The effectiveness of chemical additives on fermentation profiles, aerobic stability and in vitro ruminal digestibility of total mixed ration ensiled with Napier grass and wet distillers’ grains in southeast China

Abstract Chemical additives have been widely used to restrain the growth of undesirable microorganisms, and they were ideal additives to well preserve silages. The work aimed at evaluating the effects of four chemical additives on fermentation quality, aerobic stability and in vitro ruminal digestibility of the fermented total mixed ration (FTMR). The TMR was treated with: (1) no additive (control); (2) sodium diacetate; (3) calcium propionate; (4) sodium benzoate (SB); (5) potassium sorbate (PS) on a fresh weight (FW) basis. After 60 days of ensiling, the silos were opened and sampled to determine fermentation quality, in vitro rumen parameters, and then the 60-day FTMR were subjected to a 9-day aerobic stability test. Chemical additives significantly (p < 0.05) increased dry matter, crude protein, and water-soluble carbohydrates contents, and decreased ammonia nitrogen, ethanol contents, aerobic bacteria and yeast counts after 60 days of ensiling. During aerobic exposure, chemical additives obviously improved aerobic stability as indicated by higher lactic acid, water-soluble carbohydrates contents and lactic acid bacteria counts, and lower pH, ethanol, ammonia nitrogen contents, aerobic bacteria and yeast counts. Treatments with SB and PS had higher (p < 0.05) lactic acid content, and lower (p < 0.05) pH, ammonia nitrogen content, aerobic bacteria and yeast counts than other FTMR. After 72 h of incubation, chemical additives significantly (p < 0.05) increased cumulative gas production, in vitro digestibility of dry matter and crude protein. The SB and PS-treated FTMR had higher (p < 0.05) cumulative gas production and in vitro dry matter digestibility than other FTMR. These results indicated that SB and PS were more effective to improve aerobic stability and in vitro ruminal parameters, and they were recommended as optimal additives for FTMR. HIGHLIGHTS Four chemical additives were served as additives for fermentation total mixed ration Chemical additives markedly improved fermentation quality, aerobic stability and in vitro ruminal parameters Sodium benzoate and potassium sorbate were optimal for improving aerobic stability and in vitro parameters


Introduction
Wet brewers' grains are the main by-products of the brewing industry, and the annual production yield is more than 10 million tons in China. Wet brewers' grains are rich in crude protein (CP, 20.3%), ether extract (EE, 6.0%), vitamins and minerals (Chiou et al. 1998), and consequently it is more available for ruminants. However, it is a challenge to store and utilise such a large amount of wet brewers' grains for farmers and farms. Ensiling could be a better option to preserve wet brewers' grains, which could constantly supply high-quality protein feedstuffs for ruminants. It is noted that water-soluble carbohydrates (WSC) of grains were utilised during the brewing process, and fermentation quality tended to be poor when wet brewers' grains were ensiled alone (Ferraretto et al. 2018). Recently, wet brewers' grains and local fresh grass were typically used as ruminant feedstuffs in the form of roughage mixtures (Yuan et al. 2013). In southeast China, Napier grass (Pennisetum purpureum schum) was characterised by a perennial and high biomass production, and has been more used for silage making compared to other grass (Gusmao et al. 2018;. However, dry matter intake and ruminal digestion might decline when fed as the sole roughage (Su arez et al. 2007). Therefore, preparing total mixed ration (TMR) is an efficient practice in China to utilise roughage based on a mixture of by-products and fresh grass.
The TMR is a type of complete diet formula that can meet the nutritional requirements of ruminants. The TMR materials were ensiled in a sealed silo to create an anaerobic environment and fermented for a long time, which was defined as fermented TMR (FTMR). It avoids daily labour for TMR preparation, and also improved the palatability by altering odours and flavours through anaerobic fermentation. However, it is easy to spoil when being exposed to air or feed-out, particularly in hot and humid regions. Exposure to air or feed-out may stimulate the activity of aerobic microbes, such as yeasts and aerobic bacteria, leading to the decrease of lactic acid (LA) content, and the increase of pH and nutrient loss. Meanwhile, poor palatability and less intake may also occur. Therefore, it is crucial to seek effective methods to avoid or reduce aerobic deterioration.
Chemical additives are good choices to improve aerobic stability by inhibiting the proliferation of aerobic microorganisms (Sepp€ al€ a et al. 2016). In recent years, chemical additives mainly focus on some organic acid salts , such as sodium diacetate (SDA), calcium propionate (CAP), sodium benzoate (SB) and potassium sorbate (PS). These chemical additives can ionise to produce organic acids (mainly acetic, propionic, benzoic and sorbic acids) and salt ions, which have acidic characteristics and antibacterial property . The undissociated molecules of these organic acids can pass through the plasma membrane and liberate protons to acidify the cytoplasm, thus preventing the growth of yeasts and moulds (Kleinschmit et al. 2005). Therefore, it is hypothesised that the application of these chemical additives might contribute to improving the fermentation quality and aerobic stability of FTMR. Furthermore, little information is about the effects of these chemical additives on the fermentation profiles, aerobic stability and in vitro ruminal digestibility of FTMR.
The work aimed at investigating the effects of chemical additives on the fermentation profiles, aerobic stability and in vitro rumen parameters of FTMR ensiled with wet brewers' grains and Napier grass.

