Dietary Supplementation of Phytoncide and Soybean Oil Increases Milk Conjugated Linoleic Acid and Depresses Methane Emissions in Lactating Holstein Cows

The objective of this study was to determine whether adding phytoncide oil (PO) and soybean oil (SBO) to the dairy cow diet could increase milk conjugated linoleic acid (CLA) and depress methane (CH 4 ) emissions in lactating Holstein cows. Rumen fermentation was conducted at four levels of SBO (0, 1, 2, and 4%, on DM basis) and two levels of PO (0 and 0.1%, on DM basis) with in vitro experiment. To evaluate blood parameters, fecal microbe population, milk yield fatty acid compositions, and methane production, in vivo experiment was conducted using 38 Holstein dairy cows divided into two groups of control (fed TMR) and treatment (fed TMR with 0.1% PO and 2% SBO as DM basis). In the in vitro study (Experiment 1), PO or SBO did not affect ruminal pH. However, SBO tended to decrease ruminal ammonia-N (p = 0.099). Additionally, PO or SBO signicantly decreased total gas production (p = 0.041 and p = 0.034, respectively). Both PO and SBO signicantly decreased methane production (p < 0.05). In addition, PO signicantly increased both CLA (c9, t11 and t10, c12 CLA) (p < 0.001). Collectively, 0.1% PO and 2% SBO was selected that improved CLA and methane production most effectively and did not induce abnormal ruminal fermentation. In the in vivo study (Experiment 2), 0.1% PO with 2% SBO (PSO) did not affect CBC. However, it decreased BUN and MG levels in blood (p = 0.021 and p = 0.01, respectively). PSO treatment decreased pathogenic microbes (p < 0.05). It increased milk yield (p = 0.017) but decreased percentage of milk fat (p = 0.013) and decreased MUN level (p < 0.01). In addition, PSO treatment increased both CLA and PUFA in milk (p < 0.01). Finally, it decreased methane emissions from dairy cows. These results provide compelling evidence that PSO supplemented with a TMR-based diet can simultaneously increase CLA synthesis and decrease methane production with no inuence on the percentage of milk fat in lactating Holstein cows. , plant extracts can increase propionate and decrease acetate. Propionate can regulate in Biohydrogenation is involved in the metabolism to produce CLA from linoleic It is related to the their with Based on our previous study, we hypothesized that combining soybean oil (SBO) from nut could or by regulating and There is a lack of as to whether can affect nal and reduce study to investigate vitro vivo effects of oil plus on increasing CLA equipped with sunlight-roof and with dry manure for (4.5 2 per cow). Sunlight-roof covered the helped The cows were fed in mangers and each pen was equipped with separate water throughs. The procedures involving animals followed the ARRIVE guidelines. The ethical procedure was according to the Animal Experimental Guidelines in Konkuk University and approved the Animal Care and Use Committee (IACUC) of Konkuk Republic of (Approval KU18058). were subjected to least squares analysis of variance according to ANOVA and MIXED procedure of SAS (SAS Inst., Inc., Cary, NC, USA), respectively. Duncan multiple range test was used for ranking treatment means within a signicant F test. Animals in in vivo were randomly assigned to control and treatment groups based on completely randomized design. Statistical differences were considered signicant at P < 0.05 Differences among means with 0.05 < P < 0.10 were accepted as representing tendencies of differences.


