Six heifers (Bos indicus × Bos taurus) fitted with a permanent rumen cannula (10 cm i.d., Bar Diamond Inc., Parma, ID, USA) were adapted to the three different diets (Table 1) for 15 days, before rumen liquor was collected. The experiment was arranged as a completely randomized design with three treatments: (1) Control (CTL, no additive); (2) OEO1 at 0.25% incubated DM (IDM); and (3) OEO2 at 0.50% IDM. The substrate of each experimental treatment (1 g DM) was incubated at 39°C in ANKOM glass incubators (ANKOM Technology, USA) with 120 mL of ruminal liquor pooled from each pair of heifers and buffer solution in a 2:1 ratio. The modules were opened after 24 h of incubation, and the pH of the incubation fluid was measured immediately and evaluated using a portable pH meter (Hannah^® Instruments, Woonsocket, RI, USA). Aliquots of the incubation fluid (10 mL) were collected and mixed with 2.5 mL of metaphosphoric acid [25% weight per volume (w/v)] following a previously described method (20). The volatile fatty acid (VFA) content of the samples was quantified using a gas chromatography system (7820A GC system Agilent Technologies Inc, Santa Clara, CA, USA) equipped with a flame ionization detector. The samples were analyzed in triplicates.

The in vitro incubations were carried out using the ANKOM RF Gas Production System (ANKOM Technology, Macedon, NY, USA). Samples (ca. 1 g DM) from each treatment (CTL, OEO1, and OEO2) were incubated in triplicates in ANKOM glass modules equipped with a pressure transducer. This batch culture procedure was repeated in two consecutive runs on two different days (n = 6 per treatment). Fermentations were carried out according to the manufacturer's instructions by incubating the sample with a 2:1 mixture of buffer and ruminal inoculum solution. Four cows (Bos indicus × Bos taurus) fitted with a permanent rumen cannula (10 cm i.d., Bar Diamond Inc., Parma, ID, USA) were adapted to the diet for 15 days (70:30 forage:concentrate ratio), and rumen liquor was withdrawn from these cows in the morning before they were fed (between 0800 and 0830 h). The ruminal inoculum was collected as a pooled sample and immediately transported in thermos bottles to the laboratory, where it was mixed, flushed with CO₂, and filtered with four layers of cheesecloth (21). Buffer solutions were prepared following the instructions provided by Ankom (ANKOM Technology, Macedon, NY, USA). The pressure was measured every hour during the 96 h incubation period. The kinetics of in vitro gas production were determined by fitting data to the Gompertz function (22) using the following equation:

where GP = gas production at time t (mL); G_max= maximum gas production (mL); k = constant gas production rate (h⁻¹); and A = latency time before gas production begins (h). To measure the proportions of CH₄ and CO₂, the pressure relief valve of the modules was opened for 2 s as per the manufacturer's instructions. The expelled gas was sent through a tube to a portable gas analyzer (GEMTM5000, LANDTEC, USA) as per a previously described method (23).

Six crossbred heifers (Bos taurus × Bos indicus) with average body weight (BW) of 383 ± 21.6 kg (mean ± SD) were used in this experiment. The heifers were randomly assigned to three treatments (no additive [CTL], 0.25% OEO [OEO1], and 0.5% OEO [OEO2]) in a replicated 3 × 3 (n = 6) Latin square design. Each period lasted for 21 days. The acclimation period was between 1 and 14 days. Measurements were taken between day 15 and day 21. The heifers were acclimated to indirect open-circuit respiration chambers for CH₄ measurements before the beginning of the experiment.

Heifers were provided unlimited access to water and fed the TMR (Table 1) at 2.8% of their body weight. Heifers were fed once daily at 0800 h. The TMR was formulated to satisfy the energy and protein requirements for maintenance and growth (25). The difference between the amount of feed supplied and that rejected the next day was used to compute daily feed intake. Between day 17 and day 21 of each period, the total amount of feces was collected in metabolic crates to determine the apparent total tract digestibility of nutrients. A sub-sample of 10% of the total feces was taken after homogenization. The fecal samples were dried in a forced-air oven (55°C for 72 h) and ground through a 1-mm screen (Willey mill, Arthur H. Thomas Co., Philadelphia, PA, USA) before chemical analysis. To avoid contamination of the feces with urine, each of the metabolic cages had a metal grid that allows urine and feces to pass through, but the heifers stayed dry and comfortable. Below the first grid, two steel mesh grills with 11.5-mm diameter holes collected feces and permitted the passage of urine into a container.

Five hours after feeding on day 18 of each period, 8 mL of ruminal liquor was collected using a syringe connected to a stainless-steel tube through the rumen cannula (20-mm internal diameter), to which 2 mL of 25% of metaphosphoric acid was added. The samples were then stored at −20°C before the determination of VFA content by gas chromatography (7820A GC system Agilent Technologies Inc, Santa Clara, CA, USA). An additional sample (50 mL) of rumen liquor was taken to measure the pH using a portable pH meter (Hannah^® Instruments, Woonsocket, RI, USA).