Silage preparation
Wet brewers' grains were obtained from a commercial liquor company (Jiafurui Biotechnology Co., Ltd., Nanjing, China). The concentrate supplement was acquired from a private dairy cattle farm (Jiangsu, China). Napier grass was harvested at the late vegetative stage (6 weeks after sowing) from the experimental field of Nanjing Agricultural University (Jiangsu, China). It was fertilised with nitrogen and potassium using ammonia sulphate and potassium chloride, respectively. After a standardising cut, it was chopped into a length of 2-3 cm using a mechanical chopper (9ZT-1, Zhengzhou Hualong Agriculture and Animal Husbandry Machinery Co., Ltd., Zhengzhou, China). The concentrate supplement consisted of 17.5% cracked corn, 8% rape cake meal, 25% cotton seed meal, 22.5% distillers dried grains with solubles (DDGS), 25% wheat bran and 2% vitamin-mineral. The chemical composition of pre-ensiled TMR is shown in Table 1.  Table 2. The TMR was treated with various chemical additives based on fresh weight (FW): (1) no additive (control); (2) sodium diacetate applied at 5 g kg À1 FW (SDA); (3) calcium propionate applied at 5 g kg À1 FW (CAP); (4) sodium benzoate applied at 1 g kg À1 FW (SB); (5) potassium sorbate applied at 1 g kg À1 FW (PS). Chemical additives were dissolved in water, followed by spraying at a rate of 5 mL kg À1 TMR, and the control was added with equal distilled water. After mixing thoroughly, 6 kg mixed raw material was packed into a 10-L laboratory silo (27.5 cm diameter Â 31.6 cm height, Lantian Biological Experimental Instrument Co., Ltd., Jiangsu, China) to obtain a packing density of approximately 600 kg FW/m 3 (five silos per treatment). All silos were sealed with a screw top and plastic taps, and then all silos were stored at ambient temperature (23-26 C). After 60 days of ensiling, five silos per treatments were opened and sampled to determine fermentation quality, in vitro rumen parameters, and then the residual FTMR were subjected to a 9-day aerobic stability test.

