Effects of roughage type on particle separation, rumination, fiber mat characteristics, in situ degradation, and ruminal fermentation parameters in beef steers

Abstract

Six ruminally cannulated steers (average BW = 791 ± 71 kg) were used in a replicated 3 × 3 Latin square experiment to determine the effects of roughage type on rumination, fiber mat characteristics, and rumen fermentation variables. Three roughages were included at 7% (DM basis) in a steam flaked corn-based diet: cotton burrs (CB), wheat silage (WS), or corn stalks (CS). Steers were fitted with a sensory collar to record rumination behaviors in 2-h intervals at the beginning of the experiment. Each 30-d period consisted of 7 d of recovery, 14 d of diet adaptation, 7 d of rumination data collection (daily and bi-hourly average rumination), 1 d of rumen fluid collection, and 1 d of rumen evacuations. In situ degradation of individual roughages was determined for 4 d after period 3 evacuations. During rumen evacuations, ruminal contents were removed; the rumen fiber mat (RF) was separated from the liquid portion with a 2-mm sieve, weighed, and a subsample was dried. Data were analyzed using the MIXED procedure of SAS with steer as the experimental unit and roughage (CB, WS, and CS) as the main effect. Dry matter intake (DMI) was not different for CB and WS (P = 0.25) and greatest for steers consuming CS diet (P ≤ 0.01). Roughage type did not influence the weight of the RF dry matter (%; DM; P = 0.92), RF weight (P = 0.69), or RF:DMI ratio (P = 0.29). Daily rumination (min/d) did not differ among roughages (P = 0.40), but min of rumination/kg of DMI was greatest for CS (18.0 min), min/kg of NDF was greatest for WS (89.8 min; P = 0.02), and min/kg of peNDF was greatest for CS (132.4 min; P ≤ 0.01). Wheat silage had the greatest percentage of soluble and degradable DM. Rumen fiber mat did not differ for roughages, although rumination min/kg of DMI and peNDF was greatest for steers consuming CS and WS. In situ degradation determined that CB-R and CS-R had the greatest percentage of ruminal undegraded DM. Based on the objective of the experiment, roughage type did not influence daily rumination or fiber mat characteristics.

Introduction

To speed the time to harvest, beef cattle rations are formulated to contain high energy densities using grains and by-product feed ingredients. To prevent digestive upsets due to high-concentrate inclusion rates of finishing rations, lower amounts of roughage (5% to 15%) are included in finishing diets (Mertens, 2002). Roughages create a fibrous ruminal mat, stimulate chewing activity (Gentry et al., 2016; Jennings et al., 2019), and decrease acidosis (Galyean and Defoor, 2003). Unique physical characteristics of individual roughages contribute to the density of the ruminal mat and may influence ingredient digestibility, ruminal retention times, and particle degradation (Schulze et al., 2014 and NASEM, 2016). Decreasing fiber inclusion rate and particle size influences feedlot performance (Jennings et al., 2019) and ruminal fermentation (Weiss et al., 2017) and shortens retention time within the rumen (desBordes and Welch, 1984). Reducing retention time decreases degradation of neutral detergent fiber (NDF), increases fiber concentration in the hind gut, and reduces total starch digestion (Hales et al., 2014; NASEM, 2016).

Physically effective NDF (peNDF) is the fraction of feed ingredients or diets that stimulates chewing and may contribute to the overall ruminal fibrous mat and in dairy diets, and can be estimated based on particle size and NDF concentration of the diet (Mertens, 1997 and Mertens, 2002). The physical effectiveness factor of roughages can be determined based on the animal’s total chewing time, ruminal pH, and ruminal mat resistance (Goulart et al., 2020). Roughages utilized in the beef industry vary due to geographical location, availability, and costs. Little is known about the peNDF of these various roughages and their effect on chewing behaviors, fibrous mat, and fermentation characteristics in beef cattle. We hypothesized that the unique peNDF of roughages would contribute to the formation of the fibrous ruminal mat that would influence rumination, ruminal fermentation, fiber concentration of the fibrous mat, and particles retained on increasing sieve sizes. The objective of this experiment was to evaluate the rumination, ruminal mat fiber concentration, and ruminal fermentation of beef steers fed steam flaked corn-based diets with 7% (DM basis) corn stalks (CS), wheat silage (WS), or cotton burrs (CB).

Materials and Methods

All methods and procedures involving live animals were approved by the West Texas A&M University—Cooperative Research, Educational, and Extension Team Institutional Animal Care and Use Committee.

Ruminally cannulated steers (n = 6; average BW = 791 ± 71 kg) with a 10.2-cm diameter cannula (Bar Diamond, Parma, ID) were in used in a 3 × 3 Latin square design with three dietary treatments and three 30-d periods. Steers were housed in individual pens at the Texas A&M AgriLife Research Feedlot in Bushland, TX. Dietary treatments consisted of steam flaked corn (SFC)–based diets and differed based on three roughage sources were formulated at 7% inclusion rate (Table 1). Diets were formulated to contain 30 g/ton monensin (DM basis), 10 g/ton tylosin phosphate (DM basis; Rumensin and Tylan, respectively; Elanco Animal Health; Greenfield, IN), and adequate vitamins and minerals to meet or exceed NASEM (2016) requirements. Diets were mixed individually in a mixer wagon mounted on load cells (IV 84-8; Roto-Mix, Dodge City, KS; readability ± 0.45 kg). While in the mixer, diets were allowed 4 minutes (min) of closed door, continuous mixing to ensure adequate mixing without pulverization. After mixing, diets were unloaded into feed bins and individual steer’s feed was weighed using a platform scale (Ohaus SD Series; Ohaus Corp. Parsipanny, NJ). Orts were collected daily prior to morning feeding and bunks were always managed to contain 0.45 kg of feed.

Table 1.

