Animals and experimental design

The study was conducted at the University of Saskatchewan Livestock and Forage Centre of Excellence (Clavet, SK, Canada). Eight Simmental yearling heifers (328 ± 33.1 kg) were purchased from a single source, halter trained, and surgically fit with a 7.6-cm ruminal cannula (model 3C; Bar Diamond Inc., Parma, ID). Approximately 2 weeks post-surgery, sutures were removed, and the 7.6-cm cannula was replaced with a 10-cm cannula (Model 9C; Bar Diamond Inc.). Heifers were housed in individual tie stalls (1 m × 2 m) fit with water mattresses on the floor, automated water bowls, and an individual feed manger. Heifers were provided access to an outdoor exercise yard daily for approximately 2 h during feed preparation except during sampling periods or in cases of extreme weather (<–30 °C).

Prior to the start of the study, cattle were gradually transitioned (five intermediate diets over 48 days) to a high-grain diet composed of barley silage, barley grain, canola meal, and a barley-based mineral and vitamin supplement. The amount of feed offered during the diet transition was restricted to 2.3% of body weight (BW) on a DM basis. Heifers were abruptly switched to their treatment diet at the start of the study. Abrupt diet transitions were also used at the start of each period.

The study was conducted as a replicated 4 × 4 Latin square with a 2 × 2 factorial arrangement of treatments utilizing 21-day periods. Each square had a unique treatment sequence and sequences were balanced for carry-over effects. Diets (Table 1) were formulated using the Beef Cattle Nutrient Requirements Model (NASEM 2016) and included (DM basis) 15% barley silage, 6% of a mineral and vitamin supplement using barley grain as a carrier, 3% dry distillers’ grains with solubles, and 76% hybrid rye grain (KWS Bono, FP Genetics, Regina, SK, Canada). The hybrid rye grain was processed using dry rolling or temper-rolling methods with a narrow (0.8 mm) or wide (1.23 mm) roller gap width. The roller mill (SvenMill 4, Apollo Machine and Products, Saskatoon, SK, Canada) had two 17.8-cm diameter rollers with 6.3 grooves/cm and a groove depth of 2.4 mm. Scrapers on each roller were used to prevent grooves from plugging. The gap width of the rollers was tested repeatedly for dry-rolled hybrid rye grain until achieving the target densities prior to the start of the study. Once achieved, the roller gap width was documented, and the roller gap remained static for the duration as a single lot of hybrid rye was used for the duration of the study. The same gap widths were used for dry-rolled and temper-rolled grains as described by Nixdorff et al. (2020). For tempered hybrid rye, water was added to achieve a DM of 80%, and the grain was soaked for 24 h prior to rolling. The tempered hybrid rye was processed in batches every 2 days. A single bunker-style pit of barley silage was used, and a single lot of all other ingredients was sourced from the Canadian Feed Research Centre (North Battleford, SK, Canada; dry distillers’ grains with solubles and vitamin, and mineral supplement).

Table 1.

Diet ingredients and chemical composition of the method (dry-rolled or tempered) and severity (coarse or fine) of hybrid fall rye processing as a grain source for cattle.

The resulting severity of processing was measured using a 1.18 mm sieve and a pan from the Penn State Particle Size Separator (Kononoff et al. 2003) to determine the percentage of fine particles. In addition, the processing index (PI) was calculated, in duplicate, by measuring the weight of the processed sample in a 500 mL container and dividing that weight by the weight of the unprocessed sample filling the same container (Beauchemin et al. 2001). The PI (expressed in OM and DM) and percentage of fine particles (<1.18 mm) for each batch processed (dry-rolled coarse, dry-rolled fine, tempered coarse, and tempered fine) were recorded throughout the study.