Methane measurements were carried out in two open-circuit respiration chambers with an internal volume of 9.97 m³. Each chamber was equipped with a steel feeder (90 cm length, 70 cm height, and 50 cm width), automatic bowl-type waterer (20 cm diameter), fan (30 cm diameter), air conditioner (12,000-BTU), and dehumidifier. The temperature and relative humidity were maintained at 23 ± 1°C and 55 ± 10%, respectively. A polyvinyl chloride tube and a half-turn valve (75-mm diameter) positioned in the upper-right corner of the front wall regulated the supply of ambient air and air pressure within the chamber (−275 Pa). Ambient air was drawn from the chambers by mass flowmeters (Flowkit; Sable Systems International; Las Vegas, NV, USA), adjusted to the animal's body weight (1 L min/kg BW), and sent to the sample device using a flexible industrial pipe (50-mm diameter). A CH₄ analyzer that uses infrared light (MA-10 Sable Systems International^®, Las Vegas, NV, USA) and respiration chambers were calibrated by infusing a known amount of pure CH₄ (Praxair^® Gases Industrial Inc., Monterrey, NL, Mexico; 99.997% purity), with recovery values ranging from 97 to 102%. Every morning before each run, the calibration was verified by zeroing the apparatus with pure N₂ (Praxair^®, State of Mexico, Mexico) and then releasing CH₄ (1000 ppm, Praxair^®, State of Mexico, Mexico) diluted in N₂ until stable readings (plateau) were obtained after 5 min. The measurements were carried out for three consecutive days (day 19 to day 21 of each period), and chambers were cleaned in 1 h. Expe Data^® software was used to extrapolate the data to a 24 h period (26). According to IPCC (27), GEI/day lost as CH₄ was computed from its heat of combustion (CH₄ = 55.65 MJ/kg).

The total feed provided, refused feed, and fecal samples were oven-dried for 72 h at 60°C, ground through a 1-mm screen (Wiley mill; A. H. Thomas, Philadelphia, PA, USA), and analyzed for DM (method 930.15), ash (method 923.03), and crude protein (CP) using a C/N-analyzer (CN-2000 series 3740, LECO^® Corp., St. Joseph, MI, USA; AOAC, method number 992.15), as well as ether extract (EE, AOAC, method number 920.39). An adiabatic bomb calorimeter was used to determine gross energy (model 6400 Parr Instrument Company, Moline, IL, USA). The concentrations of NDF and acid detergent fiber (ADF) (AN 3805 ANKOM, ANKOM Technology, Wayne County, NY, USA) were determined following the method proposed by Van Soest et al. (28). Briefly, NDF and ADF were determined using ~0.5-g samples with fiber analysis bags (F57). Samples were treated with a neutral detergent solution and rinsed with a heat-stable amylase to remove the soluble non-structural carbohydrates from the matrix; the remaining compounds were cell wall compounds, such as cellulose, hemicellulose, and lignin. Thereafter, hemicellulose was removed from the matrix by making it soluble with an acid detergent solution. Cellulose and lignin remaining in the matrix were treated with sulfuric acid to remove cellulose using an ANKOM²⁰⁰ fiber analysis system and commercially available solutions (ANKOM Technology, Wayne County, NY, USA).

The inclusion of OEO had no effect (P > 0.05) on in vitro pH, which ranged between 6.7 and 6.8. The proportions of acetate, propionate, isobutyrate, butyrate, isovalerate, and valerate were not altered by OEO supplementation (P > 0.05); the acetate:propionate ratio ranged from 1.9 to 2.0 (Table 3) and was also not affected by OEO supplementation.

No differences in G_max (mL/g IDM) and GP24 (mL/g IDM) between the OEO treatments and the control were observed (P > 0.05; Table 4). However, A (h) and K (% h⁻¹) were higher in the OEO1 and OEO2 treatments compared with the control (P < 0.05). The production of CH₄ decreased by 41.7% (mL at 24 h), 41.6% (mL/g IDM), and 45.9% (mL/g degraded DM), respectively, when OEO was supplemented at 0.25% DM (P < 0.05). There was a quadratic relationship between the dose of OEO and CH₄ production (ml/g IDM and ml/g of degraded DM; P < 0.05).

Supplementation of OEO had no effect on DM, OM, CP, NDF, and ADF intake (P > 0.05; Table 6); however, there was a negative quadratic relationship between GE intake and the level of OEO supplementation (P < 0.05; Table 6). The apparent total tract digestibility of DM decreased quadratically (P < 0.05; Table 7) when 0.5% OEO was added. The apparent total tract digestibility of OM, CP, NDF, and ADF and digestible energy were not affected (P > 0.05) by OEO supplementation.

