative to the control silage, and at 60 d of storage, only the cinnamaldehyde-treated silage mantained lower pH values than the control silage (Table 2). There was an increase in pH with the extension of storage length from 30 to 60 d for all strategies.

Table 2

Item1	DM (g/kg FM)		DM losses (g/kg)		pH
Storage length (days)	30	60	30	60	30	60
CT	622	608	4.28bcA	4.20cdA	3.97bcB	4.00cdA
CA200	630	609	4.93abA	5.33bcA	3.99abB	4.03bcA
CA600	619	610	5.63aA	5.98bA	4.02aB	4.07bA
CA1000	601	602	5.50abB	9.43aA	4.03aB	4.21aA
CI200	617	608	3.70cdA	3.48deA	3.92defB	3.96efghA
CI600	610	607	2.13efA	2.65efA	3.90efB	3.96fghA
CI1000	612	612	1.28fA	1.73fgA	3.89fB	3.94hA
LI200	608	610	3.03deA	3.40deA	3.94dB	3.98defgA
LI600	616	611	4.17bcdA	4.38cdA	3.94dB	3.99defA
LI1000	610	609	3.20cdeA	3.75deA	3.94dB	3.98defgA
CR200	609	614	3.45cdA	4.00dA	3.94cdB	3.98defgA
CR600	618	614	3.38cdA	3.68deA	3.92defB	3.98defghA
CR1000	608	609	3.10cdeA	3.60deA	3.93deB	3.97defghA
SB	615	613	1.38fA	1.30gA	3.93deB	3.95ghA
MI	602	607	3.55cdA	4.13cdA	3.94dB	3.99cdeA
SEM	0.88		0.15		0.01
P: Strategy	0.01		<0.01		<0.01
P: Storage length	0.03		<0.01		<0.01
P: Strategy × Storage length	0.23		<0.01		<0.01

DM - dry matter; FM - fresh matter; SEM - standard error of the mean.

There was effect of interaction between strategy and storage length (P<0.01) on LAB counts (Table 3). The SB silage presented the lower counts at both storage lengths, while the MI silage had greater LAB counts at 30 d of storage, but lower counts at 60 d in comparison to the control silage (Table 3). At 30 d of storage, LAB counts for all essential oil compound-based strategies did not differ from the control silage, and at 60 d of storage, all strategies for cinnamaldehyde, D-limonene and citral reduced LAB counts relative to the control silage (Table 3). All strategies reduced LAB counts as storage length was extended from 30 to 60 d. There was an effect of interaction between strategy and storage length (P<0.01) on yeast counts (Table 3). The SB strategy decreased yeast counts to below the minimum threshold for counting at both storage lengths, while the MI silage did not differ from the control silage. At 30 d of storage, CA600 and CI1000 silage also presented lower yeast counts than the control silage, with CI1000 not differing from the SB silage. At 60 days of storage, only the CA1000 silage presented lower yeast counts than the control silage, also not differing from the SB silage. When storage length was extended from 30 to 60 days, there was an increase in yeast counts in CA200, CA600, CI1000, LI600, CR200 and CR600 silage, and a decrease only on the CA1000 silage. There was an effect of strategy (P<0.01) on filamentous fungi counts (Figure 1). The SB strategy decreased fungi counts relative to the control silage, while MI silage did not differ from the control silage. The CI1000 strategy also decreased fungi counts, and it did not differ from the SB strategy. There was a decrease in fungi counts when storage length was extended from 30 to 60 d (P<0.01). There was an effect of strategy (P<0.01) on clostridial spore counts. Neither the SB nor the MI silage differed from the control silage, and there was an increase in spore counts in the CA1000, CI200, and LI1000 silages relative to the control silage. There was an interaction effect between strategy and storage length (P<0.01) on lactic acid concentration (Table 4). Both the SB and MI strategies did not differ from the control silage at either storage length. At 30 d of storage, the CA200, CA600, CA1000, LI600, and LI1000 silages presented lower lactic acid concentrations than the control silage. At 60 d of storage, the CA600, CA1000, CI1000, LI200, LI600, and LI1000 silages presented lower lactic acid concentrations than the control silage. There was a decrease in lactic acid concentration as storage length was extended from 30 to 60 d in the CI1000, LI200, and LI600 silages, and an increase in the CI200, CI600, and CR200 silages. For NH₃-N concentration (Figure 2), there was only a strategy effect (P<0.01). The SB strategy reduced NH₃-N concentration relative to the control silage, whereas the MI strategy and all essential oil compound-based strategies did not differ from the control silage. There was an interaction effect between strategy and storage length (P<0.01) on acetic acid concentration (Table 4). The SB strategy did not differ from the control silage, whereas the MI strategy increased acetic acid concentration only at 30 d of storage. All essential oil compound-based strategies increased acetic acid concentration relative to the control silage at 30 d of storage, whereas at 60 d, only the CI200, LI200, LI1000, and CR200 silages still presented greater concentrations than the control silage. Extending storage length from 30 to 60 d increased acetic acid concentration in the control, SB, MI, CA200, CI200, LI1000, and CR200 silages. There was an interaction effect between strategy and storage length (P<0.01) on propionic acid concentration (Table 4). Both the SB and MI strategies did not differ from the control silage. At 30 d of storage, the CA600, CA1000, CI1000, and CR200 strategies increased propionic acid concentration relative to the control silage. At 60 d of storage, only the CA600 and CA1000 silages still presented greater concentrations than the control silage. Extending storage length from 30 to 60 d increased propionic acid concentration in the CA600 and CA1000 silages and decreased it in the CI1000 and LI200 silages. There was an interaction effect between strategy and storage length (P<0.01) on butyric acid concentration (Table 4). Both the SB and MI strategies did not differ from the control silage. An increase in butyric acid concentration was observed in the CA600 silage at 30 d of storage and in the CA600 and CA1000 silages at 60 d of storage. Extending storage length from 30 to 60 d increased butyric acid concentration in the CA600, CA1000, and CI200 silages. There was an effect of strategy (P<0.01) on ethanol concentration (Figure 3). The SB strategy decreased ethanol concentration, whereas the MI strategy did not differ from the control silage. Both the CA600 and CA1000 strategies increased ethanol concentration relative to the control silage, whereas a decrease was observed with the CI600, CI1000, LI200, and CR1000 strategies, although only the CI1000 strategy did not differ from the SB strategy. Extending storage length from 30 to 60 d increased ethanol concentration (P<0.01).