Ingredients and calculated chemical composition of experimental diets

Ingredient, % DM	Diet1
	CB	WS	CS
Steam flaked corn	54.3	53.8	52.8
Wet corn gluten feed2	18.5	17.7	17.9
Wet distillers’ grains with solubles	14.4	15.3	15.6
Wheat silage	–	6.89	–
Corn stalks	–	–	7.21
Cotton burrs	7.33	–	–
Urea	0.30	0.32	0.30
Limestone	0.69	0.86	0.93
Corn oil	1.60	1.61	1.81
Supplement3	3.48	3.51	3.48
Calculated chemical composition, % DM
DM, % as-fed	66.4	60.9	63.8
CP, %	15.5	15.5	15.1
NDF, %	20.0	19.3	20.5
ADF, %	11.2	11.8	11.1
Fat, %	5.23	5.32	5.40
Ca, %	0.80	0.74	0.77
P, %	0.50	0.50	0.51
S, %	0.23	0.21	0.20
Na, %	0.17	0.17	0.17
ME, mcal/kg4	3.14	3.16	3.14
NEm, mcal/kg	2.05	2.07	2.06
NEg, mcal/kg	1.53	1.57	1.62

¹CB, cotton burrs; WS, wheat silage; CS, corn stalks.

²Sweet Bran, Cargill Corn Milling.

³Supplement was formulated to meet or exceed vitamin and mineral requirements established by NASEM (2016). Supplement formulated to provide 30 g/ton of monensin and 10 g/ton of tylosin phosphate (Rumensin and Tylan, respectively; Elanco Animal Health, Greenfield, IN).

⁴Energy values were obtained from Preston (2016).

Each period consisted of 7 days (d) of recovery, 14 d of diet adaptation, 7 d of data collection (daily feed intake and rumination), and 2 d of rumen fluid sampling. During the diet adaptation period (days 8 to 21), steers were fed one third of their daily allotment of feed, three times daily at approximately 0700, 1300, and 1900 h. During data collection (days 22 to 28), steer’s rumens were not disturbed, and cattle were observed for a 2-h interval (minute/hour; min/h) and daily (min/d) rumination values. Only on sampling days, steers were offered 100% of their daily allotment of feed at 0700 h before rumen fluid sample (day 29) or evacuations (day 30). After evacuations, steers were moved to a pen for ad libitum access to a recovery diet (78.4% roughage and 11.5% concentrate) for 2 d and then transitioned onto their next diet by decreasing the portion of the recovery diet and replacing it with the next treatment diet (days 3 to 7). The day following the recovery period, steers began the next 14-d adaptation period. For calculation of DMI, data collected during adaptation period were used (days 8 to 28) and DMI during sampling days was removed from the analysis.

Data and Sampling Collection

Prior to the start of the first period, steers were fitted with a sensory collar (AllFlex Livestock Intelligence, Madison, WI) to continuously record swallowing and regurgitation of feed boli. Rumination time (min/h) was recorded 24 h a d (days 8 to 21) in 2-h increments to quantify individual steers rumination behaviors. Daily rumination was calculated by summing daily 2-h intervals and averaged across 24-h periods.

Rumen fluid samples were collected to analyze concentrations of ammonia (NH₃) and volatile fatty acids (VFA). At each collection, rumen contents (1 to 2 liters) removed from the rumen and immediately strained through five layers of cheese cloth. After straining, the liquid portion was placed into three 50-mL conical tubes and frozen until further analysis. Rumen fluid was collected at 0 h (1 h prior to feeding) and 4, 6, 10, 14, and 24 h post-feeding.

After rumen fluid collections, steers were offered 100% of their daily allotment of feed and given 5 to 7 h to consume it. Feed remaining was removed and weighed to determine that days DMI. After feed was removed, ruminal evacuations were performed on 1 steer per treatment and the ruminal contents were processed. After the ruminal contents or the first 3 steers were processed, ruminal evacuations of the remaining 3 steers were performed. Total rumen contents were removed and placed in a large bucket (18.9 liters). Rumen contents were dumped onto a 2-mm sieve where the fibrous portion (RF) was separated from the liquid. The total time RF spent in the sieve was dependent on how long it took to separate the liquid from RF. After the RF was separated, total RF was weighed, and a subsample was stored. Remaining RF and liquid portion were returned to the rumen. The RF subsample was placed into 1 of 3 metal pans (35.6 × 35.6 cm) and placed in an oven (Despatch Model KBB2-18-1; Despatch Industries, Minneapolis, MN) at 42 °C for approximately 72 h. While drying, samples in pans were mixed daily to prevent clumping. Based on the DMI, the ratio of rumen fiber (RF) to DMI (RF:DMI; kg:kg) was calculated.

After the third period evacuation, steers were returned to their individual pens and continued to receive their experimental diets for an additional 6 d. Based on the diet the steers were consuming, in situ digestion was used to determine the percentage of disappearance of the ground roughage samples. Samples of individual roughages, cotton burrs (CB-R), wheat silage (WS-R), and corn stalks (CS-R), were collected and dried at 55 °C for approximately 48 h and ground in a Wiley mill (model 4; Thomas Scientific, Swedesboro, NJ) through a 2-mm screen. Approximately 1.25 g of dried sample was weighed into 5- × 10-cm nylon bags (three bags/timepoint; 50 ± 10 µm pore size; Ankom Technology, Fairport, NY). One blank bag/timepoint was used to account for microbial attachment. Before insertion in the rumen, in situ bags were placed in 31- × 20-cm laundry bags by timepoint and incubated in 40 °C tap water for 20 min. Except for 0-h bags, laundry bags were placed in the rumen of steers receiving the respective roughage (2 steers/roughage). Laundry bags were suspended in the rumen 1 h before the 0700-h feeding and remained in the rumen 24, 48, 72, or 96 h. To determine the soluble fraction, a separate set of bags were incubated in 40 °C tap water for 20 min, rinsed in cold water, and placed in the oven to dry (0 h). Immediately after removal from the rumen, laundry bags were rinsed in cold water. In situ bags were then removed and gently rinsed in cold water approximately three times or until water was clear. Bags were dried in 55 °C oven for approximately 72 h. After drying, triplicate bags were weighed and composited by timepoint within steer. Composited samples were sent to a commercial laboratory (Servi-Tech Laboratories) for NDF, ADF, and OM analysis.

The in situ ruminal DM and NNDF degradation parameters were calculated based on Mertens first-order exponential model with discrete lag (1977). The model is

$${\displaystyle \begin{array}{l}{\displaystyle {\mathrm{R}}_{(\mathrm{t})}=\left(\mathrm{b}\right){\mathrm{e}}^{-\mathrm{k}}{{}_{\mathrm{d}}}^{(\mathrm{t}-\mathrm{L})}+\mathrm{c}}\end{array}}$$

where “R_t” is the total digested residue at a given time “t,” “b” is the potentially degradable fraction, “k_d” is the fractional rate of degradation (h⁻¹) of b, “t” is the time incubated in the rumen, “L” is the discrete lag time in h, and “c” is the residue remaining after 96 h. The soluble “a” and potentially degradable “b” fractions were determined as a + b by the difference of the residue remaining from the 96-h bags. The slope of the natural log of residue remaining after lag through 96 h is the k_d. Lag was calculated as the difference in intercepts of the linear regression equation after lag minus the actual intercept at time zero divided by k_d.