Introduction
Recently, research on various health and eco-friendly substances is being conducted due to the ever-increasing interest in consumer health. Several studies have shown that plant or plant extracts can increase production quality in cows, including conjugated linoleic acid (CLA) 1 . CLA is known as a functional natural product that is produced in the rumen. It has been proven to have many bene cial effects on the health of humans. In addition, CLA has been shown to be an effective anticancer agent 2 . It is an antioxidant 3 . It can improve atherosclerosis 4 and immune response 5 . CLA can also lower cholesterol 6 and prevent diabetes 7 . According to previous research, although CLA form may vary greatly, cis-9, trans-11 CLA is the main form and its role is the greatest among all forms. Interestingly, cis-9, trans-11 CLA levels in ruminant products are higher than those in other animal products. This is because the cis-9, trans-11 CLA can be produced by microbes in the rumen of ruminant. To increase cis-9, trans-11 CLA levels in ruminant products, lots of soybean and sun ower compounds have been used because they contain high levels of linoleic acid 8, 9 . Their extract in oil form is known to be the direct source that can increase CLA 10 and inhibit methane production 11,12 . Although such sources of plant oil contain considerable linoleic acid, they also have a negative effect of increasing trans-10, cis-12 CLA which inhibits the synthesis of milk fatty acids [13][14][15] . Phytoncide oil (PO) has attracted attention as an additional source to deal with such problem.
POs are secondary metabolites from many plants 16 . PO has been reported to show antimicrobial properties similar to tannin and saponin 17,18 . This is because PO contains a large amount of β-carotene which is effective in regulating Gram-negative and Gram-positive bacteria 19 known to participate in the metabolism of fatty acid in rumen, termed bio-hydrogenation.
From what we discussed above and according to Oh et al. 20 , plant extracts can increase propionate and decrease acetate. Propionate can regulate methane production in rumen 20,21 . Biohydrogenation is involved in the metabolism to produce CLA from linoleic acid. It is related to the formulation of methane and volatile fatty acid (VFA) because of their competition with hydrogen ion 12,22 . Based on our previous study, we hypothesized that combining soybean oil (SBO) with phytoncide oil from pine nut cone (PO) could increase CLA yield and decrease methane production, while maintaining or increasing milk fat, by regulating the formulation of methane and VFA in rumen. There is a lack of information as to whether phytocides can affect nal ruminant production and reduce methane emissions. Therefore, the objective of this study was to investigate in vitro and in vivo effects of phytocide oil plus soybean oil on increasing milk CLA and depressing methane emissions in lactating Holstein cows.

Experimental materials and procedures for in vitro experiment
The procedures involving animals in this study followed the ARRIVE guidelines. All cannulation procedures involving animals were performed according to the Animal Experimental Guidelines in Konkuk University and the ethical approval was obtained by the Animal Care and Use Committee (IACUC) of Konkuk University, Republic of Korea (Approval No.: KU18058).
The ruminal uid was collected after ltering with two layers of cheese cloth through rumen cannula at 2 hours before feeding. The collected ruminal uid was kept in a 2 L thermos bottle that was lled with preheated CO 2 gas and with oxygen (O 2 ) completely removed, maintained at 39 °C, and then moved to the research lab. After removing feed particles by vacuum pump, the supernatant was used for the test.
The basal diet was dried in a drying oven (JS-CO2-AT100, Johnsam Co., Ltd. Bucheon, Korea) at 60°C for 24 hours. The forage and concentrate were then crushed with a 2 mm screen and used as a TMR at a ratio 6:4. Then 0.2 g of TMR and 20 mL of anaerobic McDougall buffer 23 were added to a 60 mL serum bottle containing 10 mL of ruminal uid. After adding 30 mL of culture solution to the culture bottle, two levels of PO (0 and 0.1% on DM basis arranged by Phylus Co., Ltd. Seoul, Korea) were added according to each treatment and four levels of SBO (0, 1, 2, and 4% on DM basis arranged by Haepyo Co., Ltd) as a precursor of CLA in the rumen were added, respectively. Each ruminal sample was kept at the maximum anaerobic conditions by ushing with CO 2 gas and capping with a rubber cap and an aluminum cap to prevent gas leakage as much as possible. The experiment was conducted with 8 treatments (three replicates per treatment) at time intervals of 0, 6, 12, 24, 48 h in a shaking incubator (JSP Corp, SI-900R; 100 rpm) at 39°C.

Analysis
After each incubation, total gas and methane were detected and the portion of ruminal sample was distributed for analysis of pH, VFA, ammonia-N, and fatty acid. Measurement of total gas was conducted using a 50 mL gas cylinder (Habdong Co., Anyang, Korea). Methane was detected using a Gas Chromatography/TCD (HP 6890, Hewlett Packard Co., CA, USA) equipped with an HP-PLOT-A column (dimension: 30 m × 0.32 mm × 20.00 μm) according to Owen et al. 24 . Culture solution of 2 mL was collected for ammonia (1 mL) and VFA analyses (1 mL). All samples collected were kept frozen at -20 °C until analyzed. Ammonia concentration was determined with the method of Fawcett and Scott 25 using a spectrophotometer (Model 680, BIO-RAD, CA, USA).
Culture solution of 1 mL was mixed with 0.1 mL 25% phosphoric acid. Then 0.2 mL pivalic acid solution (2%, w/v) was added as an internal standard. The mixed solution was centrifuged at 15,000 × g for 15 min and the supernatant was used to determine the concentration and composition of VFA using a gas chromatograph (HP 6890, Agilent Technologies, Santa Clara, CA, USA). Incubation solution of 20 mL for long chain fatty acids was also collected at each incubation time and freeze dried. Lipids were extracted using Folch's solution and analyzed with the gas chromatograph/FID (HP 7890 series GC System, Agilent technologies, California, USA) equipped with an Sp-2560 capillary column (dimension: 100 m × 0.25 mm × 0.2 μm lm thickness).

