Effects of complete substitution of dietary grain and protein sources with by-products on the production performance of mid-lactation dairy cows fed diets based on barley silage under heat stress conditions

This study evaluated the effect of replacing cereal grains and soybean meal with by-products (BY) on production performance, nutrient digestibility, ruminal fermentation, nutrient recovery, and eating and chewing behavior of moderate-producing dairy cows under heat stress conditions. Twelve multiparous Holstein cows (116.7 ± 12.01 d in milk; 42.7 ± 5.06 kg/d of milk yield; 665 ± 77 kg of body weight; mean ± SD) were used in a replicated 3 × 3 Latin square with 28-d periods (21 d for diet adaptation and 7 d for sampling and data collection). Cows were fed a total mixed ration containing a 39.2:60.8 ratio of forage to concentrate throughout the experiment. All diets were formulated to be isoenergetic and isonitrogenous, with different concentrates. Diets were 1) control diet based on cereal grains (ground corn and ground barley, plus SBM; CON), 2) sugar-rich BY (beet pulp, citrus pulp, and liquid molasses, plus canola meal; S-BY-CM), 3) cereal grain BY (rice bran, corn germ meal, wheat bran, barley sprout, and broken corn; CG-BY). Our results showed that replacing grains with BY increased neutral detergent fiber intake and digestibility, whereas starch intake, human edible energy, and human edible protein decreased. Milk yield and dry matter intake decreased in cows fed with the CG-BY diet compared with the other 2 treatments. In contrast, there were no significant differences observed between the CON and S-BY-CM diets in terms of milk yield and dry matter intake. The S-BY-CM diet increased energy-corrected milk production compared with the CG-BY diet (36.2 vs. 34.3 kg/d) but CG-BY enhanced feed conversion efficiency compared with the other 2 treatments. Although the S-BY-CM diet prolonged the eating and sorting of small particles, neither of the dietary treatments affected chewing activity or ruminal pH 4 h after feeding. Furthermore, both diets containing BY contributed to an increase in milk fat content in comparison to the CON group. Additionally, the CG-BY and S-BY-CM diets demonstrated greater performance than the CON diet in terms of human-edible feed conversion efficiency for protein and energy. The results indicated that S-BY-CM can completely replace barley and corn grain in the diet of mid-lactating dairy cows exposed to heat stress conditions without any negative effect on production and ruminal pH. However, the inclusion of CG-BY impaired dry matter intake, milk yield, and digestibility of nutrients and is not recommended during heat stress conditions.