Laboratory analysis

Diet and individual ingredient samples were taken weekly for DM determination. Dry matter content of the ingredients and orts was determined using a forced-air oven at 55 °C for approximately 48 h. After drying, ingredients were composited by month and experimental diets were composited by period. A subsample of individual ingredients was sent to a commercial laboratory (Servi-Tech Laboratories; Amarillo, TX) monthly for nutrient analysis of fat (AOAC # 920.39 and 2003.5), mineral concentrations (AOAC # 990.08), NDF (ANKOM procedure 6.2017), ADF (ANKOM procedure 5.2017), and ash (AOAC # 942.05). A subsample of the monthly roughage samples, corn stalks (CS-R), cotton burrs (CB-R), and wheat silage (WS-R), were composited into two samples and sent to a commercial laboratory (Table 2; Cumberland Valley Analytical Services, Waynesboro, PA) for the additional analysis of crude protein (CP; AOAC # 990.03), soluble protein (SP), neutral detergent insoluble crude protein (NDICP; AOAC # 990.03), acid detergent insoluble crude protein (ADICP; AOAC # 990.03), starch (AOAC # 996.11), lignin (AOAC # 973.18), ethanol soluble carbohydrate (ESCHO; Dubois et al., 1956), NDF, ADF, ash, and mineral analysis. Net energy for maintenance (NE_m) and gain (NE_g) were calculated based on equations presented in Weiss et al. (1992) and Galyean et al. (2016).

Table 2.

Nutrient analysis of individual roughages

Item1, % DM	Roughage2
	CB	WS	CS
DM, %	89.0	35.3	89.1
CP	9.35	9.70	5.65
SP	4.60	7.00	2.35
NDICP	5.26	1.20	1.61
ADICP	4.80	0.91	1.51
ESCHO	0.20	3.30	5.75
NDF	70.5	43.1	77.2
ADF	61.9	61.8	51.8
Lignin	19.6	7.44	7.54
Starch	0.70	4.30	0.40
Crude fat	1.93	2.61	1.26
TDN	26.1	51.3	46.7
NEm (Mcal/kg)3	0.82	1.21	0.96
NEg (Mcal/kg)	0.28	0.65	0.41
ME (Mcal/kg)	1.64	2.17	1.81
Nonfiber carbohydrates	7.25	16.9	9.15
Nonstructural carbohydrates	1.60	7.60	6.15
Mineral
Organic Matter, % DM	16.3	10.3	8.35
Calcium, % DM	1.91	0.37	0.30
Phosphorus, % DM	0.15	0.32	0.05
Magnesium, % DM	0.34	0.16	0.17
Potassium, % DM	1.90	2.31	1.56
Sulfur, % DM	0.31	0.18	0.07
Sodium, % DM	0.07	0.03	0.03
Chloride, % DM	0.19	1.14	0.70
Iron, ppm	670.5	275.0	278.7
Manganese, ppm	39.5	41.0	61.5
Zinc, ppm	15.0	41.5	26.0
Copper, ppm	5.50	6.00	8.50

¹CP, crude protein; SP, soluble protein; NDICP, neutral detergent insoluble crude protein; ADICP, acid detergent insoluble crude protein; ESCHO, ethanol soluble carbohydrates; NDF, neutral detergent fiber; ADF, acid detergent fiber; NE_m, net energy for maintenance; NE_g, net energy for gain; ME, metabolizable energy.

²CB, cotton burrs, WS, wheat silage, CS, corn stalks.

³Energy values were calculated based on equations presented in Galyean et al. (2016). ME = (0.929 × digestible energy) − (0.0056 × CP) + (0.0343 × fat) + (0.0042 × starch) − 0.3612. NE_m = (1.1104 × ME) − (0.0946 × ME²) + (0.0065 × ME³) − 0.7783. NE_g = (1.1376 × ME) − (0.1198 × ME²) + (0.0076 × ME³) − 1.2979.

Diets, roughages, SFC, and WDGS samples were collected weekly during each period and separated as-is using the Penn State Particle Separator (PSPS) methods described by Konoff et al. (2003). After drying, RF samples were separated by animal within period using PSPS methods and stored by animal, sieve size, and period. The PSPS is composed of a stack of four sieves and the sieve size decreases with the bottom pans: 19, 8, 4, and less than 4 mm. Sample was placed on the 19-mm sieve and the stack of sieves shook back and forth 5 times, then shifted 90°, and shaken another 5 times. This was repeated until the stack of sieves was rotated in a complete circle. After the sample was separated, the sample retained on each sieve was weighed. The sample retained on each sieve was summed and the percentage of the sample retained on each sieve was calculated. For individual ingredients, diets, and RF, physical effective NDF (peNDF) was calculated by multiplying the percentage of the samples > 4 mm (percentage of particles retained in the top 3 sieves) by NDF concentration of the diet. A subsample from each sieve for roughages, diets, and RF was collected and analyzed for NDF concentration using an automatic fiber analyzer (Model 200/220; ANKOM procedure 6.2017). Samples were analyzed for DM (AOAC # 934.01) by drying 2.0 g of sample for 24 h at 100 °C and organic matter (OM) was determined by ashing samples at 600 °C for approximately 6 h.

Statistical analysis

For particle separation of individual ingredients, data were analyzed using PROC MIXED of SAS to compare LSMEANS of the percentage of particles retained on individual sieves. The model included ingredient (WS-R, CS-R, CB-R, WCGF, WDGS, and SFC) as the main effect and day collected as the random effect.

Data were analyzed as a replicated 3 x 3 Latin square using PROC MIXED of SAS (SAS Inst. Inc., Cary, NC). The model included diet (CB, WS, and CS) as the main effect and steer as a random effect. For RF characteristics, particle separation, and sieve NDF concentration, diet was used as the main effect and steer as the random effect. For VFA, NH₃, and rumination (min/d and min/h), the model included diet, time, and their interaction as a main effect.

In situ degradation, parameters were analyzed using PROC MIXED of SAS with a model that included roughage (CB-R, WS-R, and CS-R), when significant LSMEANS were compared with Tukey’s range test.