Table 3

Item1	LAB (log cfu/g)		Yeasts (log cfu/g)		Fungi (log cfu/g)		Clostridia (log cfu/g)
Storage length (days)	30	60	30	60	30	60	30	60
CT	6.89bcA	6.19abB	4.02abA	3.64abA	3.59	2.85	ND	1.92de
CA200	6.79cA	6.05bB	3.29bB	4.36abA	3.18	3.56	ND	1.80e
CA600	7.01abcA	5.98bcB	1.70cB	4.09abA	3.03	2.20	ND	2.03bcde
CA1000	6.86bcA	6.42aB	3.00bcA	<1.00cB	3.00	2.45	ND	2.82abc
CI200	7.16abA	5.46deB	3.99abA	4.19abA	2.60	1.23	ND	2.87ab
CI600	6.86bcA	4.99fB	4.41aA	4.62aA	1.88	2.21	ND	2.70abcd
CI1000	6.88bcA	4.52gB	<1.00dB	3.55bA	1.80	<1.00	ND	1.92de
LI200	7.16abA	5.27efB	3.97abA	4.09abA	2.83	2.20	ND	2.20bcde
LI600	7.10abcA	5.47deB	3.90abB	4.48abA	2.27	1.23	ND	2.18bcde
LI1000	7.07abcA	5.68cdB	4.42aA	4.64aA	2.80	1.94	ND	3.21a
CR200	7.13abA	5.40deB	3.36bB	4.21abA	2.66	1.83	ND	1.96cde
CR600	7.00abcA	5.27efB	3.62abB	4.46abA	2.43	1.85	ND	2.01bcde
CR1000	7.09abcA	5.35eB	4.33aA	4.29abA	2.51	<1.00	ND	2.73abcd
SB	6.09dA	<3.00hB	<1.00dA	<1.00cA	1.55	<1.00	ND	1.76e
MI	7.28Aa	5.42deB	3.32bA	3.58abA	2.45	2.41	ND	1.93de
SEM	0.17		0.15		0.11		0.08
P: Strategy	<0.01		<0.01		<0.01		<0.01
P: Storage length	<0.01		<0.01		<0.01		-
P: Strategy × Storage length	<0.01		<0.01		0.61		-

LAB - lactic acid bacteria; ND - not determined; SEM - standard error of the mean.