Chemical and microbial analyses
The chemical and microbial compositions of raw materials, pre-ensiled TMR and FTMR were analysed, respectively. About 100-g sample was oven-dried at 65 C for 48 h to determine dry matter (DM) content. The dried sample was ground to pass a 1-mm screen with a laboratory knife mill (FW100, Taisite Instrument Co., Ltd., Tianjin, China) for later analysis. Total nitrogen (TN, 984.13), EE (963.15) and ash (942.05) were determined according to the description of the Association of Official Analytical Chemists (AOAC 2012). The content of CP was determined as TN Â 6.25. The modified phenol-sulfuric acid method (Thomas 1977) was used to determine WSC content. The methods of Van Soest et al. (1991) were used to analyse neutral detergent fibre (NDF), and acid detergent fibre (ADF) with an ANKOM 200i fibre analyser (ANKOM Technologies, Inc., Fairport, NY, USA). Hemicellulose (HC) was calculated as NDF minus ADF. Non-fibrous carbohydrate (NFC) was calculated by the formula: NFC ¼ 1000 À CP À NDF -EE À ash (NRC 2001).
Approximate 30 g sample was blended with 90 mL distilled water and macerated at 4 C for 24 h. The extract was filtered through four layers of medical gauze and a filter paper (pore size of 11 lm, Xinhua Co., Ltd., Hangzhou, China). The pH of filtrate was measured with a glass electrode pH metre (HANNA pH 211; Hanna Instruments Italia Srl, Padova, Italy). The ammonia nitrogen (NH 3 -N) content was determined by the phenol-hypochlorite reaction method of Chen et al. (2014). Buffering capacity (BC) of raw materials and pre-ensiled TMR were determined according to Playne and McDonald (1966). The analyses of organic acids and ethanol were conducted in high performance liquid chromatography system (1260 HPLC, Agilent Technologies, Inc., Waldbronn, Germany) equipped with a refractive index detector (column: Carbomix V R H-NP5, Sepax Technologies, Inc., Newark, DE, USA; eluent: 2.5 mmol L À1 H 2 SO 4 , 0.5 mL min À1 ; temperature: 55 C).
About 10 g of pre-ensiled TMR and FTMR were serially diluted tenfold with sterilised saline solution (0.85% sodium chloride). Lactic acid bacteria (LAB), aerobic bacteria (AB) and yeasts were enumerated according to the description of Chen et al. (2017).

Aerobic stability test
After 60 days of ensiling, 100 silos (5 replicates Â 4 exposure days Â 5 treatments) were opened for a 9-day aerobic stability test. In brief, the FTMR sample was removed from each silo to determine fermentation quality and in vitro parameters, and the residual FTMR were loosely loaded into a bigger 15 L open-tap and sterile polyethylene bottle. All bottles were stored at ambient temperature (18-22 C) and covered with a double layer of gauze. The probes of multi-channel temperature recorder (MDL-1048A, Shanghai Tianhe Automation Instrument Co., Ltd., Shanghai, China) were placed in the centre of the bottle for measuring temperature variation. Five probes were placed in the environment as blanks. A temperature recording was conducted every 30 min for 9 days. The aerobic stability was defined as the needed time (h) when the temperature of FTMR samples remained 2 C above the ambient temperature (Wilkinson and Davies 2013). The FTMR was sampled to determine pH, LA, acetic (AA), propionic (PA) and butyric acids (BA), ethanol, WSC, NH 3 -N contents, LAB, AB and yeast counts during aerobic exposure.