For NDF concentration by sieve, the model for individual roughages was analyzed for the main effects of roughage (CB-R, WS-R, and CS-R) and experimental diet (CS, CB, and WS) period as a random effect. For RF, diet was used as the main effect and steer or period used as random effects, respectively.

For all analysis, separation of significantly different means was determined using the LSMEANS statement with PDIFF option. Means were considered significant when P < 0.05 and a trend when 0.05 < P ≤ 0.10.

Results and Discussion

Particle separation for feed ingredients is presented in Table 3. The ingredient with the greatest (P ≤ 0.01) percentage of particles retained on the 19-mm sieve was CB-R; WS-R and CS-R were intermediate; and WCGF, WDGS, and SFC had the least. For particles retained on the 8-mm sieve, SFC had the greatest and CB-R had the least (P ≤ 0.01). The percentage of particles on the 4-mm sieve was greatest for WCGF, WDGS, followed by SFC, CS-R, WS-R, and CB-R (P ≤0.01). Cotton burrs, CS-R, and WCGF had the greatest percentage of particles that were less than 4 mm and WS-R had the least (P ≤ 0.01). Wheat silage had the greatest percentage of particles greater than 4 mm, whereas CS-R and CB-R had the least (P ≤ 0.01). Estimated peNDF was greatest for WS-R and least for SFC (P ≤ 0.01).

Table 3.

Particle separation analysis and estimated physical effective NDF (peNDF) of feed ingredients

Item	Ingredient1						SEM	P-value
	WS-R	CS-R	CB-R	WCGF3	WDGS	SFC
NDF, % DM	58.2	70.6	63.3	35.4	36.8	7.9
	Retained/Screen, %
Sieve screen size
19 mm	37.8b	36.0b	58.7a	0.05f	8.61d	4.11e	5.14	≤ 0.01
8 mm	53.1b	25.2c	8.43d	52.4b	53.4b	61.8a	4.63	≤ 0.01
4 mm	6.77c	14.1b	6.09c	27.6a	28.7a	17.1b	2.50	≤ 0.01
Less than 4 mm	2.30d	24.7a	26.8a	20.0a	9.23c	17.0b	2.71	≤ 0.01
Greater than 4 mm	97.7a	75.3d	73.2d	80.0c	90.8b	83.0c	2.71	≤ 0.01
peNDF2, %	56.9a	53.1c	46.3b	28.3e	33.4d	6.6f	0.89	≤ 0.01

¹WS-R, wheat silage; CS-R, corn stalks; CB-R, cotton burrs; WCGF, wet corn gluten feed; WDGS, wet distillers’ grains with solubles; SFC, steam flaked-corn (12.7 kg/L flake).

²Physical effective NDF = (percentage of particles retained on top 3 sieves) × ingredient NDF, %.

³Sweet Bran; Cargill Corn Milling.

^a–f Unique superscripts differ when P≤ 0.05.

One disadvantage of the PSPS method is the lack of ability to identify uniqueness of individual ingredients based on what they contribute to the nutrient composition of the diet. For example, the percentage of particles retained on the greater than 4 mm sieve was > 80% for most of the ingredients, including SFC, WCGF, and WDGS. The percentage was < 80% for corn stalks and cotton burrs. Although SFC, WCGF, and WDGS had a greater percentage of particles retained, they contribute to the NDF concentration of the diet, but not to the lignin concentration that contributes to the peNDF. The concentrate ingredients contribute to the starch, protein, and energy of the diet. The fiber concentration of the roughages allows them to contribute to the physical coarseness of the diet and will aid in stimulating rumination and ruminal health (NASEM, 2016). When using the PSPS for individual roughages with a high percentage of larger particles (> 19 mm), results may indicate the contribution of that roughage to the formation of an extensive fibrous mat within the rumen. Currently, the peNDF calculation based on results of the PSPS is based on particle size and NDF concentration but does not take into consideration daily rumination (Yang and Beauchemin, 2006; Jennings et al., 2019). During the development of peNDF calculations, Mertens (1986) proposed a roughage value unit that would take into consideration the chewing stimulated by a roughage. This unit would be based on the chemical measurement of fiber and physical measurement of the particle size of the roughage. Further development determined that the peNDF of a feed ingredient is based on its physical properties, stimulation of chewing, and the biphasic stratification of particles within the ruminal contents (Mertens, 2002). When finishing diets are separated using PSPS methods, the particle size of the roughage may be diluted because of the lower roughage inclusion rate compared to the inclusion rates of the concentrates.

Particle separation of diets is presented in Table 4. The WS and CB diets had the greatest (P ≤ 0.01) percentage of particles retained on the 8-mm sieve. The CS had the greatest (P = 0.02) percentage of particles retained on the sieve less than 4 mm, WS was intermediate, and CB had the least. The CB had the greatest percentage (P = 0.02) of particles greater than 4 mm, and WS and CS were not different. Physically effective NDF was greatest for CB and least for WS and CS (P ≤ 0.01). Particle separation of RF is presented in Table 4. For particles retained on 19-mm sieve, CB had the greatest (P = 0.01), CS were the intermediate, and WS had the least. Wheat silage had the greatest percentage of particles retained on the 8-mm sieve, CB were the median, and CS had the least (P ≤ 0.01). Corn stalks had the greatest percentage of particles retained on the 4-mm sieve followed by WS and CB (P ≤ 0.01). For the percentage of the smallest particles (< 4-mm sieve), CS had the greatest (P = 0.02), WS were the intermediate, and CB was the least. Cotton burrs and WS had the greatest percentage of particles greater than 4 mm and CS had the least (P ≤ 0.01). Estimated peNDF of the RF was greatest (P = 0.02) for CB, and WS and CS were not different (Table 5).

Table 4.

Effects of roughage source on dietary particle separation

	Diet1
	CB	WS	CS	SEM	P-value
Dietary NDF, % DM	20.0	19.3	20.5	0.25	0.20
	Retained/Screen, %
Sieve screen size
19 mm	5.26	3.90	5.19	0.81	0.43
8 mm	48.7a	49.4a	40.6b	1.85	≤ 0.01
4 mm	19.6	17.9	20.4	0.94	0.20
Less than 4 mm	26.3c	28.6b	33.7a	1.71	0.02
Greater than 4 mm	73.6a	71.3b	66.2b	1.71	0.02
peNDF2, %	14.7a	13.7b	13.6b	0.43	≤ 0.01

¹Roughage was included at 7% (DM basis) of a steam flaked corn-based diet; CB, cotton burrs; CS, corn stalks; WS, wheat silage.