Figure 1 CT - control; carvacrol at 200 (CA200), 600 (CA600) and 1,000 (CA1000) mg/kg FM; cinnamaldehyde at 200 (CI200), 600 (CI600) and 1,000 (CI1000) mg/kg FM; D-limonene at 200 (LI200), 600 (LI600) and 1,000 (LI1000) mg/kg FM; citral at 200 (CR200), 600 (CR600) and 1,000 (CR1000) mg/kg FM; SB - sodium benzoate at 2 g/kg FM; MI - microbial inoculant. Effect of strategy, P<0.01; effect of storage length, P<0.01; interaction strategy × storage length, P = 0.61. SEM = 0.11. Strategies with different letters differ by Tukey test (P<0.05).

Table 4

Item1	NH3-N (g/kg DM)		Lactic acid (g/kg DM)		Acetic acid (g/kg DM)		Propionic acid (mg/kg DM)		Butyric acid (mg/kg DM)			Ethanol (g/kg DM)
Storage length (days)	30	60	30	60	30	60	30	60	30	60		30	60
CT	0.6	0.61	23.5abcA	23.8abA	4.38eB	6.20defA	10.5dA	13.8bA	4.00bA	8.50cA		5.85	6.75
CA200	0.61	0.53	19.8deA	20.7bcdA	5.65dB	6.75abcdA	19.3cdA	16.5bA	18.8bA	36.3cA		6.45	6.98
CA600	0.61	0.61	19.9deA	18.7deA	6.25cdA	6.23cdefA	34.3abB	70.0aA	131aB	643bA		7.35	8.43
CA1000	0.71	0.58	18.3eA	19.2cdeA	5.60dA	5.47efA	33.8abB	58.0aA	119abB	4345aA		6.90	8.88
CI200	0.59	0.54	19.9cdeB	22.6abcA	6.48cdB	7.20abcA	19.3cdA	20.5bA	3.50bB	68.8cA		5.40	5.83
CI600	0.58	0.59	20.3bcdeB	23.4abA	6.73abcA	6.40bcdeA	23.5bcdA	16.5bA	3.25bA	5.00cA		3.30	3.77
CI1000	0.60	0.58	21.3abcdeA	18.9deB	7.60aA	7.18abcdA	45.5aA	21.8bB	3.00bA	3.00cA		1.65	1.98
LI200	0.61	0.61	22.9abcdA	17.0efB	7.48abA	7.63aA	23.0bcdA	15.0bB	6.33bA	3.00cA		5.20	4.93
LI600	0.59	0.56	19.6deA	14.9fB	6.68abcA	6.98abcdA	24.0bcdA	18.8bA	2.75bA	8.50cA		5.58	7.10
LI1000	0.50	0.59	19.6deA	19.2cdeA	6.70abcB	7.28abA	22.8bcdA	17.8bA	2.50bA	5.25cA		5.38	6.25
CR200	0.65	0.61	22.3abcdB	25.0aA	6.58bcdB	7.30abA	26.5bcA	19.8bA	2.75bA	2.33cA		5.20	6.40
CR600	0.59	0.60	24.6aA	24.1abA	6.68abcA	6.95abcdA	23.8bcdA	18.0bA	3.50bA	2.67cA		5.03	5.70
CR1000	0.55	0.66	23.8abA	24.5aA	6.45cdA	6.85abcdA	20.8bcdA	19.8bA	3.75bA	2.25cA		4.50	5.53
SB	0.46	0.44	23.2abcdA	23.4abA	4.13eB	5.35fA	16.5cdA	17.3bA	2.33bA	3.00cA		1.40	1.70
MI	0.57	0.57	23.6abA	23.1abA	6.15cdB	6.95abcdA	13.5cdA	17.5bA	1.50bA	2.50cA		5.53	6.45
SEM	0.01		0.25		0.08		1.11		55.7			0.18
P: Strategy	<0.01		<0.01		<0.01		<0.01		<0.01			<0.01
P: Storage length	0.41		0.32		<0.01		0.81		<0.01			<0.01
P: Strategy × Storage length	0.10		<0.01		<0.01		<0.01		<0.01			0.19

NH₃-N - ammonia nitrogen; DM - dry matter; SEM - standard error of the mean.

Figure 2

Figure 3 CT - control; carvacrol at 200 (CA200), 600 (CA600) and 1,000 (CA1000) mg/kg FM; cinnamaldehyde at 200 (CI200), 600 (CI600) and 1,000 (CI1000) mg/kg FM; D-limonene at 200 (LI200), 600 (LI600) and 1,000 (LI1000) mg/kg FM; citral at 200 (CR200), 600 (CR600) and 1,000 (CR1000) mg/kg FM; SB - sodium benzoate at 2 g/kg FM; MI - microbial inoculant. Effect of strategy, P<0.01; effect of storage length, P<0.01; interaction strategy × storage length, P = 0.19. SEM = 0.18. Strategies with different letters differ by Tukey test (P<0.05).