Animals and design for in vivo experiment
A total of 38 Holstein dairy cows (milk fat: 5.0 ± 0.30%; milk yield: 37.1 ± 2.85 kg/d/head; days in milk: 99.4 ± 24.32; parity: 3.1 ± 0.68) were randomly assigned to two groups to analyze milk fatty acids, methane production, fecal microbe population, and blood parameters and studied for 35 days. The cows were housed in a roofed shelter equipped with sunlight-roof and bedded with dry manure for comfort (4.5 m 2 per cow). Sunlight-roof covered the barn and winch curtain helped ventilation. The cows were fed in mangers and each pen was equipped with separate water throughs. The procedures involving animals in this study followed the ARRIVE guidelines. The ethical procedure was according to the Animal Experimental Guidelines in Konkuk University and approved by the Animal Care and Use Committee (IACUC) of Konkuk University, Republic of Korea (Approval No.: KU18058).

Diets and treatments
Dairy cows were fed total mixed ration (76.8%), roughage (8.4%), and concentrated feed (14.8%). These experimental diets were formulated to meet or exceed the NRC recommendations 26 . They were provided for ad libitum intake. Supplemental Table S1 shows ingredients and chemical compositions of the experimental diet. According to results of in vitro experiment, PSO (0.1% PO and 2% SBO on DM basis) were mixed with the vehicle and added to the TMR by top dressing. The control group (Control) was fed TMR added with the same amount of vehicle (not including PO or SBO).

Blood collection and analysis
For blood cell counting and metabolite analysis, blood samples were collected from each Holstein dairy cow via a jugular vein at 3 h before feeding and at 0, 21, and 35 days after feeding. Using EDTA tube (BD 367844 Vacutainer, Becton Dickinson, NJ, USA), whole blood was subjected to complete blood count (CBC) test using a VetScan HM2 Hematology System (Abaxis, Union City, CA, USA). In addition, the plasma portion of blood was obtained by using heparin tube (BD 367874 Vacutainer, Becton Dickinson, NJ, USA) and centrifuging at 3,000 rpm for 15 min in order to determine blood metabolites using a Chemistry analyzer (Furuno CA-270, Nishinomiya, Japan).

Fecal collection and analysis of microbes
Feces were collected on the same day at the same time as blood sampling. They were obtained from each lactating Holstein cow from the rectum, kept in liquid nitrogen, transferred into ice box at -4°C, transported to laboratory, and maintained in a deep freezer (-80°C) until analysis. One gram of fecal sample was diluted with 10 mL of 0.85% NaCl solution. Then 1 mL of the mixture was subjected to ten-fold serial dilution (10 1 to 10 11 ). The dilution was vigorously shaken with a vortex mixer. Then 0.1 mL of the dilution was spread onto a agar plate. Enumeration of bacteria was performed on MRS agar plates for Lactobacillus spp. and LB agar (Luria-Bertani broth diluted with bacto agar) plates for bacillus spp. Plates were incubated in anaerobic condition at 37°C for 24 h. Under the same condition, enumeration was done for pathogenic microbes Salmonella and E. coli on MacConkey agar plates. The number of colonyforming unit (CFU) was expressed as Methane emission analysis At 0, 7, 21, and 35 d, measurement of methane emission was conducted using Laser Methane mini-Green (LMm-G; Tokyo gas engineering solution, Tokyo, Japan). Between 10:00 and 11:00 hours when gas emission was the most active during feed intake, methane emission levels from all dairy cows in each group were measured in triplicates. To increase the reliability of measurement, a xed distance was used for each measurement and external environment (temperature, humidity, wind speed) was also measured. Finally, the total ppm value was calculated for each treatment by dividing the value of ppm-m (parts per million × meter) to the distance (meters) measured after derivation.