INTRODUCTION
By-products (BY) from different food industries are widely available, promoting interest in animal nutrition as substitutes for cereal grains (Bradford, 1999, Eastridge, 2006, Naderi et al., 2022).Genetic advancements for higher milk yields in modern dairy cows (Oltenacu and Algers, 2005) and challenge of poor or medium quality forage harvested in some countries (Nemati et al., 2020) have driven nutritionists to formulate starchrich diets to meet the energy needs of high-producing cows.However, the use of these grain-based diets with low NDF from forage can bring adverse effects, such as acidosis (Plaizier et al., 2008), poor NDF digestibility, and milk fat depression (NRC, 2001).Another disadvantage associated with using grains in ruminant nutrition is that grains like corn and barley are humanedible ingredients, and their reduction in feeding ruminants has recently become more important to increase net food production, defined as human-edible livestock products (output) minus human-edible feeds (input; Ertl et al., 2016).
Prolonged exposure of dairy cows to hot weather is to blame for approximately 50% of milk production Effects of complete substitution of dietary grain and protein sources with by-products on the production performance of mid-lactation dairy cows fed diets based on barley silage under heat stress conditions loss due to reduced voluntary DMI (Wheelock et al., 2010).A practical approach to boost heat-stressed cow performance involves optimizing nutrient utilization through enhanced digestion.While lowering dietary fiber and increasing concentrates is a common practice to increase fDMI, partly due to the decreased heat increment (Reynolds et al., 1991), dairy cows exposed to heat stress are at a higher risk of developing ruminal acidosis (Sammad et al., 2020), which can be exacerbated by low-fiber, high-starch diets (Drackley et al., 2003).
To sustain net food production and address heat stress impacts and nutrition-induced acidosis in dairy cows, a viable approach involves replacing high-starch grains with low-starch, highly digestible, and energyrich BY.However, while starch digestibility is significantly higher than NDF (almost twice), substituting NDF (from BY) for starch (from barley and corn) has the potential to decrease the energy concentration of the diet (Weiss et al., 2009).This decrease in energy concentration may, in turn, have a negative impact on milk yield during heat stress.However, when grains are replaced with BY, not only is the starch content reduced, but the forage NDF is also concurrently diminished due to the low-forage-based diets.In such cases, if replacing starch with NDF reduces digestibility, it could potentially increase DMI as high-fiber BY rapidly pass from the rumen (Bradford and Mullins, 2012).Therefore, the increase in DMI as a result of the rapid passage rate of less bulky BY can potentially maintain the energy intake, mitigating any negative effects on milk yield (Bradford and Mullins, 2012).Previous studies (Kanjanapruthipong et al., 2015;Halachmi et al., 2004) reduced forage fiber by substituting roughage NDF from rice straw or corn silage with soy hulls or cassava residues, attributing the benefit to extended meal length and size, indicating cows allocate more time to rest under severe heat stress.Few studies incorporated both low forage NDF and starch levels in the diet of heat-stressed dairy cows.For example, Heydari et al. (2021) found that simultaneously reducing dietary forage NDF and starch, achieved through partial replacements of corn silage and ground barley grain with beet pulp up to 24% (DM), could benefit lactating dairy cows through greater DMI and lower signs of heat stress such as respiration rate.
The other advantage of using BY instead of grains in a low-forage-based diet relates to the prevention of ruminal acidosis and milk fat depression.Although both BY and grains have a small particle size, BY-based diets contain less starch.In addition, BY contains a greater amount of NDF, which is approximately half as effective as forage NDF at maintaining ruminal function and milk fat yield (Swain and Armentano,199).Krause and Oetzel (2006) found that including fiber-rich BY in dairy cow diets lowered the risk of acidosis.Malekkhahi et al. (2023) found a dietary starch-fiber interaction on milk fat in which BY-based diets increased milk fat yield and concentration with low-starch (22% DM) compared with high-starch diets (32% DM), indicating the significance of not only physically effective forage NDF but also the dietary starch content in influencing ruminal health, DMI, and milk production.Low-forage diets with starch levels as low as 12 to 17% were formulated by substituting cereal grains with high-fiber, low-starch BY feedstuffs (Mahjoubi et al., 2009, Naderi et al., 2022, and Rezac et al., 2012).However, these studies were conducted to lower starch and increase NDF in diets.Hence, extrapolating these findings to low-starch, low-forage fiber diets with increased sugar and fat levels due to cereal grain replacement with BY during heat stress might not be appropriate.There is limited research into the performance of heat-stressed dairy cows fed such low-starch, low-forage fiber diets.
Cereal grain BY, including rice bran, corn germ meal, wheat bran, and barley malt sprout, offer costeffective energy alternatives in livestock diets.Rice bran typically contains 11-14% CP, 16-21% NDF, and 12-18% EE (dos Santos Oliveira et al., 2011, Criscioni andFernández, 2016).Corn germ meal features 20-23% CP, 37-54% NDF, and 16-18% EE (Widmer et al., 2008, Lakshmi et al., 2017), while wheat bran offers 16-18% CP, 53% NDF, and 2-5% EE (Šramková et al., 2009).Barley malt sprout typically includes 25-30% CP, 42.7-48.8%NDF, and 2.4-2.8%EE (Šidagis et al., 2013).These BY, enriched with digestible NDF, CP, and EE, offer an alternative energy or protein source compared with grains and soybean meal (SBM).Sugar-rich BY like sugar beet pulp, dried citrus pulp, and beet molasses are widely used in livestock feed for their high energy content.Sugar beet pulp typically holds 35.5-44.5% NDF, 23% pectin, 50-60% sugar, and 7-10% CP (Naderi et al., 2016, NRC, 2001).Dried citrus pulp has 25% NDF, 25% pectin, 23% sugar, and 3-6% CP (López et al., 2014).Beet molasses contains 45.2-62% sugar and 3-6% CP (Palmonari et al., 2020).While these BY offer significant dietary energy through their degradable NDF, pectin, and sugar content, they are considered poor sources of protein and must be supplemented with a protein source such as canola meal or SBM.Several studies have demonstrated that replacing SBM with canola meal in dairy cow diets does not negatively affect milk production (Huhtanen et al., 2011, Martineau et al., 2013).Given that a significant portion of feed resources, such as grains and SBM, could be more efficiently used in human foods, or for poultry and pig production, it is essential to find cost-effective and high-quality protein supplements for dairy cows.In Erfani et al.: Effects of complete substitution… addition, both SBM and canola meal are crucial protein sources for ruminants in Iran, but SBM is significantly more expensive due to its higher demand from the dairy and poultry industries.Therefore, there is a need for alternative protein sources to reduce dependence on SBM in dairy cattle nutrition and potentially alleviate concerns about the shortage of SBM for humans and monogastric animals.
Consequently, a reduction in starch content in BY diets can be compensated by incorporating fat-rich ingredients and digestible fiber for cereal grains BY or by introducing a combination of pectin, sugar, and digestible fiber for sugar-rich BY.We hypothesized these energy sources would not impair milk yield and DMI in mid-lactation dairy cows during heat stress, while maintaining ruminal pH and net food production.We aimed to study the impact of different BY sources in low-forage diets on net food production, yields of milk and milk components, apparent total-tract digestibility of nutrients, and ruminal pH coupled with monitoring the eating, ruminating, and feed-sorting behavior of heat-stressed dairy cows.

MATERIALS AND METHODS
The experiment was carried out at the Lavark Research Station's dairy facility (Isfahan University of Technology, Isfahan, Iran).The experiment received approval from the Institutional Animal Care Committee for Animals Used in Research, and animals were cared for in accordance with the Iranian Council of Animal Care (1995) guidelines; the protocol number of the ethics committee was #2020/B48.