There was an interaction effect between strategy and storage length (P<0.01) on aerobic stability (Table 5). The SB strategy served as the positive reference, achieving 240 h at both storage lengths, whereas the MI strategy did not differ from the control silage. At 30 d of storage, the CA1000, CI1000, and LI600 strategies also enhanced aerobic stability, although the improvement was still lower than that achieved by the SB strategy. At 60 d of storage, the CA600, CA1000, and CI1000 strategies enhanced aerobic stability, and only the CA1000 strategy achieved 240 h, not differing from the SB strategy. Extending storage length from 30 to 60 d increased aerobic stability in the CA600 and CA1000 silages and decreased it in the CI1000, LI600, CR200, CR600, and MI silages. There was an interaction effect between strategy and storage length (P<0.01) on the maximum temperature (Tmax) achieved during aerobic exposure (Table 5). The SB strategy achieved a lower Tmax at both storage lengths, whereas the MI strategy did not differ from the control silage. At 30 d of storage, all essential oil compound-based strategies did not differ from the control silage. At 60 d of storage, the CA600 and CA1000 strategies achieved lower Tmax values than the control silage, although only the CA1000 strategy did not differ from the SB strategy. When storage length was extended from 30 to 60 d, Tmax decreased in the CA600, CA1000, CI200, and LI200 silages and increased in the LI600 and CR200 silages. There was an effect of strategy (P<0.01) on the time required to achieve Tmax during aerobic exposure (Figure 4), with only the CA1000 and CI1000 silages requiring more time to achieve their respective Tmax than the control silage. There was an interaction effect between strategy and storage length (P<0.01) on pH variation (ΔpH) during aerobic exposure (Table 5). The SB strategy decreased ΔpH at both storage lengths, whereas the MI strategy did not differ from the control silage. At 30 d of storage, the CI1000 strategy reduced ΔpH relative to the control silage, although it remained higher than that of the SB silage. At 60 d of storage, the CA600 and CA1000 strategies also reduced ΔpH relative to the control silage, and only the CA1000 strategy did not differ from the SB strategy. Extending storage length from 30 to 60 d reduced ΔpH in the CA600 and CA1000 silages and increased ΔpH in the CI600 and CI1000 silages. There was an effect of strategy (P<0.01) on post-opening DM losses during aerobic exposure (Table 5). The SB strategy decreased post-opening DM losses, whereas the MI strategy did not differ from the control silage. The CA600, CA1000, and CI1000 strategies also reduced post-opening DM losses, although only the CA600 and CA1000 strategies did not differ from the SB strategy.

Table 5

Item1	Aerobic stability (h)			Tmax (°C)		Time to Tmax (h)		ΔpH		DM losses (g/kg)
Storage length (days)	30	60		30	60	30	60	30	60	30	60
CT	67defgA	61deA		50.2aA	50.4abA	148	130	2.73abA	2.75aA	ND	189a
CA200	63efgA	64deA		49.0aA	49.8abA	126	120	2.29abA	2.69abA	ND	189a
CA600	99bcdefB	140bA		47.9aA	44.7cB	175	229	2.24abA	1.58bB	ND	91.6bc
CA1000	117bcB	240aA		46.7aA	23.9dB	211	202	2.96abA	0.01cB	ND	46.7c
CI200	84cdefgA	69deA		50.1aA	47.3bcB	184	194	2.69abA	2.30abA	ND	158a
CI600	59fgA	71deA		48.8aA	50.2abA	181	186	1.91bcB	2.54abA	ND	179a
CI1000	141bA	114bcB		47.1aA	49.4abA	236	204	1.07cB	2.15abA	ND	110b
LI200	55gA	49eA		51.4aA	48.5abcB	125	140	2.34abA	2.69abA	ND	187a
LI600	133bA	96cdB		47.4aB	51.5abA	180	137	2.34abA	2.83aA	ND	189a
LI1000	94bcdefgA	75cdeA		48.1aA	50.4abA	166	160	2.86abA	2.82aA	ND	179a
CR200	97bcdefA	69deB		48.3aB	51.8aA	187	133	1.90bcA	2.87aA	ND	194a
CR600	107bcdA	63deB		46.4aA	49.3abA	177	151	3.29aA	2.71abA	ND	175a
CR1000	69defgA	62deA		46.8aA	48.9abA	172	151	2.07bcA	2.54abA	ND	179a
SB	240aA	240aA		24.5bA	24.3dA	192	187	−0.04dA	0.02cA	ND	53.2c
MI	99bcdeA	67deB		48.2aA	50.7abA	190	143	3.17aA	2.87aA	ND	170a
SEM	6.53			0.73		4.04		0.09		6.77
P: Strategy	<0.01			<0.01		<0.01		<0.01		<0.01
P: Storage length	0.32			0.05		0.07		0.71		-
P: Strategy × Storage length	<0.01			<0.01		0.30		<0.01		-