Kernel characteristics (width, thickness, and length) were determined using 10 randomly selected kernels from each source in each period according to Zinn (1993) and Johnson et al. (2020). Amyloglucosidase reactive soluble starch (AGR) and amylase reactive insoluble starch (ARIS) were analyzed according to Zinn (1990) and Rodríguez et al. (2001). Briefly, dried samples were ground to pass through a 1-mm screen. For AGR, 0.5 g was incubated in an acetate-based (24 mmol/L) buffer solution containing toluene and amyloglucosidase (26 IU/sample; Sigma-Aldrich, Oakville, ON, Canada). Samples were incubated at 39 °C for 4 h with mixing occurring every 20 min. Immediately after incubation, samples were placed in an ice bath for 10 min, 2 mL of a zinc sulfate solution (15% wt/vol) was added, and samples were filtered. To each sample, 4 mL of o-toluidine solution (0.6 mol/L) was added and the amount of glucose released was determined using a photospectrometer at a wavelength of 630 nm. The amount of glucose released was determined using a colorimetric assay. The ARIS was determined using 0.15 g of sample and enzymatic digestion using amyloglucosidase (26 IU/sample), α-amyloglucosidase (240 IU/sample), porcine pancreatin (8 mg/sample), yeast-based lyticase (400 IU/sample), and protease from Bacillus licheniformis (2 mg/sample) in a phosphate buffer solution. Samples were incubated at 39 °C for 6 h with mixing occurring every 30 min. After 10 min in an ice bath, 2 mL of zinc sulfate solution was added, and samples were maintained at room temperature for 10 min. Samples were then filtered, 4 mL of o-toluidine solution (0.6 mol/L) was added, heated at 100 °C for 10 min, followed by 5 min in an ice bath, and 10 min at 27 °C. Absorbance was then read at 630 nm. Values for AGR and ARIS were then used in calculations described by Corona et al. (2006). Insoluble reactive starch (IRS, %) was calculated as ((ARIS – AGR)/6), where 6 represents the number of hours of in vitro incubation. Insoluble starch that was potentially digestible (ISD, %) was calculated as (100 – AGR) × ((IRS)/(IRS + 0.06)), where 0.06 represents the passage rate of grain from the rumen as estimated by Ørskov and McDonald (1979). Predicted ruminal starch digestion (PRSD, %) was calculated as (1.32 × AGR) + (0.93 × ISD), with the maximum value set at 100% according to Johnson et al. (2020). In addition, 7 h in vitro starch digestibility was determined at Cumberland Valley Analytical Services (Waynesboro, PA) as described by Johnson et al. (2020).

Throughout the study, heifers were fed to achieve ad libitum intake with fresh feed provided once daily at 09:00 h and were allowed ad libitum access to water. Refusals (targeting 5%–10% relative to the weight of the feed offered) were removed, weighed, and collected daily at 08:00 h. While the amounts of feed offered and refused were recorded daily, data used for statistical analysis were restricted to those measured from day 16 to day 21 of each period. Refusal samples from day 16 to day 21 within each sampling period were composited by heifer proportionally to the amount refused each day. Representative feed ingredients samples of silage (1 kg/d), hybrid rye (0.5 kg/d), dry distillers’ grains with solubles (0.5 kg/d), and mineral (0.5 kg/d) were collected daily from day 16 to day 21 at 08:00 h and were composited by ingredient. The hybrid rye samples were collected from each processing treatment and the original unprocessed grain source. Composited feed ingredient and refusal samples were dried in a forced-air oven at 55 °C for 72 h to determine the DM concentration. The DMI was then calculated based on the amount of feed offered and refused when corrected for DM concentration. In addition, silage DM was measured twice weekly, and the DM of all other ingredients was measured once weekly. Dietary feed ingredient inclusion rates (as-fed basis) were updated to reflect most recent DM of ingredients. The dried feed ingredient and refusal samples were ground using a hammer mill to pass through a 1 mm sieve (Christy and Norris Ltd, Chelmsford, UK) and stored until being analyzed. Water was supplied ad libitum.

The particle size distribution of each diet was determined in duplicate using representative 1 L samples of each ingredient with the Penn State Particle Separator (Nasco Education, Fort Atkinson, WI) according to Heinrichs (2013). The particle size separator included sieves with aperture sizes of 19, 8, and 4 mm and a pan; the distributions are reported in Table 1. Particle size distribution of refusals were also determined using the composited refusal sample for each heifer within each period. All particle size measurements were conducted on an as is basis. The sorting index for particles retained on each sieve was calculated as (actual nutrient intake)/(theoretical nutrient intake) × 100%, as described by Leonardi and Armentano (2003). Particle sorting values equal to 100 indicate no sorting, those <100 indicate selective refusals (sorting against), and those >100 indicate preferential consumption (sorting for).