Milk collection and analysis
During the experimental period, milk was collected twice a day (at 3 am and 3 pm). Total milk yield was then calculated and the mean milk yield for 7 days was measured. At the end of the experiment, average milk yield was calculated. After mixing the collected milk in the morning and afternoon at each sampling day (0, 21 and 35 days), samples were stored at 4°C for milk composition analysis. Subsamples were stored at -20°C for fatty acids analysis.

Statistical analyses
Results obtained from in vitro and in vivo experiments were subjected to least squares analysis of variance according to ANOVA and MIXED procedure of SAS (SAS Inst., Inc., Cary, NC, USA), respectively. Duncan multiple range test was used for ranking treatment means within a signi cant F test. Animals in in vivo were randomly assigned to control and treatment groups based on completely randomized design. Statistical differences were considered signi cant at P < 0.05 Differences among means with 0.05 < P < 0.10 were accepted as representing tendencies of differences.

In vitro experiment
The Changes in ruminal pH, ammonia-N (NH 3 -N) and gas production is shown in Table 1. Oil supplementation materials did not affect the pH of the culture solution over the incubation time. No signi cant differences were observed in ammonia-N concentration by phytoncide oil. However, numerically lower values were observed in phytoncide oil treatments. In contrast, ammonia concentration was increased by SO treatment (p = 0.099).
The production of total gas and methane was detected by using GC at 0, 12, 24 and 48 h incubation time. There were signi cant decreases in the production of total gas and methane among treatment groups (p < 0.05). Total gas production was lower (p = 0.041 and 0.034, respectively) when PO or SBO was added, especially from 48 h incubation. Additionally, from 6 h incubation, methane production was lower (p < 0.05) in treatment groups containing PO or soybean oil. Except for 12 h incubation, methane was much lower (p < 0.05) when both PO and SBO were added to the diet, compared with separate supplementation and the most decreases in methane was observed when 0.1% PO and 2% SBO supplementation.
The properties of volatile fatty acids at 48 h incubation time was analyzed by using GC (Table 2). There was no signi cant difference in individual VFAs.
In the variation of long chain fatty acids. From 6 to 48 h, the trans vaccenic acid (TVA) was higher (p < 0.05) in treatment groups containing PO or SBO, except for 4% dosage (Table 3). Additionally, linoleic acid was higher (p < 0.001) at all time in the treatment groups containing PO or SBO. From 24 h incubation, both conjugated linoleic acids (c9, t11 CLA, t10, c12 CLA) were higher (p < 0.001, p < 0.05), and total CLA was signi cantly higher (p < 0.001) in the PO treatment groups than the non-PO treatment groups. Additionally, soybean oil signi cantly increased CLA at 24 and 48 h after incubation (p < 0.05 at 24 h, p < 0.001 at 48 h). Moreover, more higher CLA level from 24 to 48 h was observed in the treatment group containing more SBO with PO (p < 0.05).
Collectively, 0.1% PO and 2% SBO supplementation on TMR was selected for in vivo study because its dosage indicated most activation of inducing hydrogen ion from methane production to biohydrogenation of fatty acids when considering the result of most decrease in methane and increase in CLA simultaneously with no in uence on ruminal fermentation.

Blood analysis
The result of CBC analysis was conducted by using whole blood (Supplemental Table S2). There were no signi cant differences in white blood cells, red blood cells, hemoglobin, hematocrit, mean corpuscular volume, and mean corpuscular hemoglobin or mean corpuscular hemoglobin concentration between the PSO treatment and control groups. In the result of the blood metabolic parameters (Supplemental Table S3), There was no signi cant difference (p > 0.05) in most of the parameters, except for blood urea nitrogen (BUN) and magnesium (MG). The two metabolites, BUN and MG, were lower (p = 0.021, p = 0.01) in the PSO treatment group.

Fecal microbes
The bene cial microbes in the fecal were not affected by the treatment. However, the pathogenic microbes were signi cantly lower (p < 0.05) from 21 to 35d in the fecal of the PSO treatment group (Figure 1).

Milk yield, composition and fatty acids
In the result of milk yield. There was no difference from the initiation of the study until 2 weeks. However, from 15 d, an increase (p < 0.01) in milk yield was observed in the PSO group (Table 4). Thus, the total average of milk yield was higher (p = 0.017) in the PSO group than in the control group. Although milk fat content (%) was lower in PSO group, 4% FCM was not difference between two groups. In addition, lower value of MUN was observed (p < 0.01) in the treatment group (Table 5). Although the metabolites of CLA, stearic acid, or trans vaccenic acid were not changed consistently, cis-9, trans-11 CLA in milk was higher at nearly double (p < 0.01), and trans-10, cis-12 CLA was higher (p < 0.05) in the treatment group. Additionally, the total CLA was higher (p < 0.01) in the PSO group (Table 6).