In vitro ruminal incubation
The experiment was approved by the Ethics Committee of the Nanjing Agricultural University (Jiangsu, China). The ground TMR silages (1 g) was put into a nylon bag (8 Â 12 cm, 42-lm pore size). The rumen fluid was obtained from 3 Boer goats through rumen fistulas before the morning feeding. The goats were fed twice a day with a diet comprising 5% alfalfa hay, 50% guinea-grass hay and 45% concentrate. The rumen fluid was immediately filtered through 4 layers of gauze, immediately transported to the laboratory, and incubated at 39 C in a water bath. Before use, the rumen fluid and the buffer were mixed at the ratio of 1:2 (v/v). The artificial buffer solution was prepared as the instruction of Menke and Herbert (1988). The nylon bag was previously washed with acetone, dried at 65 C to a constant weight. One nylon bag was prepared for each silo, and a total of 25 samples (5 treatments Â 5 silos replicates Â 1 nylon bag for per replicate Â 1 run) were prepared. Another five serum bottles with empty nylon bags were served as blanks.
All nylon bags were heat-sealed and placed into preheated serum bottles with the mixture fluid at 39 C. Continuous CO 2 was flushed to the serum bottle to keep anaerobic condition. Gas production (GP) was measured using a pressure transducer after 4, 8, 12, 24, 36, 48 and 72 h and corrected with the blank bottles. The cumulative GP was fitted to the exponential equation: y ¼ b (1 À e Àct ), where y is the cumulative volume of GP (mL) at time t, b is the potential GP (mL) and c is the rate constant of GP, t is the incubation time (h). The metabolisable energy (ME) was estimated with the method of Menke and Herbert (1988).
After 72 h of incubation, all nylon bags were rinsed with tap water and then dried at 65 C for 48 h. The in vitro digestibility of dry matter (IVDMD), crude protein (IVCPD) and neutral detergent fibre (IVNDFD) were determined by the relative amounts of different weights before and after incubation to initial weights, respectively.

Statistical analyses
Analysis of variance (ANOVA) was performed using the general linear model (GLM) procedure of SAS 9.2 (SAS Inst. Inc., Cary, NC, USA). Data on fermentation quality, in vitro rumen parameters, chemical and microbial compositions after 60 days of ensiling were subjected to one-way ANOVA. During aerobic exposure, the data of pH, organic acids, ethanol, WSC, NH 3 -N content, LAB, AB and yeasts counts were subjected to two-way ANOVA with the fixed effects of treatment, aerobic exposure day and treatment by aerobic exposure day interaction using GLM procedure of SAS. Day (0, 3, 6 and 9 days) during aerobic exposure was included in the repeated statement to account for repeated measures. Statistical difference between means were determined by Tukey's multiple comparisons and significance was declared at p < 0.05. The Pearson correlation analyses between fermentation parameters and microbial counts of FTMR during aerobic exposure were analysed using the statistical packages for the social sciences (SPSS 25.0 for windows; SPSS Inc.).

Results
Fermentation quality of total mixed ration after 60 days of ensiling The fermentation quality of TMR is shown in Table 3. Chemical additives significantly (p < 0.05) decreased ethanol, NH 3 -N contents, AB and yeast counts. The addition of SDA and CAP significantly (p < 0.05) increased LA content and decreased the pH compared with other FTMR. The SDA treatment had the highest (p < 0.05) AA content among all treatments. The addition of CAP significantly (p < 0.05) increased PA content, whereas SDA and PS significantly (p < 0.05) decreased PA content compared with the control. None or tiny amounts of BA content was observed in all FTMR. The yeast counts were not more than 1 Â 10 2 cfu g À1 FW, except that in the control. The LAB counts in all FTMR exceeded 1 Â 10 6 cfu g À1 FW, and no significant (p > 0.05) differences were found on LAB counts among all FTMR.

Chemical compositions of FTMR after 60 days of ensiling
The NDF, ADF, HC, EE and ash contents were not significantly (p > 0.05) affected by chemical additives, except DM, CP and WSC contents (Table 4). Chemical additives significantly (p < 0.05) increased DM, CP and WSC contents compared with the control.