Ruminal fermentation

Indwelling ruminal pH measurement was initiated on day 17 of each period and proceeded for the subsequent 96 h as described by Penner et al. (2006). Indwelling ruminal pH measurement systems (Dascor, Escondido, CA) were placed in the ventral sac of the rumen and were maintained in that location using 2 kg of weight. The pH systems were programmed to record a measurement every 5 min. Prior to insertion into the rumen and following removal, the pH systems were standardized in pH buffers 7 and 4 at 39 °C. The raw data collected during measurement were transformed from mV recordings to pH using beginning and ending linear regressions derived from the starting and ending standardizations while accounting for drift that was assumed to be linear over time (Penner et al. 2006). The pH values were then used to calculate the minimum, mean, and maximum pH along with the duration and area that pH was <5.5.

Samples of ruminal digesta (250 mL/region) were collected from three regions of the rumen (cranial, central, and caudal) at the ruminal fluid–ruminal mat interface beginning at 07:00 h on day 18 and every 12 h thereafter with a 3 h offset on subsequent days until day 21. The sampling approach resulted in a total of eight samples collected for each heifer that represented every 3 h of a 24 h feeding cycle. The digesta from each region were pooled and samples were strained through two layers of cheesecloth. The resulting ruminal fluid was mixed, and two 10 mL aliquots were collected. One 10 mL sample was preserved in 2 mL of metaphosphoric acid (25% wt/vol) for measurement of SCFA concentration (Khorasani et al. 1996) and the other was preserved in 2 mL of 1% sulfuric acid for ammonia-N analysis (Broderick and Kang 1980). These samples were sealed and stored at –20 °C until analysis.

Apparent total tract digestibility and chemical analysis

Titanium dioxide (TiO₂) was used as a marker to predict fecal output and was included at 0.2% of dietary DM from day 13 to day 20 of each period. Fecal samples (200 g) were collected from the rectum beginning at the same time as ruminal digesta collection. Fecal samples were dried in a forced-air oven at 55 °C to a constant weight and were ground to pass through a 1-mm screen using a hammer mill (Christie-Norris Laboratory Mill, Christie-Norris Ltd, Chelmsford, UK). All dried and ground feed, refusal, and fecal samples were analyzed for DM, organic matter (OM), crude protein (CP), neutral detergent fiber measured using α-amylase, sodium sulfite, and corrected for ash content (aNDFom), starch, ether extract, calcium, and phosphorus, and the NEm and NEg were calculated at Cumberland Valley Analytical Services (Waynesboro, PA) as described by Pereira et al. (2021). Samples were also analyzed for titanium concentration according to Myers et al. (2004).

Statistical analysis

Heifer was considered as the experimental unit (n = 8). The model for all variables included fixed effects of method (dry-rolled vs. tempered), severity (coarse vs. fine), and the method × severity interaction. Heifer, period, square, and the period × treatment interactions were included as random effects. All data were analyzed using the PROC MIXED procedure of SAS (SAS Institute, Inc., Cary, NC) with Tukey's test to compare means. Tests for normality (Shapiro–Wilk and Kolmogorov–Smirnov) and heterogeneity of treatment variances (GROUP option of SAS) were performed before analyzing data. Results were considered significant when the P-value was ≤0.05 and trends were considered when 0.05 < P < 0.10.

Kernel characteristics, starch reactivity, and 7 h in vitro starch digestibility

Interactions were observed (P ≤ 0.01) between method and severity of processing for all kernel characteristics evaluated (Table 2). For PI (whether reported as a %OM or %DM), tempered grain had greater PI values than when dry-rolled, regardless of severity, and increasing the severity of processing had a greater effect to decrease the PI for dry-rolled than tempered grain. The percentage of fine particles (<1.18 mm) was greater when dry-rolled than when temper-rolled, regardless of the severity of processing, and increasing the severity of processing increased percentage of particles <1.18 mm for dry-rolled hybrid rye only. Kernel width was greater when tempered than dry-rolled and increasing the severity of processing increased kernel width with tempering but not when dry-rolled. The thickness of processed kernels was also greater when tempered than dry-rolled, regardless of processing severity, and increasing the severity of processing decreased thickness to a greater extent for tempering than for dry rolling. Kernel length was greater for tempered than dry-rolled hybrid rye, and increasing the severity of processing reduced the length for dry-rolled but not for tempered hybrid rye.