Methane emission analysis
The analysis of methane emission from the cows was conducted by using laser detector in each group ( Figure. 2). There were 4 detection times during the experiment at (0, 7, 21, 35) d. Although the data could not be analyzed statistically because we measured in total for each group with 3 replicates, we could nd that lower methane gas emissions was observed in the treatment group than in the control group from 7 to 35 d.

Discussion
In vitro experiment This in vitro experiment was conducted to select optimal dosage of PO and SBO which could be used in in vivo experiment by considering the effect of the oils on production of methane, CLA and in uence on ruminal fermentation.
In this experiment, ruminal pH was not affected by PO or SBO (Table 1). Some studies have reported that ruminal pH could be variously changed by highconcentrate diets in feed and the addition of fat source. High-concentrate diets supplemented with sun ower oil can promote the fermentation of rumen microbes and consequently decrease ruminal pH 27 . Doreau and Ferlay 28 have reported that ruminal pH does not change even in high-energy diets such as SBO. As mentioned before, the variety of ruminal pH could be affected by changes of ruminal fermentation patterns. Considering that neither SBO nor PO affected ruminal pH value, these sources might not directly affect the environment of ruminal fermentation.
Results of our experiment showed that there was no signi cant difference in the amount of ammonia concentrate in the group treated with 0% PO at 48 h (Table 1). However, ammonia was affected by SBO (increase of ammonia in SBO supplementation group with PO at 0.1%). The synthesis of ruminal ammonia nitrogen is also known to be in uenced by the composition of fat source. Beauchemin et al. 29 reported that the supplemented fat sources can be used as an energy source for rumen microbes, affecting ruminal fermentation patterns and the synthesis of microbial protein. However, it is also known that fat additives, which can be used as an energy source, have various effects on ruminal pH and microbial activity, depending on the fat composition 28 . In this study, when considering that the ammonia-N concentrate was numerically higher in 0% than in 0.1% PO treatments and the ammonia was increased by the SBO at 0.1% PO treatments, it is speculated that PO and SBO affected the synthesis of microbial protein and ammonia-N differently.
Generally, total gas production and methane levels in rumen are negatively correlated in in vitro experiments with fat supplement. This is because the fat source can facilitate ruminal lipolysis, resulting in decreased methane and increased total gas production by affecting ruminal microbes 12,22 . Wettstein et al. 30 have reported that fat supplement in diet can reduce total gas production, which means exhausted energy. Thus, they suggested that fat supplement could improve the energy availability of microorganisms by reducing the exhausted energy in rumen. On the other hand, it has been reported that the addition of unsaturated fatty acids in feeds can act as a hydrogen sink in the rumen besides lipolysis, inducing hydrogen ions that could be used for methane production, to the process of converting unsaturated fatty acid into saturated fatty acid 22 .
In addition, as with PO, the essential oil has antimicrobial effects in the rumen, and results in lowering gas production during ruminal fermentation 31 .
Considering all these phenomena, the results of this study indicate that the reduced total gas and methane production were affected by the hydrogen sink effect and especially accelerated the lipolysis effect of the oil; the effect could be con rmed more severely at 0.1% PO and 2% SBO addition (Table 1).
Factors affecting VFA compositions in association with the rumen fermentation environment are various, including the ratio of forage to concentrate, high protein, carbohydrate and fat content in feed, supplemented fat composition, and so on 30,32 . Depending on the fermentation rate of feed in rumen, ruminal pH and VFA composition changes are associated with each other. For example, the rate of fermentation in the rumen in the case of overfeeding concentrate could be faster than that in the case of feeding low concentrate, resulting in lower ruminal pH, lower levels of acetate and butyrate, and increased propionate ratio in the VFA composition. In particular, it has been reported that oil supplement in feed can be fermented by ruminal microbes and used as an energy source, thereby modifying the VFA composition by reducing the pH in the rumen 33 . As previously mentioned, the addition of unsaturated fatty acids in the rumen can increase the production of propionate, act as a hydrogen sink, and reduce methane production. Castillejos et al. 34 have reported that antimicrobial materials such as essential oil can increase the ratio of acetate to propionate in calves fed a 6: 4 alfalfa hay: concentrate diet. In contrast, Cardozo et al. 35 have observed that the ratio of acetate to propionate is decreased in calves fed a 1: 9 straw: concentrate (based on corn, barley, and soybean meal) diet supplemented with essential oil. It has been reported that effects of essential oil on ruminal pH and feed composition are very diverse 31 . Considering these results, further research is needed to determine the effect of plant-derived essential oil on ruminal VFAs.
In results of long chain fatty acids in rumen, consecutive changes of linoleic acid were observed (Table 3). In the early stage of fermentation, TVA from linoleic acid was increased by the SBO supplement. CLA was increased by SBO and PO. Linoleic acid, which is found in SBO, is known to produce CLA and TVA through isomerization and biohydrogenation by microorganisms in the rumen 36 . According to Kairenius et al. 37 , fatty acid metabolism is affected by the composition of the added fat. Higher degree of unsaturation of the added fat causes more extensive metabolism of fatty acids. In addition, Kim et al. 16 have reported that the addition of not only fat, but also antimicrobial essential oil such as PO, can induce an increase of CLA in the milk of ruminants. Taking all these studies into account, the increase of CLA in the present study might be due to effects of SBO containing a large amount of unsaturated fatty acids and PO. Furthermore, it was speculated that the most activation of inducing hydrogen sink might be occurred by 0.1% PO and 2% SBO addition when considering most decreases in methane and VFAs and increase in CLA production simultaneously.