Aerobic stability of FTMR
Fermentation parameters and chemical compositions of FTMR during aerobic exposure are shown in Tables 5 and 6, respectively. Chemical additives, aerobic exposure days, and their interaction had significant (p < 0.05) effects on pH, LA, AA, PA, ethanol, WSC and NH 3 -N contents. All FTMR significantly (p < 0.05) decreased LA content, and increased the pH until the end of aerobic exposure. Chemical additives significantly (p < 0.05) increased LA content and decreased the pH compared with the control. Treatments with SB and PS had higher (p < 0.05) LA content and lower (p < 0.05) pH than other FTMR. Chemical additives significantly (p < 0.05) increased AA content compared with the control, and SDA treatment had the highest (p < 0.05) AA content among all FTMR. The control, CAP, and SB-treated FTMR significantly (p < 0.05) decreased PA content, and SDA and PS-treated FTMR numerically (p > 0.05) decreased PA content during aerobic exposure. Chemical additives significantly increased PA content compared with the control, and CAP treatment had the highest (p < 0.05) PA content among all FTMR. The ethanol content in all FTMR showed a significant (p < 0.05) decrease till the end of aerobic exposure. All FTMR significantly (p < 0.05) increased NH 3 -N content during aerobic exposure. Chemical additives significantly (p < 0.05) decreased ethanol and NH 3 -N contents compared with the control, and SB and PS treatments had lower (p < 0.05) NH 3 -N content than other FTMR. The WSC content in all FTMR began to decrease (p < 0.05) after 6 days, and chemical additives significantly increased (p < 0.05) WSC content compared with the control during aerobic exposure. None or tiny amounts of BA content were found in all FTMR.
Chemical additives, days of aerobic exposure, and their interaction had significant (p < 0.05) effects on LAB, AB and yeast counts (Table 7). All FTMR significantly (p < 0.05) decreased LAB counts, and increased AB and yeast counts during aerobic exposure. Chemical additive-treated FTMR had higher (p < 0.05) LAB counts, and lower (p < 0.05) AB and yeast counts than the control. Both SB and PS treatments had lower (p < 0.05) AB and yeast counts than other FTMR.   .127 a a-d Day means in a row with different superscripts differ (p < 0.05), A-C FTMR means in a column with different superscripts differ (p < 0.05). w-z Means with different small letters show significant differences between the aerobic exposure day effect within a treatment (p < 0.05), W-Y Means with different capital letters show significant differences between the treatment effect within an aerobic exposure day (p < 0.05). DM: dry matter; SEM: standard error of the mean; Control: no additive; FTMR: fermented total mixed ration; SDA: sodium diacetate; CAP: calcium propionate; SB: sodium benzoate; PS: potassium sorbate; D: aerobic exposure days; T: treatments; D Â T: interaction between aerobic exposure days and treatments. .43 a a-c Day means in a row with different superscripts differ (p < 0.05), A-C FTMR means in a column with different superscripts differ (p < 0.05). w-y Means with different small letters show significant differences between the aerobic exposure day effect within a treatment (p < 0.05), W-Z Means with different capital letters show significant differences between the treatment effect within an aerobic exposure day (p < 0.05). DM: dry matter; TN: total nitrogen. SEM: standard error of the mean; Control: no additive; FTMR: fermented total mixed ration; SDA: sodium diacetate; CAP: calcium propionate; SB: sodium benzoate; PS: potassium sorbate; D: aerobic exposure days; T: treatments; D Â T: interaction between aerobic exposure days and treatments.
The aerobic stability of FTMR is shown in Figure 1. Chemical additives significantly (p < 0.05) improved aerobic stability compared with the control, and SB and PS prolonged the days of aerobic stability more than 9 days.
The Pearson correlation between fermentation parameters and microbial counts of fermented total mixed ration during aerobic exposure The positive or negative correlations between fermentation parameters and microbial counts of the FTMR during aerobic exposure are summarised in Table 8. There is a negative correlation between pH and LA content, AB count and LA content, AB count and AA content, AB count and PA content, ethanol content and yeast count, respectively (p < 0.01). However, there is a positive correlation between pH and AB count, pH and NH 3 -N content, respectively (p < 0.01).