Animals, Experimental Design, and Treatments
Twelve multiparous Holstein cows (DIM = 116.7 ± 12.01; milk yield = 42.7 ± 5.06 kg/d; BW = 665.6 ± 77.02 kg; parity = 2.5 ± 1.24) were randomly assigned in a replicated 3 × 3 Latin square design with 28-d periods (21 d for diet adaptation and 7 d for sampling and data collection).Each cow was housed in a sandbedded box stall (4 × 4 m) and a roofed barn for the duration of the trial.Each box stall had automated water troughs and a concrete feed bunk.
The experimental treatments (Table 1) differed in the type of concentrate derived from BY, with the same forage sources.The control diet (CON) was based on cereal grains (ground corn and ground barley) and SBM, whereas the BY-based diets included: (1) sugar-rich feedstuffs (beet pulp, dried citrus pulp, and molasses) and canola meal (S-BY-CM), or (2) cereal grain by-products (rice bran, corn germ meal, wheat bran, barley sprout, and broken corn; CG-BY).Throughout the experiment, diets fed as TMR comprised 60.8% concentrate, 24% barley silage, 12% alfalfa hay, and 3.2% wheat straw (on a DM basis).The CG-BY ingredients were purchased from Shahd Zagros Jahanbin Co., Shahre-kord, Chaharmahal VA Bakhtiari, Iran.The S-BY-CM ingredients were purchased from Ramsar Citrus Gardeners Co. (Ramsar, Mazandaran).Both sets of BY components were purchased in a single shipment to avoid any variability in composition.The diets were formulated using the Cornell Net Carbohydrate and Protein System (version 5.0;Fox et al., 2000) for dairy cows producing 45 kg/d of milk with 650 kg BW, 3.2% fat, and 3.0% CP, and were isocaloric and isonitrogenous.The barley, corn, and beet pulp were ground using a hammer mill with a 3-mm sieve (model 5543 GEN; Isfahan Dasht, Isfahan, Iran), whereas other ingredients had fine textures and were not ground.The TMR for each animal was individually weighed and then mixed manually in the troughs once a day and offered twice daily (0900 h and 1600 h) in sufficient quantities to ensure 5 to 10% orts.After morning milking, half of the prepared TMR was offered to the cows, while the other half was stored in a large sealed container under a roof until evening milking.The objective behind preparing the TMR once daily and providing it twice daily was to mitigate the risk of overheating and spoilage of the ration in the troughs.The chemical composition and estimated human-edible fraction of each ingredient according to Wilkinson (2011) are shown in Table 1.
The dietary energy precursors were changed from more glucogenic to more lipogenic as a result of substitution of cereal grains with BY from different sources.In comparison to CON,starch (19.5 and 9.5% vs. 32.5%),non-fiber carbohydrate content (33.4 and 35.1 vs. 37.4%), human edible energy (HEE) (2.06 and 1.84 vs. 7.67 Mcal of gross energy/ kg of DM), and human edible protein (HEP) (2.0 and 1.9 vs. 7.9% of DM) were lower for CG-BY and S-BY-CM, respectively, vs. CON, whereas NDF concentrations (42.6 and 39.1% vs. 31.8%)increased for CG-BY and S-BY-CM vs. CON.Given the lower starch content of sugar-rich compounds, a fat supplement was added to the S-BY-CM treatment to compensate for the energy deficiency of this group, while a fat supplement was also added to the other 2 dietary treatments to prevent the effect of the fat supplement on animal responses.Accordingly, by adding fat supplements to treatments, the dietary fat concentration increased from 4.3 to 5.7% on a DM basis due to the inclusion of fat-rich ingredients, including rice bran, wheat bran, and corn germ meal in the CG-BY diet.In addition, the sugar content in the S-BY-CM group increased from 4.1 to 13.7% due to the inclusion of sugar-rich feedstuffs.Therefore, the  1).