In vivo experiment
This in vivo experiment was carried to con rm the effect of selected supplementation (0.1% PO and 2% SBO) in in vitro experiment on increasing milk CLA and depressing methane emission by using dairy cows.
As expected from results of in vitro experiment (Table 6), CLA percentage and CLA yield were higher in the PSO group. According to Wang et al. 36 , CLA in milk could be increased when linoleic acids, the same as SBO, are added to the diet. The problem is that increased t10, c12 CLA often decreases milk fat as a result of increased total CLA. In the current study, milk fat percentage was decreased. However, milk yield was higher in PSO group than control thus, amount of milk fat (kg/d) was not difference between two groups. In addition, total CLA yield was not decreased, although t10, c12 CLA was higher in the treatment group than that in the control group. Considering all these phenomena, PO might have induced more c9, t11 CLA isomerization than t10, c12 CLA isomerization by affecting related microbes (e.g., Butyrivibrio brisolvens). In addition, increased UFA and PUFA levels were observed at the same time in the treatment group. All these differences between the two groups could be explained by the amount of linoleic acid in added SBO and PO inducing hydrogen ion to isomerization of linoleic acid from what might be methanogenesis results of in vitro experiment. Thus, induced isomerization in the treatment group might have increase UFA, PUFA, and CLA compared to the control group. In addition, stearic acid was higher in the PSO group at 35 d. This higher stearic acid could be explained by greater microbe biohydrogenation of linoleic acids in the PSO group fed more linoleic acid than the control group.
In the methane production results, one lack was the measuring methane per pen which was not individual animal, even though we tried to redress this limitation by measuring the methane 3 times during the experiment. However, interestingly, Figure 2 shows the adaptability of cows to the PSO treatment, as it shows considerable increase in 7 d, and gradual increase after 7 d to the end of the experiment. Generally, the production of methane in ruminant is decreased when dietary fat is added, and it is reported that the effect is varies greatly, depending on the constituents or unsaturation of added fat 12,22 . As mentioned earlier, the addition of unsaturated fatty acids is known to induce fatty acid metabolism and inhibit methane production in rumen microbes. In the result of gas production in in vitro experiment, it is observed that the addition of SBO, which contains a large amount of unsaturated fatty acid, decreased methane production, and PO addition resulted in more decrease in methane. Considering the results of fatty acid in rumen and milk, the CLA enhancement effect of the two oil sources con rmed in this study may ultimately lead to a decrease in methane production, by competition of hydrogen ion between biohydrogenation and methane production.
Complete blood cell counting (CBC) is a representative physiological index of animals. It was used in this study for the purpose of verifying the safety of experimental additives through animals. Some researchers have shown that oil supplements can affect animal disease because of the negative effect of unsaturated fatty acid in rumen 38 . However, in this study, there were no signi cant differences in results of CBC (white blood cell, lymphocyte, monocyte, or other indices) between control and treatment groups containing PO and SBO. Thus, the addition of PO and SBO supplement to the diet did not lead to diseaserelated chemical changes in dairy cows (Supplemental Table S2). Some researchers have found a relationship between fat supplement and blood metabolites 39,40 . According to Henderson et al. 39 , fat addition in ruminant diet affects the synthesis of microbial protein in rumen. The added fat is used as an energy source for protein synthesis, resulting in increasing amino acid-N and decreasing ammonia-N in the rumen, eventually reducing nitrogen content in the blood. It is known that fat degraded by lipolysis in rumen is fermented to increase propionate content in VFA compositions, thereby increasing glucose