Tmax - maximum temperature; ΔpH - pH variation after 10 days of aerobic exposure; DM - dry matter; ND - not determined; SEM - standard error of the mean.

Figure 4 CT - control; carvacrol at 200 (CA200), 600 (CA600) and 1,000 (CA1000) mg/kg FM; cinnamaldehyde at 200 (CI200), 600 (CI600) and 1,000 (CI1000) mg/kg FM; D-limonene at 200 (LI200), 600 (LI600) and 1,000 (LI1000) mg/kg FM; citral at 200 (CR200), 600 (CR600) and 1,000 (CR1000) mg/kg FM; SB - sodium benzoate at 2 g/kg FM; MI - microbial inoculant. Effect of strategy, P<0.01; effect of storage length, P = 0.07; interaction strategy × storage length, P = 0.30. SEM = 4.0. Strategies with different letters differ by Tukey test (P<0.05).

The treatment with sodium benzoate at 2 g/kg FM improved the quality of conservation by decreasing DM losses and suppressing yeast and mold populations, which also markedly enhanced the aerobic stability of RCGS. Although the lower pH in the SB treatment was an unexpected outcome (Morais et al., 2017), it may be explained by the decreased NH₃-N content and the unaffected lactic acid production. The treatment with a microbial inoculant combining L. buchneri, L. plantarum and P. acidipropionici did not achieve better outcomes for conservation and aerobic stability in RCGS within the storage lengths evaluated. Longer storage lenghts may be required for the practical effects of the inoculant treatment to become apparent. Thus, the SB strategy was chosen as the positive control reference treatment for discussing the results for the essential oil compound-based strategies evaluated in this study.

Carvacrol is one of the most active monoterpenes in terms of antimicrobial activity, exhibiting a broad spectrum of action against both Gram-positive and Gram-negative bacteria, as well as yeasts and fungi (Dorman and Deans, 2000; Chao et al., 2000). Carvacrol did not inhibit LAB, but it decreased lactic acid production especially at the higher application rates (600 and 1,000 mg/kg FM) and compromised the acidification of RCGS, with the CA600 and CA1000 silages presenting the highest pH values in our study. Foskolos et al. (2016) reported similar results for ryegrass silage treated with carvacrol at application rates varying from 50 to 2,000 mg/kg FM, including decreased lactic acid content and greater pH values without any change in LAB counts. The impaired acidification probably favored undesirable fermentations in RCGS, as greater DM losses were observed, along with evidence of enhanced clostridial development occurred in the CA1000 silage, which also presented an exceptionally high butyric acid concentration. Aerobic stability was enhanced by CA600 and CA1000 strategies and by extending the storage length to 60 d. Only the CA1000 strategy was able to emulate the effects the SB strategy, managing to keep aerobic stability throughout the whole 240-h trial, decreasing DM losses in aerobic exposure and keeping pH practically unchanged, suggesting minimum or no degradation of organic acids by yeasts and molds. This outcome may be explained by greater concentrations of acetic, propionic and butyric acids promoted by the CA600 and CA1000 strategies, which are related to antifungal properties in silage during aerobic exposure (Daniel, 2019). An inhibitory effect of carvacrol itself on yeast and fungal populations would also be expected (Chao et al., 2000). However, fungi counts were not affected by carvacrol strategies, and despite microbial counts at 60 d suggest reduced persistence of yeasts on CA1000 silage, both CA600 and CA1000 silages exhibited the highest contents of ethanol. It is possible that carvacrol is more effective in reducing the persistence of viable yeast cells than in inhibiting their activity at the initial stages of silage fermentation, a hypothesis that should be investigated in future studies. Overall, it seems that increasing application rates of carvacrol in RCGS may replicate the enhanced aerobic stability promoted by the SB strategy, but at the expense of greater DM losses if the grains are contaminated with undesirable organisms, such as clostridia bacteria. Based these results, carvacrol cannot be safely recommended as an additive for RCGS.

