INTRODUCTION

There is a trend in global demand for garlic consumption, which has generated employment in food, particularly ruminants, of its main by-products such as husks, leaves and straws (^{Kallel and Ellouz, 2017}). In various countries, the garlic leaves obtained during harvest are incinerated, and consequently there are problems associated with disposal costs and environmental contamination (^{Han et al., 2013}). In this sense, it can be said that the methane emissions by ruminants only cause serious environmental problems, but also represent an important source of energy loss for animals (^{Gallegos et al., 2019}; ^{Lan and Yang, 2019}). The production of greenhouse gases (GHG) in different livestock systems and their impact on climate changes are a major concern throughout the world. Enteric methane has been reported as the most GHG emitted (50-60%), in the production systems in ruminants (^{Tapio et al., 2017}; ^{Haque, 2018}).

Because of this problem, many attempts have been to modify ruminal fermentation and decrease methane production made; among them is the incorporation into ruminant diets of good quality forages, such as alfalfa hay. Garlic and some of its components have also been shown to decrease methane production under in vivo and in vitro conditions (^{Kamra et al., 2012}). In fact, so far few studies have investigated the effects of garlic and its harvest by-products on ruminal fermentation and methane production.

Based on the above, it assumes that the substitution of garlic leaves for alfalfa hay in beef cattle diets produces similar changes in the rumen fermentation patterns and methane production during in vitro fermentation. Therefore, the objective of the present work was to evaluate the substitution of alfalfa hay for garlic leaves in the production of in vitro gas and methane, as well as in the ruminal fermentation patterns.

MATERIAL AND METHODS

The experiment was in the metabolic unit and in the Animal Nutrition laboratory of the Faculty of Veterinary Medicine and Zootechnics of the Juárez University of Durango state (Mexico) carried out.

Raw garlic leaf sampling. The samples (25 cm long leaves) of raw garlic (Allium sativum) used in this study were collected in the northern region of Mexico. To ensure representative sampling, the samples were collected five times, between January and May 2019. Before chemical analysis and in vitro tests, the samples were dried and then ground through a 1 mm mesh.

Experimental treatments. In each experimental treatment, alfalfa hay and raw garlic leaf hay were used as forage sources. The chemical composition of raw alfalfa hay and garlic leaf hay is shown in Table 1.

Four treatments were evaluated: (T1) alfalfa hay (50% DM) + raw garlic leaves (0% DM); (T2) alfalfa hay (33% DM) + raw garlic leaves (17% DM); T3 alfalfa hay (17% DM) + raw garlic leaves (33% DM); T4 alfalfa hay (0% DM) + raw garlic leaves (50% DM). Similar proportions of ground corn, cottonseed and minerals were used in all treatments. The nutritional composition of the experimental treatments is shown in Table 2.

Nutritional composition. Each sample of the experimental treatments was analyzed in triplicate for dry matter (DM), crude protein (CP), ether extract (EE) and organic matter (OM) (^{AOAC, 2000}). Neutral detergent fiber (NDF), acid detergent fiber (ADF) and acid detergent lignin (ADL) analyzes were determined using the filter bag technique with a fiber analyzer (ANKOM Technology, Fairport, NY, USA). The non-fibrous carbohydrate content (NFC) was calculated using the difference between the total carbohydrates (TCH) and the % of NDF. The in vitro digestibility of dry matter (DIVDM) and organic matter (DIVOM) was determined using the Daisy incubator (ANKOM Technology, Fairport, NY, USA).

In vitro gas, ruminal fermentation patterns and methane production. In vitro gas production was measured using the ANKOM gas production system. Ruminal fluid was collected approximately 3 h after the morning feeding of two fistulated steers in the rumen, fed open access with alfalfa hay. The ruminal fluid was immediately through four layers of gauze filtered and it was transported to the laboratory in a thermos sealed. The resulting ruminal fluid was purged with CO2 before use as an inoculum. Approximately 1 g of dried and ground samples from each treatment were in triplicate weighed and placed in glass modules. Ruminal liquid was into each module pipetted and mixed with McDougall's saliva (120 mL).

Gas production was recorded after 2, 4, 6, 8, 12, 16, 24, 36, 48, 72 and 96 h of incubation. The gas production kinetics (PD) was adjusted to the non-linear model proposed by ^{Ørskov and McDonald (1979)}. It is the following: PD (t) = a+b × (1-exp-ct); where “a” is the gas production from the soluble fraction (mL), “b” is the gas production from the insoluble fraction, (mL), “c” is the constant rate of gas production ( mL/h) and "t" is the incubation time.

For the determination of methane, after 24 h of incubation, 10 mL of gas were collected from the upper space of the modules of the gas production system, from which 3 mL subsamples were taken and the methane production was then determined in a gas chromatograph (^{Kara, 2015}).

Likewise, after the end of the 24 h incubation and before filtration, two samples (5 mL) of the liquid from each module were collected, and the pH was immediately measured and stored at -20 °C until subsequent analysis of ammoniacal nitrogen (N-NH3) and volatile fatty acids (VFA) (^{Ammar et al., 2005}).

Statistical analysis. An analysis of variance was carried out for a completely randomized design to compare the parameters of in vitro gas production, methane and ruminal fermentation patterns with the MIXED procedure of SAS (2002). Individual means were separated by Tukey's multiple range test. The mean differences were considered significant at P <0.05.