Feed Intake and Nutrient Digestibility
To measure DMI for each cow, the offered TMR and refused feed amounts were measured and sampled daily from d 21 to d 28 of each period.All samples were kept in a freezer with a temperature of −20°C until the end of the collection period, at which time they were com-bined by cow, sub-sampled, and then stored at −20°C for further analysis.Following thawing, the DM content of representative samples was measured by drying for 48 h at 60°C in an oven with forced air (Townshend, 1995).A Wiley mill with a 1-mm screen was used to grind all samples (Arthur H. Thomas, Philadelphia, PA) for analysis of CP (AOAC International, 2006;method 955.04), ether extract (AOAC International, 2006;method 920.39), ash (AOAC International, 2006; method 942.05), aNDF using heat-stable α-amylase (Mertens, 2002) and sodium sulfite, and ADF (Van Soest et al., 1991).The non-fibrous carbohydrate content was calculated as 100 − (CP + NDF + ether extract + ash) according to NRC (2001).To determine the starch concentration in both feed and fecal samples, we utilized a modified glucoamylase procedure as described by Zhu et al. (2016).First, samples were gelatinized by boiling them in a water bath for 20 min before being allowed to cool to ambient temperature.Once the samples had cooled, acetate buffer was added to them, and the starch was hydrolyzed using glucoamylase.Next, glucose concentration was measured using glucose oxidase with commercial kits (Pars Azmoon Co., Tehran, Iran) according to the manufacturer's instructions.Fecal grab samples were collected from each cow for 4 consecutive days at 12-h intervals so that 8 samples were taken from each cow each period (Cooke et al., 2008).Sampling time was forwarded by 2 h each day, so samples were collected at 0500-1700, 0700-1900, 0900-2100, and 1100-2300 h.The samples were combined by cow and frozen at −20°C until analysis.After thawing, fecal samples were dried in a forced-air oven (Binder Avantgarde Line FD series Hot Air Oven, Binder GmbH, Germany) at 60°C for 72 h, milled to pass through a 1-mm screen, and analyzed for DM, ash, ADF, aNDF, starch, and CP by procedures previously described.Apparent total-tract digestibility of nutrients was determined (in triplicate) using undigested NDF after 288 h of in situ ruminal incubation as an internal marker in 2 ruminally cannulated nonlactating Holstein dairy cows fed a high-forage TMR diet (25% barley silage, 10% alfalfa hay, 40% wheat straw, 25% concentrate mix composing BY; DM basis) in accordance with the recommendations of Bender et al. (2016) and Kahyani et al. (2019).Before incubation, this diet was fed to the cows for 2 wk.

Particle Size Measurement, Sorting Behavior and Chewing Activity
The amounts of fresh TMR by treatment and amounts of refusals for each cow were collected and recorded daily during the 8-d collection period for analysis of particle size distribution.The Penn State Particle Separator (PSPS), equipped with 3 sieves and a bottom pan, was used to quantify particle size in triplicate (Kononoff et al., 2003).After sieving, samples were dried in a forced-air oven (Binder Avantgarde.Line FD series Hot Air Oven, Binder GmbH, Germany) at 60°C to measure the DM of each sieved fraction.Physically effective factor (pef) values were obtained by summing the amount of DM that remained on 2 sieves (pef8; Lammers et al., 1996) or 3 sieves of the Penn State Particle Separator (pef1.18;Kononoff et al., 2003; Table 2).The physically effective NDF of 2 and 3 sieves (peNDF8 and peNDF1.18,respectively) were estimated by multiplying the NDF content (DM basis) of the diet by pef8 and pef1.18,respectively.The ASAE (1995) procedures were used to compute the geometric mean particle size of TMR and orts.Sorting of particles was assessed by comparing the actual intake of each fraction to the expected intake of the same fraction if the diet had been consumed as formulated (Leonardi and Armentano, 2003).Sorting index values > 100 suggested that cows sorted in favor of these particles, while values less than 100 indicated sorting against these particles, and a value of 100 indicated that no sorting occurred (Leonardi and Armentano, 2003).Eating and ruminating behaviors were visually evaluated for each cow for 24 h on d 25 of each period.Three individuals took part in the observation, in which a single observer monitored cows continuously for 4 h before being replaced by a new observer.For the entire 24-h period, cows were noted every 5 min, with the observer requiring roughly 1 min to perform observations for all cows, and it was assumed that the activity of each cow lasted the entire 5-min interval between observations (Beauchemin et al., 2003).The sum of ruminating and eating time was used to determine total chewing time.

Ruminal Sampling and Analysis
On the last day of each period (d 28), ruminal fluid samples (approximately 3 mL) were obtained from the ventral sac via rumenocentesis (Nordlund and Garrett, 1994) 4 h after the morning feeding.A portable digital pH meter (HI 8318; Hanna Instruments) calibrated at pH 4 and 7 at the beginning of each testing day was used to immediately determine the pH of the ruminal fluid.After that, 2 mL of the collected sample was acidified with 0.4 mL of 25% meta-phosphoric acid and immediately frozen at −20°C for later VFA analysis.Following thawing and centrifuging ruminal fluid at 10,000 × g at 4°C for 20 min, the Hashemzadeh-Cigari et al. ( 2014) procedure was used to measure the VFA using gas chromatography (Chrompack, model CP-9002; Chrompack International BV).Ruminal NH3-N concentration was measured according to the method described by Broderick and Kang (1980).