In vitro parameters of fermented total mixed ration
The gas production parameters and in vitro digestibility of FTMR are summarised in Table 9 and Figure 2. Chemical additives significantly (p < 0.05) increased cumulative GP, IVDMD and IVCPD. The PS treatment had higher (p < 0.05) potential GP and rate constant of GP than the control, while these were not different (p > 0.05) among the control, SDA, CAP and SB treatments. There was no significant (p > 0.05) difference in IVNDFD and ME among all FTMR.

Discussion
Chemical composition of pre-ensiled total mixed ration It was noted that the Napier grass used in the experiment was only a sample of Napier grass another sample harvested with different soil fertility and at different maturity might vary. It is known that proper DM content (>20%DM), sufficient WSC content (>5%DM), low buffering capacity, and a suitable initial 2.00 c 4.28 b 5.00 a 5.39 a a-d Day means in a row with different superscripts differ (p < 0.05), A-C FTMR means in a column with different superscripts differ (p < 0.05). w-z Means with different small letters show significant differences between the aerobic exposure day effect within a treatment (p < 0.05), W-Y Means with different capital letters show significant differences between the treatment effect within an aerobic exposure day (p < 0.05). FW: fresh weight; SEM: standard error of the mean; Control: no additive; FTMR: fermented total mixed ration; SDA: sodium diacetate; CAP: calcium propionate; SB: sodium benzoate; PS, potassium sorbate; D: aerobic exposure days; T: treatments; D Â T: interaction between aerobic exposure days and treatments. LAB count (>1 Â 10 5 cfu g À1 FW) are the key factors to ensure good fermentation quality (McDonald et al. 1991;Weinberg 2008). The DM content (53.2.7%DM), WSC content (15.91%DM), BC (188.88 mEq kg À1 DM) and the epiphytic LAB counts (2.63 Â 10 8 cfu g À1 FW) in the experiment suggested that TMR met the prerequisite for satisfied fermentation quality.

Fermentation quality of total mixed ration after 60 days of ensiling
The pH is one of the critical factors that reflects fermentation quality, and the threshold value to ensure good quality is less than 4.20 (Chen et al. 2017). In the experiment, the pH in chemical additives treatments ranged from 4.19 to 3.86, indicating these FTMR were well preserved. The addition of SDA and CAP markedly decreased the pH compared with other FTMR, which might be caused by the increasing LA content. The low NH 3 -N content (<100 g kg À1 TN) indicated the occurrence of weak proteolysis (McDonald et al. 1991). All FTMR had NH 3 -N content for less than 55.0 g kg À1 TN, indicating less proteolysis. Chemical additives evidently decreased NH 3 -N content and increased CP content, which might be due to the inhibition effects on activity of proteolytic clostridia Zhang et al. 2020). The SDA treatment had higher AA    content, and CAP treatment had higher PA content than other FTMR. These results might result from ionisation of SDA and CAP to produce corresponding organic acids. Chemical additives increased DM content, and decreased AB and yeast counts. A probable reason was that the antibacterial properties of chemical additives can restrain the growth of aerobic bacteria, and result in the decrease of DM loss (Bernardes et al. 2015). Though SDA and CAP treatment had obviously lower pH than other FTMR, they did not perform better antibacterial properties. The probable explanation for this was that the antibacterial effects of organic acids increased as the increasing molecular weight, implying the weak bacteriostatic function of acetic and propionic acids (McDonald et al. 1991). The lower ethanol content in chemical additives treatments was probably related to the decline in the activity of yeasts (Chen et al. 2017).