²Physical effective NDF = (percentage of particles retained on top 3 sieves) × dietary NDF, %.

^a–c Means with unique superscripts differ when P≤ 0.05.

Table 5.

Effects of roughage source on particle separation of dried rumen fiber (RF) mat

Item	Diet1			SEM	P-value
	CB	WS	CS
Dietary NDF, % DM	20.0	19.3	20.5	0.25	0.20
RF NDF, % DM	40.6	45.1	46.5	2.80	0.42
	Retained/Screen, %
Sieve screen size
19 mm	13.1a	5.7c	8.4b	1.66	0.01
8 mm	48.7b	51.1a	38.9c	1.75	≤ 0.01
4 mm	13.8c	18.7b	24.2a	0.92	≤ 0.01
Less than 4 mm	23.6b	24.7b	28.6a	0.97	≤ 0.01
Greater than 4 mm	76.4a	75.3a	71.4b	0.97	≤ 0.01
peNDF2, %	15.3a	14.5b	14.6b	0.50	0.02

Rumen evacuations were performed approximately 5 to 7 h after receiving 100% of their daily call. Steers consumed their experimental diets for 22-d diet prior to evacuation. A subsample of rumen contents was weighed and split into 1 of 3 pans and dried at 42 °C for approximately 72 h prior to particle separation.

¹Roughage was included at 7% (DM basis) of the diet; CB, cotton burrs; CS, corn stalks; WS, wheat silage.

²Physical effective NDF = (percentage of particles retained on top 3 sieves) × dietary NDF, %.

^a–c Means with unique superscripts differ whenP≤ 0.05.

When finishing cattle consume a diet with a large particle roughage, the rate of passage from the rumen may increase and lead to decreased starch digestion in the small intestine due to increased fiber concentration within the intestines (Weiss et al., 2017). Weiss et al. (2017) compared the effects of CS inclusion rate and grind size on fecal output. When cattle were fed a steam flaked corn-based diet with increasing roughage particle size, fecal starch and DM output decreased. Although fecal output was not measured in the present experiment, the unique physical structure of CB and WS and their smaller particle sizes may have trapped SFC particles in the rumen and increased microbial degradation of SFC. In comparison, the percentage of particles > 19 mm was not different for all three diets. After consumption, the percentage of particles > 19 mm was 4.7 and 7.4 % greater (P ≤ 0.01) for CB than CS and WS, respectively. For CB, the greater percentage (P ≤ 0.01) of particles retained on the 19-mm sieve may indicate that the unique cotton-like consistency of the roughage may trap smaller particles and prevent them from falling to the bottom sieves. If the same is true in the rumen, the homogenous mixing of rumen contents may have increased microbial exposure to feed particles that may seem beneficial for ruminal degradation of nutrients to occur.

The NDF and OM concentrations for different particle sizes of individual roughages, diets, and RF is presented in Table 6. Average NDF concentration (% DM) was 16.6% and 6.7% greater for WS-R (P = 0.02) and CB-R (P = 0.01), respectively. For particles retained on the 19- and 8-mm sieves, CS-R had the greatest concentration of NDF, CB-R were the intermediate (P = 0.01), and WS-R (P = 0.03) had the least on both sieves. However, WS-R had the greatest NDF for particles retained on the 4-mm sieve and CS-R had the least (P ≤ 0.01).

Table 6.

Neutral detergent fiber (NDF) and organic matter (OM) concentration (% DM) of particle collected on individual sieves for individual roughages, complete diets, and rumen fiber

Item	Roughage1					Diet2					Rumen Fiber3
	CB-R	WS-R	CS-R	SEM	P-value	CB	WS	CS	SEM	P-value	CB	WS	CS	SEM	P-value
NDF, % DM
Average	67.6b	57.7c	74.3a	1.04	≤ 0.01	24.7b	20.7c	26.7a	0.75	≤ 0.01	44.7	47.5	49.2	1.90	0.25
19 mm	71.7b	61.1c	84.8a	2.27	0.01	29.4a	23.0b	32.6a	1.44	≤ 0.01	48.8	47.6	52.6	4.02	0.68
8 mm	71.4b	59.3c	77.4a	2.50	0.03	17.4a	15.4b	19.0a	0.61	≤ 0.01	43.7	44.9	47.7	4.42	0.81
4 mm	67.4b	72.4a	54.1c	1.67	≤ 0.01	21.6a	16.4b	24.4a	3.09	≤ 0.01	39.7	46.4	46.6	3.84	0.39
< 4mm	60.2	56.4	62.5	1.74	0.19	27.0a	24.7b	27.4a	0.53	≤ 0.01	46.8	51.3	50.1	2.97	0.54
OM, % DM
Average	84.6c	89.4a	88.5b	0.18	≤ 0.01	92.8	93.8	93.6	0.18	0.25	89.9b	90.6a	90.8a	0.43	≤ 0.01
19 mm	88.4c	90.5b	92.7a	0.26	≤ 0.01	92.4b	94.4a	94.4b	0.50	0.02	91.0	90.1	90.5	0.38	0.32
8 mm	91.1b	90.1c	93.2a	0.33	0.02	94.7b	94.6b	95.6a	0.41	0.01	90.9	91.3	91.3	0.47	0.76
4 mm	88.7	89.8	89.9	0.46	0.24	93.9a	94.4a	93.2b	0.58	0.03	90.9	91.8	91.9	0.41	0.17
< 4mm	70.1c	87.1a	78.4b	0.83	≤ 0.01	90.0	91.9	91.0	0.82	0.74	87.6b	89.4a	89.5a	0.51	0.04

¹Individual roughages: CB-R, cotton burrs, WS-R, wheat silage, and CS-R, corn stalks.

²Roughage was included at 7% (DM basis) of steam flaked corn-based finishing diet; CB = cotton burrs; CS = corn stalks; WS = wheat silage.

³Rumen evacuations were performed approximately 5 to 7 h after feeding. Steers consumed their experimental diets for 22-d diet before evacuation. A subsample of rumen contents was dried at 42 °C for approximately 72-h before particle separation.

^a–cMeans with unique superscripts differ when P≤ 0.05.