Milk Yield and Components, BW, and BCS
During d 21 to 28 of each period, the milk yield of each cow was recorded, and a sample from each cow was taken at each milking.Cows were milked 3 times daily, at 0800, 1600, and 2400 h in a herringbone milking parlor, so that a total of 21 milk samples were obtained per period.Samples were mixed with potassium dichromate, stored at 4°C, and submitted to Ideh Sazan Rojan Alvand Co. (Alborz, Iran).Milk samples were analyzed for fat, true protein, lactose, SNF, TS, MUN, BHB, and free fatty acids (FA) profile, utilizing the advanced Fourier-transform mid-infrared spectroscopy of CombiScope FTIR 600 HP (Delta Instruments).After obtaining the milk component analysis results from the company, composite values were generated by averaging the data obtained from each cow during each period.Calibration adjustments for all FA measures on the milk analyzer were made once a month using the calibration equations preinstalled by the company, and a set of modified milk samples described by Kaylegian et al. (2006) with reference values in FA g per 100 g of milk for each individual or groups of FA measured.Fatty acid yields were calculated as milk fat (g/d) × 0.933 × (total milk fat percentage/100), based on the estimation that total FA account for 93.3% of milk total lipids (Glasser et al., 2007).In general, 3 main categories of FA were defined as follows; de novo FA with chain length < 16 carbon (originated from mammary synthesis), preformed FA with chain length > 16 carbons (originated from blood), and mixed FA (originated from both sources, C16:0 plus C16:1 cis-9 and iso-C16:0).The yield of 3.5% FCM (kg/d) and ECM (kg/d) were calculated based on the following equations (NRC, 2001): FCM (kg/d) = 0.4 × milk yield (kg/d) + 15 × milk fat yield (kg/d); ECM (kg/d) = 0.323 × milk yield (kg/d) + 12.82 × milk fat yield (kg/d) + 7.13 × milk protein yield (kg/d).Feed conversion efficiency (FCE) for the production of milk, 3.5% FCM, and ECM were calculated as the amount of each variable per DMI (kg/kg).
Two skilled scorers determined the BCS (Edmonson et al., 1989) at the beginning and end of each experimental period.Cows were weighed at the beginning of the experiment and at the end of each experimental period, in the mornings after milking.

Environmental Measurements and Calculation of Temperature-Humidity Index
Ambient climate conditions, such as daily maximum and minimum temperatures (T db , °C) and daily relative humidity (RH, %), were measured using a digital thermo-hygrometer (TFA Dostmann Digital Thermo-Hygrometer) for 72 d (from d 12 to the last day of the experiment).Subsequently, the daily average, maximum, and minimum temperature-humidity index (THI) were calculated using the following equations:  Human-edible protein (HEP) and human-edible energy (HEE) were calculated according to recommendations by Wilkinson (2011).
erinaria glass precision thermometer with a prismatic section was used to measure rectal temperatures with a 0.1°C accuracy.Additionally, the cow's respiration rate (measured in breaths/min) was averaged during the last 5 d of each period by measuring the cow's flank movements for 1 min between 1400 and 1500 h.The experimental stalls were outfitted with 4 fans (1 fan per 8 linear m) positioned above the midline of the feeders and the bed.Throughout the experiment, we deliberately turned off the sprinklers and exclusively operated the fans from 0700 to 1900 h on a daily basis, with the intention of inducing heat stress.

Calculations and Statistical Analysis
The fraction of feedstuffs that humans can consume is referred to as human edible.The human edible proportions of each dietary ingredient were used to compute the human-edible protein (HEP) and human-edible energy (HEE) inputs (Wilkinson, 2011).The HEP and HEE outputs comprised the total gross energy and true protein of milk, respectively.The HEP output per HEP input and HEE output per HEE input were used to determine the FCE for HEP and HEE, respectively.The human edible FCE is a term used to describe the ratio of human-edible output derived from animal products to human-edible input in feedstuffs.
Data collected over multiple days were averaged per cow per period before statistical analysis.The SAS PROC MIXED procedure was used to analyze the data as a replicated 3 × 3 Latin square design (version 9.0, SAS Institute Inc., Cary, NC).In the model, square, period within square, and treatment were considered fixed factors, while cow within square was included as a random variable and was the experimental unit.According to the model Y ijkm = µ + T i + P j + C k + S m + E ijkm , where Y ijkm was the observation, µ was the overall mean, T i was the fixed effect of the treatment, P j was the fixed effect of the period, C k was the random effect of the cows, S m was the fixed effect of the square, and E ijkm was the residual error.For all variables, n = 12; no observations were excluded.Using PROC UNIVARI-ATE, the normality of distribution and homogeneity of variance of residuals were examined.Data are presented as least squares means.The Tukey's method was used to compare multiple means.Significance was declared at P ≤ 0.05, and tendencies were noted if 0.05 < P ≤ 0.10.

Nutrient Intake and Digestibility
The NDF intake was greater for the S-BY-CM and CG-BY diets, which provided nearly 3 and 1.5 times less starch, respectively, compared with the CON diet (Table 2).Cows fed CON and S-BY-CM diets had higher DMI than cows fed CG-BY diets (Table 2).The BY-based diets decreased starch, HEE, and HEP intakes of cows (P < 0.001) but tended to increase NDF intake (P = 0.07) compared with the grain-containing (CON) diet.Cows fed S-BY-CM diets consumed more OM (P = 0.002) than cows fed CON and CG-BY, while dietary treatments did not affect CP intake.The CG-BY group consumed more ether extract (P = 0.01), but S-BY-CM diets consumed more sugar (P = 0.0002).Cows offered BY-based diets tended to have increased NDF digestibility (P = 0.06), whereas the S-BY-CM and CON diets tended to increase DM digestibility (P = 0.06) and increased digestibility of CP compared with those fed the CG-BY diet (P = 0.01).Starch digestibility was lower S-BY-CM or CG-BY diets than for CON (P < 0.001).