Table 2.

Effect of method (dry-rolled vs. tempered) and severity (coarse vs. fine) of rolling hybrid fall rye on kernel characteristics, starch reactivity, and 7 h starch digestibility.

The AGR (Table 2) was less for tempered vs. dry-rolled hybrid rye but was not affected by the severity of processing. That said, there was a tendency for an interaction between method × severity (P = 0.06) where there was a greater increase in AGR for dry-rolled hybrid rye with increased processing than for tempered hybrid rye. While ARIS was not affected by method, severity, or the two-way interaction, tempering hybrid rye increased IRS relative to when dry-rolled, and increasing the severity of processing tended (P = 0.06) to decrease IRS. The ISD was increased with tempering and tended to be reduced as the severity of processing increased. As a result, the PRSD was less for tempered than dry-rolled hybrid rye grain but was not affected by severity or the method × severity interaction. The 7 h in vitro starch digestibility was not affected by the method of processing (P = 0.92), although increasing the severity of processing increased starch digestibility by 3.9% (P < 0.01).

Dry matter intake, particle sorting, and ruminal fermentation

Initial BW and final BW were not affected by method, severity, or the interaction between method and severity (P ≥ 0.06; Table 3). Heifers fed tempered hybrid rye had greater (P < 0.01) DMI when compared to cattle fed dry-rolled hybrid rye, but the processing method had no effect on DMI. There was no effect of the method, severity, or the interaction between method and severity (P ≥ 0.29) on the sorting index for particles retained on sieves with aperture openings of 19.0 and 8.00 mm. However, heifers fed dry-rolled hybrid rye sorted more against (P < 0.01) particles retained on the 4.0 mm sieve than those fed tempered hybrid rye. Cattle fed finely processed hybrid rye increased their sorting against (P < 0.01) particles <4.0 mm than those fed coarsely processed hybrid rye.

Table 3.

Effect of method (dry-rolled or tempered) and severity (coarse or fine) of rolling for hybrid fall rye on body weight (BW), dry matter intake (DMI), ruminal fermentation, and particle sorting of ruminally cannulated beef heifers (n = 8).

Mean, minimum, and maximum ruminal pH and the duration that ruminal pH was <5.5 were not affected by method, severity, or the interaction between method and severity (P ≥ 0.12; Table 3). Cattle fed dry-rolled hybrid rye had a greater (P = 0.05) area that ruminal pH was <5.5 when compared to heifers fed tempered hybrid rye with no effect of severity of processing. Total SCFA concentration and the molar proportions of individual SCFA and ruminal ammonia were not affected by method, severity, or the interaction between method and severity (P ≥ 0.23).

Apparent total tract nutrient digestibility

Heifers fed dry-rolled hybrid rye had greater DM (P < 0.01), OM (P < 0.01), and CP (P = 0.01) digestibility when compared to those fed tempered hybrid rye (Table 4). Increasing the severity of processing did not affect apparent total tract DM, OM, or CP digestibility (P ≥ 0.12). There was no effect (P ≥ 0.51) of the method, severity, or the interaction between method and severity on apparent total tract digestibility of aNDFom and ether extract. An interaction was observed (P = 0.03) between the method and severity of processing, where increasing the severity of processing for dry-rolled hybrid rye had no effect on starch digestibility but increasing the severity of processing for temper-rolled hybrid rye increased starch digestibility and resulted in values that were not different from those for dry-rolled hybrid rye.

Table 4.

Effect of method (dry-rolled or tempered) and severity (coarse or fine) of hybrid fall rye on apparent total tract nutrient digestibility of ruminally cannulated beef heifers (n = 8).