Average OM concentration (% DM) was similar for WS-R and CS-R (P = 0.56) and least for CB-R (P ≤ 0.01). Organic matter concentration of particles retained on the 19 mm sieve was greatest (P ≤ 0.01) for CS-R, intermediate for WS-R (P ≤ 0.01), and least for CB-R. For the 8-mm sieve, CS-R had the greatest (P = 0.03) OM concentration and CB-R and WS-R did not differ (P = 0.24). Organic matter concentration for the smallest particles (< 4 mm) was greatest (P ≤ 0.01) for WS-R, CS-R was median, and CB-R was the least.

Average NDF concentration (%DM) was greatest (P ≤ 0.01) for diets with CS, intermediate for CB, and lowest for WS. Diets with CS and CB had the greatest NDF concentration for 19-, 8-, 4-, and < 4-mm sieves (P ≤ 0.01), whereas WS had the least across the four sieves. Organic matter concentration (% DM) was greatest (P = 0.02) for WS and CS and least for CB for particles on the 19-mm sieve. Diets with CS had the greatest (P ≤ 0.01) OM concentration on the 9-mm sieve and but did not differ from CB in the 4-mm sieve (P = 0.23). Wheat silage had the greatest (P = 0.03) concentration of OM in the 4-mm sieve.

Roughage type did not affect the average NDF (% DM; P = 0.25) or NDF concentration of different particles sizes for steer’s RF (P > 0.39). Average OM concentration (% DM) in RF was greatest (P ≤ 0.01) for steers consuming diets with CS and WS and least for CB. The OM concentration of the smallest particles (< 4 mm sieve) was greatest (P = 0.04) for WS and CS and least for CB. Based on results in the present experiment, the various particle size’s NDF concentration for individual roughages and diets differed. In contrast, feed particles in RF lacked those differences that may be associated with the dilution of roughage when separating diet as previously stated. When looking at NDF concentration of the diets across the sieves, NDF concentration of WS decreased as the sieve size decreased. The sieve that retained the largest particles (> 19 mm) also had the greatest NDF concentration. The physical characteristics of the roughages could have contributed to this. According to the nutrient analysis of the roughages (Table 1), CB-R had the greatest percentage of lignin, whereas CS-R and WS-R were similar. Also, the ESCHO or water-soluble carbohydrates was greatest in CS followed by WS, and least for CB. Based on the increased lignin of the CB-R compared to the other roughage, it may have resulted in an increase in retention of smaller particles to distribute them throughout the rumen for digestion (NASEM, 2016). Luginbuhl et al. (1990) observed the particle distribution of the fibrous ruminal mat of steers fed hay for 19 d. The percentage of small particles was two times greater than large and medium particles and four times greater than fine particles. Results in the present experiment indicate that the NDF concentration of particle sizes in individual roughages and diets differs due to unique nutrient analysis of the roughages. However, the unique roughage characteristics did not affect the NDF concentration in the RF.

Steers consuming CS diet had a 0.4 kg/d greater (P ≤ 0.01) DMI than CB and WS (Table 7). Roughage source did not affect dry matter percentage (P = 0.92), weight (P = 0.69) of the RF, or the RF:DMI ratio (P = 0.29). Calculated ruminal passage rate was not different for CB and WS but least for CS (P = 0.05). Previous research has shown that as weight or bulkiness is added to the rumen of beef steers, overtime, DMI decreases over time (Whetsell et al., 2004). Whetsell et al. (2004) reported that the weight of the ruminal contents was increased from 7.5 to 11.6 kg with plastic balls, DMI decreased linearly for steers consuming a high concentrate diet (15.1 to 5.1 kg, respectively). Results in the present experiment did not follow this trend. Numerically, steers consuming CB diet had the greatest RF weight and passage rate, but DMI was lower than CS. Steers consuming WS had similar DMI and passage rate compared to CB, whereas RF weight was 0.8 kg less than for CB. When performing rumen evacuations, visual observations indicated that steers consuming WS and CS had a defined RF mat, whereas RF in CB steers was a homogenous mixture. Visually, the RF mat for WS and CS trapped smaller particles of WCGF and WDGS while SFC dropped to the bottom of the rumen, while steers consuming CB, had smaller and larger particles trapped throughout the rumen in a homogenous mixture.

Table 7.

Effects of roughage source on dry matter intake (DMI), passage rate (K_p), percent of nutrient remaining, and rumination

Item	Experimental Diet1			SEM	P-value
	CB	WS	CS
DMI2, kg/d	11.6b	11.6b	12.0a	0.27	≤ 0.01
Kp, %/h3	2.06a	2.04a	1.84b	0.17	0.05
Rumen fiber4
DM, %	20.0	19.7	19.7	0.51	0.92
Dry weight, kg	7.7	6.9	6.9	0.94	0.69
RF:DMI5	1.51	1.68	1.73	0.16	0.29
Rumination
Daily, min/d	182.2	177.1	213.1	17.84	0.40
DMI, min/kg	15.4b	17.3a	18.0a	1.20	0.03
NDF intake, min/kg	77.0b	89.8a	86.6a	5.72	0.02
peNDF6 intake, min/kg	104.3c	126.4b	132.4a	7.93	≤ 0.01

¹Roughage was included at 7% (DM basis) of steam flaked corn-based finishing diet; CB, cotton burrs; CS, corn stalks; WS, wheat silage.

²Recorded during days 15 to 21 of the period.

³Passage rate (K_p) = 2.904 + 1.375 × [DMI, % BW] − 0.020 × [% concentrate in diet, DM]. Equation adapted fromSchwab et al. (2003).

⁴Rumen evacuations were performed approximately 5 to 7 h after receiving 100% of their daily allotment of feed. Steers consumed their experimental diets for 22-d before evacuation. A subsample of rumen contents was weighed and split into 1 of 3 pans and dried at 42°C for approximately 72-hbeforeparticle separation.

⁵Ratio of rumen fiber mat (RF) to DMI prior to rumen evacuation

⁶Estimated physical effective NDF = (percentage of particles retained on top 3 sieves) × dietaryNDF, %.

^a,b Means with unique superscripts differ when P≤ 0.05.