Sorting and Chewing
The most sorting behavior against long particles remaining on the top sieve (19 mm) occurred when cows were fed the S-BY-CM diet (Table 3).Cows fed CG-BY and CON sorted against medium particles (<19.0,> 8.0 mm), but the highest sorting behavior was found when cows were fed the CG-BY diet (P = 0.02).Cows fed CON and CG-BY had the most sorting for short particles (<8, > 1.18 mm) and against fine particles remaining on the pan compared with S-BY-CM (P = 0.003).Cows fed the CON diet spent longer eating per kilogram of NDF intake but less eating time per kilogram of starch intake than BY-based diets (P < 0.001; Table 3).The time eating associated with peNDF > 1.8 intake was not affected by treatments (P = 0.57).The total eating time increased for the S-BY-CM group compared with the other dietary treatments (P = 0.01).The minutes of ruminating associated with kilograms of NDF intake tended to decrease (P = 0.07), but the duration of ruminating associated with kilograms of starch intake increased when S-BY-CM and CG-BY were fed (P < 0.0001) compared with CON.Feeding S-BY-CM decreased the minutes of ruminating associated with kilograms of peNDF > 1.8 intake compared with CG-BY (P = 0.04).However, BY inclusion did not affect total daily ruminating time.The BY of either source did not affect the duration of total chewing per kilogram of peNDF > 1.8 intake (P < 0.41).However,

Temperature-Humidity Index and BW
The average ambient temperature and RH were 30.1°C and 33.6% during the trial, and the average THI was 75.8 (Figure 1).Although BCS was not affected by diets (Table 4), cows fed S-BY-CM tended to have the greatest BW, followed by those fed the CG-BY and then those fed CON (P = 0.09).The rectal temperature and respiration rate were not affected by treatments.

Ruminal Fermentation
No effects of diets were observed for ruminal pH, ruminal NH 3 , valerate, and total VFA concentration (Table 5).The treatments based on BY feedstuffs increased the molar proportion of acetate (P = 0.02) but decreased the butyrate proportion in ruminal fluid (P = 0.001).The CON diet increased the molar proportion of propionate compared with both BY-based diets.The isobutyrate and isovalerate proportions were greater in cows fed CON and CG-BY diets (P < 0.05) than when cows were fed the S-BY-CM diet.

Milk Production and Composition
Milk production was lowest when cows were fed the CG-BY-based diet (P < 0.001), whereas there were no differences between the other 2 treatments.The BYbased diets increased the content of fat and the fat-toprotein ratio (P < 0.001) compared with the CON diet.Although the FCE for HEP and HEE outputs increased for the BY-containing diets compared with the CON diet (P < 0.001), FCE for 3.5% FCM and ECM production was not affected among treatments.Cows fed CON and S-BY-CM treatment had lower FCE for actual milk/DMI than those fed CG-BY (P = 0.01).Feeding CON increased lactose yield (P < 0.001) compared with BY-based diets.The net food production in terms of protein and energy increased when cows were fed BY-based diets rather than the CON diet (P < 0.001).