Aerobic stability of fermented total mixed ration
Silages were prone to aerobic deterioration after exposure to air or feed out, especially in high temperature and humid regions. Consequently, organic acids and WSC were oxidised by yeasts and AB, and released heat. Therefore, monitoring the changes of fermentation parameters and microbial counts are essential to evaluate the aerobic stability of the FTMR. In the experiment, the yeast counts in chemical additives-treated FTMR were more than 5.0 log 10 cfu g À1 FW after 3 days of aerobic exposure, indicating the proneness to aerobic deterioration. The silages with yeast count in excess of 5.0 log 10 cfu g À1 FW were susceptible to aerobic deterioration (McDonald et al. 1991). All FTMR evidently decreased LA content and increased the pH until the end of aerobic exposure. Kleinschmit et al. (2005) reported that lactate-utilizing yeasts can oxidise LA into carbon dioxide and water, leading to the rise of pH. Compared with the control, chemical additives notably increased LA content and decreased the pH. It was more likely that higher WSC content of chemical additive-treated FTMR might contribute to the establishment of WSC-utilizing yeasts rather than lactate-utilizing yeasts. In comparison with lactate-utilizing yeasts, WSC-utilizing yeasts can dominate the conversion of WSC into LA, resulting in the drop of pH .
The decrease of AB and yeast counts in chemical additive-treated FTMR was probably related to the increasing AA and PA contents. It has been proven that AA and PA can effectively inhibit the growth of aerobic bacteria, thus improving aerobic stability of silages (Kung et al. 1998;Danner et al. 2003). Both SB and PS treatments had lower AB and yeast counts than other FTMR. It might be related to intensive antibacterial properties of benzoic and sorbic acids, which were ionised from SB and PS. McDonald et al. (1991) suggested that antimicrobial functions of fatty acids increased with the increasing molecular weight, indicating intensive antimicrobial functions of benzoic and sorbic acids. Chemical additives evidently decreased NH 3 -N content, indicating that proteolysis was inhibited. Both SB and PS treatments had lower NH 3 -N content, might be also related to intensive inhibition effects.
The production of ethanol from WSC was mainly dominated by yeast activity (da Silva et al. 2020), and high ethanol content was commonly accompanied with high yeast counts. However, the negative relation between ethanol content and yeast counts was found in the experiment. Therefore, the decrease of ethanol content was assumed to volatilisation rather than microbial metabolisms.
During aerobic exposure, microbial activities should account for the alteration of organic acid fractions in the FTMR (Courtin and Spoelstra 1990). Pearson correlation analysis indicated that the activity of aerobic bacteria was negatively related to the LA content and positively related to the pH. It was possibly ascribed that the growth of yeasts consumed residual WSC and LA, resulting in the rise of pH (Muck et al. 1991). The AB and yeast counts in chemical additive-treated FTMR were lower than that in the control, which indicated that aerobic deterioration could be effectively inhibited.

In vitro parameters of FTMR
In vitro culture has been conducted as a common method to predict rumen digestibility and the ME of ruminants. The IVDMD is an important parameter that reflects the rate of feed utilisation, and the digestion of DM mainly includes carbohydrates, proteins and other substances (Li et al. 2014). In the experiment, a significant increase in IVDMD was found in chemical additive-treated FTMR, which might be related to the increase of WSC and CP contents.  suggested that high CP content was conducive for promoting IVDMD and increasing the availability of nutrients. Higher IVCPD in chemical additive-treated FTMR was probably related to more soluble CP, which resulted in an increase of total degradable CP fractions. Chemical additives significantly increased cumulative GP, which might be explained that the inhibition effects on the growth of undesirable bacteria, and reduced the loss of nutrients, thus providing sufficient energy and substrates for microbial metabolisms (Chen et al. 2014).

Conclusions
The results suggested that SB and PS obviously improved aerobic stability and in vitro ruminal digestibility compared with other chemical additives, as indicated by higher LA, WSC content, LAB counts, cumulative GP, potential GP and IVDMD, and lower pH, ethanol, NH 3 -N contents, AB and yeast counts. By comprehensive consideration, SB and PS were recommend to improve fermentation quality, aerobic stability and in vitro ruminal digestibility of FTMR.

Ethical approval
All animal experimental protocols were approved by the Animal Care and Use Committee of Nanjing Agricultural University.

Disclosure statement
There are no conflicts of interest in this work.

Data availability statement
The data that support the findings of this study are available from the corresponding author, ST, upon reasonable request.