Rumen pH, VFA, and the acetate:propionate ratio are shown in Table 8. Rumen pH was similar across treatments (P > 0.05) and ranged from 6.57 to 6.68. There was a quadratic response in acetate proportion when OEO was added (P < 0.05). The proportion of valerate decreased quadratically in response to OEO addition (P < 0.05). Feeding OEO1 resulted in an increase of valerate by 21% relative to the control, whereas valerate decreased by 14% in response to OEO2 compared to the control (P < 0.05). Supplementation of OEO had no effect on the C3:C2 ratio (P > 0.05).

When 0.5% OEO was fed, CH₄ (g/day) was reduced by 12% as compared to control (P < 0.05; Table 9). The CH₄ yield (g/kg DMI) decreased linearly with the level of OEO supplementation. No changes were observed in emissions per kilogram of fermented OM in the rumen (P > 0.05). There was a linear decrease in CH₄ yield (% GE/day) in response to OEO addition. There was a quadratic decrease in energy loss in form of CH₄ (MJ of GEI/day) in response to OEO2 supplementation (P < 0.05).

Changes in rumen pH in our study were consistent with those observed by Castillejos et al. (35), who used the most abundant monocyclic monoterpene in OEO (limonene) in doses of 5–500 mg/L rumen liquor in a basal diet with 60:40 forage:concentrate ratio. However, our findings differ from those of Kamalak et al. (16), who found that increases in OEO supplementation (200, 400, 800, and 1,200 mg/L) increase rumen pH. In our experiment, VFA (% of total) was not significantly affected by OEO supplementation, suggesting that the doses evaluated did not affect rumen bacteria (OEO1: 2,500 mg/L and OEO2: 5,000 mg/L). Kamalak et al. (16), who used rumen fluid from sheep, reported that high doses of OEO (800 and 1,200 mg/L) reduce the production of VFA (mmol/L) and alter the molar proportion of VFA; specifically, high OEO doses increase the proportion of acetate and the acetate:propionate ratio. Castillejos et al. (36) found that EO supplementation of a diet containing 10% barley straw and 90% concentrate with thyme oil (Thymus vulgaris), savory oil (Satureja montana), and Lavandin oil (Lavandula hybrida) lowered the pH. Furthermore, they found that clover leaf oil (Eugenia caryophyllus) and oregano oil (Origanum vulgare) increased pH when supplied at 500 mg.

Essential oils, which are typically obtained by steam distillation, are known to have antibacterial properties due to their ability to modify the permeability of cells (37, 38). A previous study (16) showed that OEO supplementation at doses ranging from 100 to 1,200 mg/L reduces CH₄ production (mmol/L), which is consistent with the results of our study. Methane production was reduced by 41.7% (mL), 41.6% (mL/g incubated DM), and 45.9% (mL/g degraded DM) in response to OEO1. García-Rodríguez et al. (18) evaluated the effect of substituting extruded corn for dry citrus pulp (20%) and found that reductions in CH₄ were associated with the antimicrobial effects of bioactive compounds such as terpenoids, limonene, and citral. However, this was not observed for the OEO2 treatment, as there were no differences in CH₄ between OEO2 and the control. The quadratic response in CH₄ production might stem from the fact that a higher concentration of OEO might have induced toxicity to the microbiota. This finding confirms that the effects of EO on ruminal microbiota are dose-dependent. An alternative explanation for the differences between the results of our study and the results of previous studies might be associated with differences in the concentrations of bioactive compounds, the substrates used (i.e., soybean meal vs. forage: concentrate:forage ratio 70:30), and the techniques used for quantification.

Increases in the level of OEO supplementation did not affect DMI as has been shown in a previous study that applied 200–1,200 ppm of OEO in a swamp grass:concentrate ratio 60:40 (41). A different study found that supplementation of 4.5 g of citrus extract per day did not improve the DMI of the dairy cows (42). Other EOs have been reported to increase feed intake (carvacrol, eugenol, thymol, coriander seed oil, geranyl acetate, and geraniol), but some studies have demonstrated that blends of thymol, eugenol, carvacrol, garlic, citral, and cinnamaldehyde can induce reductions in feed intake and digestibility (43, 44). This suggests that the effects of different EOs or their constituents on feed intake are variable and will depend on the type and dosage of EO (37). Other studies have suggested that the palatability of feed can be impacted by the addition of EO; however, other factors that also affect DMI—such as animal growth stage, body weight, or the specific properties of diets such as the fiber content and particle size—also require consideration (41, 45, 46). One possible explanation for the reduced rumen degradability associated with the OEO2 treatment might be the broad and non-specific antimicrobial capacity of OEO on rumen microorganisms. The use of 0.5% OEO reduced rumen microbial activity, which affected the fermentability, degradability, and CH₄ emissions (47). This is consistent with findings of in vitro studies showing that 200 and 300 ppm of orange peel oil inhibits CH₄ generation by more than 50% due to the low digestibility of DM and NDF, which was possibly associated with a microbial imbalance in the rumen (48). The discrepancies among studies might be caused by different types of EOs that were used in the respective studies, dosages tested, and differences in diet and host interactions (49).