Cinnamaldehyde is a highly active phenylpropanoid against a broad spectrum of bacteria and fungi (Chao et al., 2000). There was no change in LAB counts at 30 d of storage, but a marked suppression of LAB persistence occurred at 60 d, especially in the CI600 and CI1000 silages. Despite the absence of differences in lactic acid and NH₃-N concentrations relative to the control silage, all cinnamaldehyde-based strategies promoted lower pH values in RCGS at both storage lengths. This greater acidification probably enhanced silage conservation quality, as the CI600 and CI1000 strategies were able to emulate the reduced DM losses observed with the SB strategy. Although the CI200 strategy favored clostridial development, increasing application rates prevented this effect. Only the CI1000 strategy was able to enhance the aerobic stability of RCGS, although the improvement was still lower than that observed in the SB silage. The CI1000 strategy was effective in suppressing yeasts and was the only strategy, besides SB, capable of suppressing fungi. Both the CI600 and CI1000 strategies were able to decrease ethanol concentration, although only the CI1000 strategy matched the response obtained with the SB strategy. Interestingly, extending the storage length from 30 to 60 d reduced the positive effect of the CI1000 strategy on aerobic stability, as well as its efficiency in suppressing yeast persistence, and this was accompanied by a decrease in lactic acid concentration. Overall, increasing application rates of cinnamaldehyde may emulate the enhanced RCGS conservation quality promoted by the SB strategy and improve aerobic stability. However, storage length appears to be an important factor influencing the efficiency of cinnamaldehyde as an antimicrobial agent in grain silage. Further studies are needed to determine whether storage length would remain a limiting factor when higher application rates of this compound are used.

D-limonene is a monoterpene with a more restricted antimicrobial spectrum, being more active against Gram-negative bacteria (Dorman and Deans, 2000) and fungi (Chao et al., 2000). In a graduate report from Brazil, Silva (2021) observed that sweet orange essential oil (940 g/kg of limonene) applied at rates ranging from 250 to 750 mg/kg FM favored crude protein preservation and reduced DM losses in RCGS stored for 60 d, despite increasing silage pH to values as high as 4.29. In the present study, D-limonene strategies were associated with a consistent decrease in lactic acid concentration, increased acetic acid production, and no effect on NH₃-N concentration relative to the control silage. All D-limonene strategies, however, promoted lower silage pH values at 30 d of storage relative to the control silage, although this effect was no longer observed when storage length was extended to 60 d. Despite a decrease in ethanol concentration with the LI200 strategy, there was no evidence of yeast or fungal inhibition by any of the D-limonene strategies. As D-limonene did not consistently prevent DM losses or improve aerobic stability at application rates ranging from 200 to 1,000 mg/kg FM, this isolated compound was not considered a promising additive for RCGS in the present study. However, limonene-rich essential oils have been associated with favorable outcomes for conservation quality and aerobic stability, not only in RCGS, as reported by Silva (2021), but also in other silage matrices (Chaves et al., 2012; Andrade et al., 2024; Silva et al., 2024). This represents a promising scenario for investigating the synergistic effects of secondary compounds on the efficacy of essential oils as silage additives.

Citral is a mixture of two isomeric monoterpenes with an aldehyde functional group, geranial and neral, being active against a moderate spectrum of bacteria and fungi (Dorman and Deans, 2000). In our study, citral-based strategies decreased LAB counts at 60 d of storage, although no effect was observed on yeast or fungal counts. Despite having no effect on lactic acid and NH₃-N yields, RCGS treated with CR600 and CR1000 strategies presented lower pH values than the control silage at 30 d of storage, which could be related greater acetic acid production. As none of the citral-based strategies led to any improvement in silage quality or aerobic stability, citral also did not prove to be a promising compound as an additive to RCGS in our study.

Carvacrol compromised RCGS acidification and increased storage losses, favoring increased clostridial activity. Cinnamaldehyde reduced storage losses in RCGS at application rates of at least 600 mg/kg FM, whereas the application rate of 1,000 mg/kg FM further reduced storage losses to levels comparable to those achieved with SB standard and enhanced aerobic stability. The extended storage length from 30 to 60 d, however, seemed to compromise the efficiency of cinnamaldehyde as an antimicrobial agent. D-limonene and citral did not consistently promote desirable outcomes for RCGS within the application rates and storage lengths evaluated in this study.