DISCUSSION
The complete substitution of cereal grains with BY from sugar-rich or cereal grain resources had a significant impact on the nutritional profile of the diets.Specifically, the BY-based diets showed higher levels of NDF, sugar, and fat, while the starch content was lower, leading to an increased contribution of lipogenic nutrients rather than glucogenic nutrients.These alterations in nutrient composition were found to have significant effects on the DMI and production of mid-lactation cows.Changes in energy availability may have played a role in the observed alterations in milk production and composition, as well as in the nutrient digestibility and ruminal fermentation of the cows.Both dietary TMR containing BY at the expense of grains reduced starch intake, HEP intake, and HEE intake but increased sugar intake for the S-BY-CM, the fat intake for the CG-BY, and the NDF intake for both BY-based diets.
In the current study, the average minimum, mean and maximum THI values were 65.6, 75.8 and 86.6, respectively.Previously, Ahangaran (2013) evaluated the effects of THI on performance responses of dairy cows in climatic conditions of 7 commercial dairy farms in Isfahan Province in Iran.They suggested that the threshold of THI to trigger heat stress in these farms was 65, which was associated with lower DMI, milk production, and milk fat content.Therefore, average THI values in our study were well above the minimum needed to induce heat stress conditions.This observation is in agreement with the THI classification developed by Zimbelman et al. (2009) for high-producing dairy cows, which considers a range of 72-79 units to represent mild-to-moderate heat stress.Furthermore, respiration rates averaged more than 60 bpm, and rectal temperature averaged more than 38°C; both are characteristic of heat-stressed dairy cattle (Kadzere et al., 2002).The treatments did not have an impact on the rectal temperature and respiration rate in the current research.Previous studies (Miron et al., 2008, Heydari et al., 2021) have suggested that substituting forage with different BY in the diet of lactating dairy cows under high ambient temperatures could help alleviate signs of heat stress.In the study conducted by Miron et al. (2008), it was reported that a decrease in forage NDF from 18% to 12%, achieved through partial replacement of soy hulls for wheat silage, led to reductions in both respiration rate and rectal temperature.In addition, Heydari et al. (2021) reported that a decrease in forage NDF from 19.2% to 13.2% DM, achieved through partial replacement of beet pulp for corn silage, resulted in reductions only in respiration rate, with no significant effect observed on rectal temperature.This effect was attributed to digestion of diets with higher forage content producing a greater amount of heat than those with lower forage fiber content.Therefore, the reason for the similar rectal temperature and respiration rate across the treatments in the current study is likely due to the provision of low-forage in the basal diets.In addition, the time at which respiration rate and rectal temperature are monitored may be crucial for determining the occurrence of heat stress signs.The current study measured the respiration rates and rectal temperature of heat-stressed cows during the peak heat hours between 1400 and 1500 h.However, Drackley et al. (2003) suggested that measuring these parameters in the early morning might provide a better indication of heat stress, as the inability of cows to dissipate heat overnight could contribute to the intensity of heat stress (West, 1999).
In the current study, the CG-BY diet, which is characterized by high fiber and high-fat content, led to a significant reduction in both DMI and milk yield in comparison to other dietary treatments.Similar to our study, previous studies have reported a decrease in DMI 3 pef > 8 and pef > 1.18 = physical effectiveness factor, determined as the proportion of particles retained on 2 sieves (Lammers et al., 1996) and on 3 sieves (Kononoff et al., 2003), respectively; peNDF > 8 and peNDF > 1.18 = physically effective NDF determined as NDF content of TMR multiplied by pef > 8 and pef > 1.18, respectively.
4 Geometric mean particle size, calculated according to the method of the American Society of Agricultural Engineers (ASAE 1995; method S424.1).
5 Geometric standard deviation of particle size, calculated according to the method of the American Society of Agricultural Engineers (ASAE 1995; method S424.1). 6The sorting index was computed as the ratio of actual to predicted particle retaining on each screen of the Penn State Particle Separator (Leonardi and Armentano, 2003) when the energy density of mid-lactation cows was increased by incorporating supplemental unprotected fat, as opposed to a high starch-based diet, under heat stress conditions.For instance, under heat stress conditions, Drackley et al. (2003) observed a reduction in DMI by 1.24 kg when the diet was supplemented with 3% DM of choice white grease.Similarly, Moallem et al. (2010) reported a decrease in DMI by 0.7 kg when utilizing 1.5% DM of calcium salts of FA from palm oil distillate, compared with a high starch-based diet, during the summer.These results are also consistent with previous research studies under normal weather conditions that have demonstrated a similar trend, where cows fed a high fiber and high-fat diet exhibited lower DMI and milk yield compared with those fed a high-starch diet (Ranathunga et al., 2010, Boerman et al., 2015).By incorporating fat-rich ingredients such as corn germ meal and rice bran, along with supplemental fat, the fat concentration of the CG-BY diet increased substantially from 4.3% to 5.7% on a DM basis.This significant increase in dietary fat concentra-   tion suggests that the intake of unsaturated fatty acids in the CG-BY diet may have been significantly greater than that of the S-BY-CM and CON diets.Feeding unsaturated fatty acids has been reported to limit feed intake by stimulating the release of cholecystokinin, creating a sense of satiety, and decreasing the ruminal passage rate.Similar to our findings, Ranathunga et al. (2010) reported that adding dried distillers grains with solubles instead of corn at 21% of DM in the diets of dairy cattle increased the fat concentration in the diet from 4.35% to 5.48% on a DM basis.This, in turn, led to a notable increase in the concentration of unsaturated fatty acids from 58% to 74.2% on a DM basis and reduced DMI by approximately 2.7 kg/d.In addition to greater fat intake, another reason for the decrease in DMI observed in the CG-BY may be the physical limitations of ruminal fill when substituting grains with high-fiber BY components (Forbes, 2007).The bulkiness and lower density of high-fiber CG-BY, such as corn germ meal, rice bran, wheat bran, and barley malt sprout, resulted in a greater associative effect than high-grain diets, leading to increased physical space occupation in the rumen.As a consequence, capacity of the rumen may have been reduced, and physical limitations to ruminal fill may have occurred, causing a decrease in DMI.Moreover, the SBM of the CON treatment was replaced with canola meal of S-BY-CM and a mix of feedstuffs like corn germ meal, barley sprout, rice bran, and wheat bran in the CG-BY diet.Accordingly, another explanation for decreased DMI for CG-BY group may be attributed to the Maillard reaction caused by the agro-industrial process used to extract specific plant components.This process may have reduced the availability of AA (Slavin et al., 2000, Pang et al., 2018) and influenced the apparent totaltract digestibility of nutrients (DM and CP)..