Average daily rumination did not differ among roughage sources (P = 0.40, Table 7). Rumination time per kg of DMI was greatest (P = 0.03) for steers consuming CS and WS and least for CB. Corn stalks had the greatest min rumination per kg of NDF intake (P = 0.02) and peNDF (P ≤ 0.01). Wheat silage and CB had similar min of rumination per kg of NDF (P = 0.25), whereas CB had the least for peNDF. desBordes and Welch (1984) reported that rumination in finishing cattle is important in reducing particle size of indigestible coarse materials and aids in increasing digestion, saliva production, and passage rate. Interestingly, the concentration of NDF was greater for CS-R and CB-R (Table 1) and least for WS-R. Based on results in the present experiment, CB-R had the greatest concentration of lignin and the largest particle size but interestingly, daily average rumination did not differ among diets. When comparing minutes of rumination per kg of NDF and peNDF, steers consuming CB ruminated 12.8 and 9.6 min less than WS and CS, respectively (P = 0.02). Based on the understanding of fiber degradation in the rumen, the roughage with a higher concentration of lignin would need more chewing to reduce its size and passage from the rumen. The increased percentage of larger particles and decreased minutes of rumination for NDF and peNDF may indicate that the steers consuming CB do not ruminate enough to break down the dense fiber. However, particle size was decreased for WS and CS and may indicate the increased rumination was beneficial in reducing particles.

Previous work by Gentry et al. (2016), Weiss et al. (2017), and Jennings et al. (2019) evaluated rumination of beef steers fed finishing diets with CS as the roughage source and can be used as a baseline for daily rumination. Average daily rumination for beef steers fed 5%, 10%, and 15% CS was 245, 225, and 298 min/d, respectively. Weiss et al. (2017) and Jennings et al. (2019) reported that steers that had the greatest daily rumination (313 min/d) had the greatest daily ruminal pH (5.92). This indicates that rumination in combination with roughage inclusion may aid in increasing ruminal pH and decreasing bouts of acidosis. Although ruminal pH was not measured in the present experiment, steers consuming CB and WS had less rumination compared to steers consuming CS which may lead to an increase in digestive upsets. Ruminal fiber of steers consuming CS had the least (P ≤ 0.01) percentage of particles greater than 4 mm compared to CB and WS. As previously stated, roughage type did not impact NDF concentration of different digesta particle sizes in RF. Rumination time per kg of daily NDF intake was not different for WS and CS diets.

There were no time × diet interactions for NH₃ (P = 0.29; Table 8) or VFA concentrations (mM; P ≥ 0.44). There was a tendency for roughage type to affect propionate concentration (mmol/L; P = 0.10) such that steers consuming WS tended to have a greater concentration followed by CS and CB. Butyrate concentration was greatest (P = 0.03) for WS and CB and least for CS. Valerate was greatest (P ≤ 0.01) for CB and least for WS and CS. Corn stalks and CB had the greatest (P = 0.02) A:P ratio and WS had the least. Diets in the present experiment were formulated to contain similar crude protein (CP) and urea inclusion rates, indicating that the steers consumed similar nitrogen concentrations daily. However, the CP of individual roughages was 9.4%, 9.7%, and 5.7% (DM basis) for CB, WS, and CS, respectively. The difference in ruminal NH₃ concentrations may be contributed to different rates in proteolysis or a difference in rumination and saliva production among steers (Marshall et al., 1992). Weiss et al. (2017) fed SFC-based diets with varying CS inclusion rates. Ruminal NH₃ concentrations were greatest for steers consuming a finishing diet with 10% CS inclusion rate compared to 5%. Marshall et al. (1992) fed a ground corn-based finishing diet and NH₃ concentrations ranged from 11 to 18 mg/dL, similar concentrations as the present experiment. Based on the nutrient analysis of the roughages, WS-R had the greatest percentage of SP (7.0%) and least ADICP (0.91%) compared to the other roughages. This may indicate that of the total protein fraction of WS-R, a greater portion degrades in the rumen, resulting in an increase in NH₃ concentration in the rumen, whereas CS-R and CB-R had the greatest percentage of unavailable protein (ADICP) and SP that may have resulted in a lower concentration of NH₃ in the rumen fluid. Ruminal concentrations (mM) of propionate, butyrate, valerate, and the A:P ratio in Warner et al. (2020) were like concentrations reported in the present experiment. When CB-R was added to the diet, A:P ratio and acetate increased, whereas concentrations of butyrate and valerate decreased. In the present experiment, when CB-R was included in the diet, valerate and A:P concentrations increased.

Table 8.

Effect of roughage source on ruminal ammonia and VFA concentration

	Experimental Diet1				P-value
	CB	WS	CS	SEM	Diet	Time2	Diet × Time
NH3, mg/dL	13.6	15.0	12.8	1.2	0.08	0.05	0.29
Total VFA, mM	127.0	131.5	129.0	3.87	0.52	≤ 0.01	0.71
VFA, mM
Acetate	58.1	59.1	60.8	1.13	0.34	≤ 0.01	0.73
Propionate	43.8	48.1	45.1	3.43	0.10	≤ 0.01	0.81
Butyrate	17.3a	17.8a	15.9b	0.77	0.03	≤ 0.01	0.22
Valerate	3.7a	2.5b	2.6b	0.42	≤ 0.01	≤ 0.01	0.93
A:P	1.40a	1.29b	1.48a	0.11	0.02	≤ 0.01	0.78

¹Roughage was included at 7% (DM basis) of steam flaked corn-based finishing diet; CB, cotton burrs; CS, corn stalks; WS, wheat silage.

²Rumen fluid was collected before feeding (0 h) and 2, 4, 6, 10, 14, and 24 h after feeding.

^a–c Means with unique superscripts differ when P≤ 0.05.

According to NASEM (2016), propionate is the only VFA that can be used for gluconeogenesis in the liver and contributes energy for growth of the animal. Butyrate is metabolized extensively in the ruminal epithelium and is used in the growth and development of the ruminal papillae and can enhance absorption of various nutrients needed for adequate growth and performance (NASEM, 2016). In the present experiment, steers consuming CB and WS diets had greater ruminal concentrations of butyrate. Branch-chained VFA’s are usually synthesized from branch-chained amino acids and are beneficial for fibrolytic bacterial growth (Allison et al., 1962). Steers consuming the CB and CS diets had the greatest concentration of valerate in the rumen. If fibrolytic bacteria require valerate for growth and function, an increased concentration may indicate a decrease in the population of fibrolytic bacteria or a shift in microbial species that may not require valerate.