content in the blood by gluconeogenesis 41 . Similar to results of previous studies, the present study also showed that the PSO group containing oil supplement in feed had lower blood BUN but higher glucose concentration than the control group (Supplemental Table S3). Blood magnesium was also higher in the oil supplement group compared to the control. In general, most of the digestive and absorption lipid is transferred itself or binding to protein. However, some proportions of the lipid binding to magnesium are presented in the blood. Thus, the higher blood magnesium in our result might be affected by the oil supplement in the diet 42 . The reason why blood Ca showed a tendency to be signi cant was unknown. This parameter was inconsistent with other plasma outcomes.
Fecal microbes are microorganisms involved in the digestive physiology of livestock. Their digestibilities have been evaluated according to their populations 43 . Currently, antimicrobial effects of essential oils have been extensively studied in rumen. However, their effects on post-ruminal digestive have not been studied in detail yet. Furthermore, although the essential oil PO used in the experiment is also known to have antimicrobial effects, it is not yet known whether such effect could affect the whole ruminant digestive system 44 . However, as shown in the present study, numbers of pathogens in fecal microbes decreased due to the addition of PO and SBO. This demonstrates that the antimicrobial effect of PO also affects post-ruminal digestive system. Thus, the result of decrease in pathogenic microbes by PSO could support that PO might also be able to improve the energy availability of feed in ruminant (Figure 1).
Generally, some studies have reported that milk production is positively correlated with the energy content in feed 45 . As expected, this study also showed that the addition of 2% fat source, a high energy source, resulted in an increase in milk yield. Bauman 45 has reported that the increase in glucose in the blood is positively correlated with milk yield. In particular, the fat supplement used in this study increased glucose in the blood by increasing propionate in the VFA composition due to the effect of SBO and PO on rumen fermentation in in vitro experiment. Furthermore, the antimicrobial effect of PO might have affected fecal microbes. Thus, PO might be able to increase milk yield by improving the energy availability of SBO and basal diet (Table 4).
From milk composition data (Table 5), milk fat content (kg/d) and 4% FCM were similar between control and treatment groups. This is different from the result of Rico and Harvatine 15 . This might be because SBO has an effect on milk fat depression by increasing trans-10, cis-12 CLA known to inhibit the synthesis of milk fat. In our result, the unchanged milk fat in the treatment group may be explained by the effect of PO on isomerization to cis-9, trans-11 CLA rather than to trans-10, cis-12 CLA. Milk nitrogen is generally in uenced by blood nitrogen and ammonia-N in rumen. Thus, decreased MUN in the PSO group could be explained by the ow of decreased BUN in blood (Supplemental Table S3).
In summary, a mixture of 0.1% phytoncide oil (PO) and 2% soybean oil (SBO) in diet can lead to an increase in CLA and a decrease in methane production in rumen by inducing hydrogen ion from methane production to biohydrogenation, especially isomerization of linoleic acid with no change in milk fat. Therefore, a combination of PO and SBO could result in increased milk CLA and depressed methane production with no in uence on milk fat.  Tables   Table 1 Effects of phytoncide essential oil and soybean oil on ruminal pH, NH 3 -N and gas production  Table 3 Effects of phytoncide essential oil and soybean oil on long chain fatty acids (LCFAs) in rumen uid      Figure 1 Effects of the mixture on fecal coliform and bene cial microbe population (35 days). Values are expressed as means. Contrasts of the same microbes between Control and PSO in each period are shown (e.g. Bene cial-Control vs Bene cial-PSO). *Differences in superscript indicate signi cance at P < 0.05. Bene cial fecal microbes: Bacillus and L. bacillus spp; Pathogenic fecal microbes: E. coli and Salmonella spp.