RESULTS AND DISCUSSION

In vitro gas production parameters. The highest values of "a", "b" and "c" (table 3), were found in T4; while the lowest values in T1 (P <0.05); but no differences were observed between T2 and T3 (P> 0.05). Differences in the value "a + b" were observed between treatments (P <0.05). The results are consistent with those reported by ^{Tag El-Dini et al. (2012)}, who found an "a" value of 0.60% when garlic pulp was fermented under in vitro conditions at a level of 3% of the concentrate.

In this study, the differences observed between treatments in the value "a" can be attributed to the different concentrations of the treatments in non-fibrous carbohydrates (^{La O et al., 2008}). The average values of "b" and a+b found in the present study were 116.0 mL 200 mg-1 DM and 116.5 mL 200 mg-1 DM, respectively. These values coincide with those reported by ^{Sahli et al. (2018)} when garlic powder, in doses of 32 mg, were incubated in vitro in a diet composed of 50% ryegrass hay and 50% commercial concentrate.

The "c" value registered with T4 (9.2 mL h-1) was higher than that found by ^{Anassori et al. (2012)} (3.5 mL h-1), who evaluated the in vitro gas production kinetics of raw garlic bulb. In the current study, the high constant rate of gas production "c" registered in T4 indicates a high availability of nutrients for ruminal microorganisms; while the lower "c" values observed in T1 could be the result of a high content of NDF, whose chemical components could decrease the fermentation rate of the substrate (^{Fievez et al., 2005}).

Ruminal fermentation patterns and methane production. Ruminal fermentation patterns and methane production are in Table 4 presented. The concentration of N-NH3 ranged from 15.1 to 18.1 mg/dL. The highest concentration of N-NH3 was recorded in T4 and the lowest in T1 (P <0.05). There were no differences between T2, T3 and T4 in total VFA concentrations (P> 0.05); but both treatments were different at T1 (P <0.05).

The highest acetate concentration was recorded in T1 and the lowest in T4 (P <0.05), while the highest concentration of propionate was registered in T4 and the lowest in T1 (P <0.05). There was no difference between T1, T2 and T3 in the acetate: propionate ratio (P> 0.05); but both treatments were different at T4 (P <0.05). The highest methane production was recorded in T1 and the lowest in T4 (P <0.05).

The ruminal concentrations of N-NH3 registered in all the evaluated treatments were kept within the suggested range for optimal microbial growth. According to ^{Wanapat and Pimpa (1999)}, a range of 15 to 30 mg/100 mL is the minimum concentrations required for optimal microbial protein synthesis.

In contrast, with the results obtained in this study, several in vitro and in vivo studies report that garlic oil and garlic powder reduce or have no effect on the concentration of N-NH3 in ruminal fluid (Cardozo et al., 2004). However, ^{Yang et al. (2007}) found an increase in the concentration of ruminal N-NH3, when raw garlic was incorporated into the diets of lactating cows. In this study, the higher concentration of N-NH3 obtained with T4 could be explained by the high content of crude protein supplied by garlic leaves (^{Panthee et al., 2017}).

The total VFA concentrations obtained with T2, T3 and T4 are consistent with other in vitro studies, where they did not differ with the addition of garlic oils (^{Klevenhusen et al., 2011}). The propionate concentration observed in T4 was higher than in the other treatments; this is evidenced by the decrease in the acetate: propionate ratio.

Furthermore, the increase in propionate concentration could be attributed to the amount of non-fibrous carbohydrates supplied by T4 (^{Van Soest, 1994}). Similarly, the reduction in the ratio of acetate to propionate registered in T4 indicates an improvement in the net energy contribution (^{Zhong et al., 2019}). Similar results are reported by ^{Mirzaei- Aghsaghali and Maheri-Sis (2011)} who found that garlic oil increased the proportions of propionate and butyrate, but reduced the proportion of acetate. In this study, the low concentration of methane was obtained with T4 (high in garlic leaves); and coincides with the results found by ^{Kongmun et al., (2010}), who report a decrease in methane production when garlic powder was evaluated in ruminant diets. In our study, the methane values recorded in the four treatments were lower than those reported by ^{Zafarian and Manafi(2013)}, who evaluated in vitro conditions, the addition of different doses of garlic powder (2, 4 and 6% DM) in diets with 50% fiber and 50% concentrate.

In this study, the reduction in methane production registered in T4 in comparison with the other treatments could be attributed to the increase in the concentration of propionate, because the formation of propionate consumes reducing equivalents; while the formation of acetate generates H2 for methanogenesis (^{Moss et al., 2000}). Any component or variable in the diet that causes a change in favor of propionate production will be by a reduction in methane production per unit of fermented diet accompanied while the opposite is observed for acetate and butyrate (^{Pinares-Patiño, 2003}). However, the differences observed in this study in methane production, compared to those reported in the literature, may be related to different genetic varieties and possibly to the phenological state of the garlic plant; which results in different contents of the cell wall and as a consequence in the ruminal fermentation patterns (^{Lee et al., 2017}).

CONCLUSIONS

The substitution of alfalfa hay for garlic leaves in beef cattle diets improved some indicators of nutritional importance, such as the production of gas from the soluble fraction and the constant rate of gas production. Similarly, it increased the concentrations of N- NH3, total VFA, and reduced the production of methane. Further research is needed to evaluate the effects of garlic leaves in ruminant diets on ruminal fermentation patterns and methane production under in vivo conditions; as well as to evaluate the sustainability of garlic leaf supplementation to mitigate rumen methanogenesis without harmful effects on animal performance.