According to nutrient composition values (Table 2), solubility of the protein in each of the roughages was 4.6%, 7.0%, and 2.4% for CB, WS, and CS, respectively. As previously stated, the CP of individual roughages were 9.4%, 9.7%, and 5.7% (DM basis) for CB, WS, and CS, respectively. The greater solubility of protein in WS compared to the other roughages may have contributed to the increased degradation within the rumen. Although DM, NDF, and ADF disappearance in the present experiment are low compared to concentrate feed ingredients, previous research has shown that roughage degradation in the rumen is low, regardless of the diet the animal is consuming. Carey et al. (1993) reported that DM and NDF disappearance of hay was approximately 15% after 72 h in steers consuming high hay diets. When the steers were supplemented with SFC in every other day, in situ degradation of medium quality hay and rates of DM and NDF were not affected. Wheat silage roughage had the greatest DM and OM degradation in the present experiment, whereas CB-R and CS-R were similar in degradation percentages. Warner et al. (2020) had similar DM and NDF degradation rates of CB. They reported that DM, NDF, and OM degradation percentages of CB-R in the rumen were 15.7%, 10.2%, and 12.0 %, respectively, after a 96-h ruminal incubation period. In the present experiment, CB-R DM, NDF, and OM percentage of degradation were 17.1%, 6.0%, and 9.2%, respectively. Microbial degradation may be limited due to the change in microbial populations, lignin concentration, or processing methods (Olubobokun et al., 1990; Carey et al., 1993).

In situ degradation of the individual roughages is included in Table 9. Soluble and potentially degradable DM fraction (a + b) of the roughage was greatest for WS-R (P ≤ 0.01) and similar for CB-R and CS-R (P = 10). The DM fraction of roughage that was undegradable (c) within the rumen was greatest for CB-R and CS-R (P ≤ 0.01) and least for WS-R. Lag time was greatest for CS-R and WS-R (P ≤ 0.01) and least for CB-R. Soluble NDF fraction of the roughage (a) was greatest for WS-R (P = 0.02), and similar for CB-R and CS-R (P = 0.32). The ruminally soluble and potentially degradable (a + b; P = 0.08) and undegradable (c; P = 0.08) NDF fractions tended to differ among roughages. Lag time for NDF was greatest for WS-R (P ≤ 0.01) and similar for CB-R and CS-R (P = 0.77).

Table 9.

In situ dry matter (DM) and neutral detergent fiber (NDF) degradation for individual roughages

	Roughage1
Item2	CB	WS	CS	SEM	P- value
a + b, % DM	18.33b	34.68a	17.67b	2.92	<0.01
c, %	81.67a	65.31b	82.33a	2.92	≤ 0.01
Lag, h	4.13b	22.60a	23.52a	1.33	≤ 0.01
Kd, −h	0.0039	0.0031	0.0042	0.000785	0.64
a + b, % NDF	11.18	20.48	16.80	2.50	0.08
c, %	88.82	83.20	79.52	2.80	0.08
Lag, h	10.36b	16.06a	9.89b	1.60	≤ 0.01
Kd, −h	0.00865	0.01085	0.0386	0.01975	0.56

¹Individual roughages ground through 2-mm sieve; CB-R, cotton burrs, WS-R, wheat silage, CS-R, corn stalks. Roughages were placed in the rumen of steers consuming respective experimental diet for 0, 12, 24, 48, and 96 h.

²a + b, soluble fraction determined by disappearance at 0 hr plus rumen degradable fraction, c, undegradable fraction determined by remaining fraction 96 hr, lag time, discrete lag time in h., Kd, fractional rate of degradation.

^a,bMeans with unique superscripts differ when P≤ 0.05.

Although DM and NDF in situ degradation in the present experiment are low compared to concentrate feed ingredients, previous research has shown that roughage degradation in the rumen is low, regardless of the diet the animal is consuming. Carey et al. (1993) reported that DM and NDF disappearance of hay was approximately 15% after 72 h in steers consuming high hay diets. When the steers were supplemented with SFC every other day, in situ rates of DM and NDF degradation of medium quality hay were not affected. Wheat silage roughage had the greatest soluble and potentially degradable DM and least undegradable fractions in the present experiment. Corn stalk and CB-R had similar soluble plus potentially degradable and undegradable fractions. According to the nutrient composition of the individual roughages, WS-R had 79.9% and 59.5% greater concentration of non-fiber carbohydrates compared to CB-R and CS-R, respectively. Nonfiber carbohydrates are defined as organic acids, monosaccharides, oligosaccharides, starches, and other carbohydrates exclusive of hemicellulose and cellulose found in NDF. The increase in nonfiber carbohydrates in WS-R could explain the increased soluble and potentially degradable fraction. Cotton burrs had 45.1% and 64.5% greater OM than WS-R and CS-R, respectively, and a greater percentage of lignin. Based on the nature of cotton burrs, they may contain a higher dirt, debris, and mineral content than most roughages, and this may have contributed to the greater undegradable fraction. Cotton burrs also had the greater percentage of lignin that would have left the rumen undegraded regardless of rumination time or processing method.

Implications

Understanding that the impacts roughage has the ruminal fibrous mat is critical for optimizing roughage source and inclusion levels in finishing diets to improve feedlot performance, feed efficiency, and carcass characteristics of finishing cattle. Results in the present experiment indicate that when cotton burrs are utilized as the roughage source, the cotton-like physical characteristic may contribute to larger particle sizes and homogenous mixing within the rumen that aids in distributing various ingredients throughout the rumen, whereas steers consuming corn stalks or wheat silage as their main roughage source had smaller particle sizes that may aid in increasing passage rate. Increased rumination per kg peNDF for steers consuming corn stalks and increased rumination per kg of NDF for steers consuming wheat silage contributed to the decrease in the percentage of larger particles in the rumen content. Results from the present experiment indicate that individual roughages have unique physical and chemical characteristics that contribute individually to formation of a ruminal fiber mat, alter rumination time, and fermentation characteristics and thus may benefit the health of the rumen. Understanding how these physical and chemical characteristics contribute to the rumen health may benefit feedlot performance, efficiency, and carcass characteristics.

Glossary

Abbreviations

ADF

acid detergent fiber

d

day

DM

dry matter

DMI

dry matter intake

NDF

neutral detergent fiber

NH₃

ammonia

OM

organic matter

peNDF

physically effective NDF

PSPS

Penn State particle separator

RF

rumen fiber mat

RF:DMI

rumen fiber mat to DMI ratio

VFA

volatile fatty acids