Feeding CG-BY to dairy cows resulted in a significant reduction in ruminal DM digestibility and, although not statistically significant, a numerical decrease in the ruminal OM digestibility compared with other dietary treatments.The decrease in digestibility can be attributed to the shift in nutrient profile resulting in an average increase of 320 g in fat consumption for the CG-BY group.Our findings are consistent with those of previous in vitro and in vivo studies (Getachew et al., 2004, Huhtanen et al., 2009), which observed lower ruminal DM and OM digestibility when feeding diets containing increased levels of fat, as was the case with the CG-BY group in this study.Lower DMI and DM digestibility can occur if the level of dietary fat is excessive or if the type of fat being fed is not easily digestible by the cow.Studies have shown that different sources of fat can have different effects on digestibility.For example, Palmquist and Jenkins (1980) found that the digestibility of total FA was greater with calcium soap of palm oil compared with tallow or corn oil.Jenkins et al. (2008) reported that the digestibility of DM and NDF was higher with canola oil compared with tallow or corn oil.In terms of the amount of dietary fat, several studies have reported a decrease in digestibility with increasing levels of fat intake.For example, Jenkins (1997) found that increasing the total dietary fat (animal and vegetable fats, or blends thereof) from 2.5% to 10% reduced the DM digestibility in dairy cows.Likewise, the impact of fat saturation on the digestibility and production responses to a diet containing 20% soybean hulls was examined by Pantoja et al. (1994).The diets consisted of saturated tallow, tallow, or a blend of animal and vegetable fats, with relatively high concentrations, resulting in an average dietary FA concentration of 6.1% of DM.Under these circumstances, an increase in the proportion of unsaturated dietary FA concentration led to a linear reduction in DMI, a quadratic decrease in microbial CP flow, and a tendency for a linear decrease in DM digestibility (Pantoja et al., 1994).
The inclusion of a fat source in non-forage fiber-based diets may facilitate further reductions in NFC content, potentially enhancing productivity.Caution should be exercised to limit the availability of ruminally accessible unsaturated FA concentration to avoid adverse effects on ruminal fermentation and biohydrogenation (Lock, 2010).However, some studies have reported no negative effects on digestibility with moderate levels of dietary fat.For instance, a study by Boerman et al. (2015) reported that increasing the level of dietary fat from 3% to 5% (2.5% palmitic acid-enriched FA) had no significant effect on DM digestibility.Additionally, under normal circumstances in ruminants, a reduction in DMI is known to lead to an increase in diet digestibility and a decrease in ruminal passage rate (Warren et al., 1974, Mulligan et al., 2001).However, under heat stress conditions, several studies have found that a decrease in DMI has no effect on diet digestibility or ruminal passage rate (Attebery and Johnson, 1969, Miaron and Christopherson, 1992, Bernabucci et al., 1999).This is due to the direct impact of heat stress on ruminal motility, which decreases either contraction amplitude or frequency, and is not mediated by changes in DMI (Attebery and Johnson 1969;Bernabucci et al., 2012).Consequently, heat stress may have had a detrimental effect on ruminal function of CG-BY-fed groups, independent of alterations in DMI.
Similar to in vivo (Münnich et al., 2018) and in vitro (Ertl et al., 2015a) studies, our research findings indicate that the ruminal DM digestibility of the S-BY-CM diet, which was high in sugar and fiber, was comparable to diets containing cereals.Indeed, previous studies have consistently demonstrated that NSC such as sugar, pectin, and starch are digested more rapidly than structural carbohydrates such as cellulose and hemicellulose, which is consistent with our findings (Marounek et al., 1985, Oba, 2011, Boerman et al., 2015).We found no significant difference in milk yield between the CON and S-BY-CM treatments.This finding agrees with the results by Karlsson et al. (2018), who replaced SBM with canola meal and cereal grains with a combination of sugar beet pulp and wheat bran for Swedish Holstein and Swedish Red dairy cows with an average milk yield of 31 kg/d.Huhtanen et al. (2011) conducted a meta-analysis investigating the replacement of SBM with CM in isonitrogenous diets based on grass silage.They found that cows fed CM had higher DMI and milk yields compared with those fed SBM.In another study by Mulrooney et al. (2009), replacing dried distillers grains with solubles with CM in varying proportions (100, 66, 33, and 0%) did not affect milk yields and milk components.However, they concluded that diets with higher proportions of CM may be more desirable due to a reduction in MUN and better concentration of blood AA, despite no significant difference in milk yields and components.Both in vitro (Broderick et al., 2016) and in vivo (Broderick et al., 2015) studies suggested that replacing SBM with CM did not significantly alter ruminal fermentation.Therefore, it is postulated that the advantages of feeding CM may be attributed to increased DMI or improved postruminal utilization, such as a better AA profile.Based on these findings, both CM and SBM can perform equally in meeting the post-ruminal utilization of nutrients, given that there were no significant differences in DMI and MUN concentration between the S-BY-CM and CG-BY diets in our study.Furthermore, a study conducted by Sánchez-Duarte et al. (2019) showed that increasing the starch content in the diet from 21% to 27% along with CM led to an increase in DMI, which resulted in a slight increase in milk yield.However, in our study, there were no significant differences in DMI or milk yield between the control group and the S-BY-CM treatment, even when the starch content in the diet decreased from 32.5% to 9.4% of DM.This suggests that the higher concentrations of NDF and sugar in the S-BY-CM treatment may have improved the supply of MP to the dairy cows, as has been previously suggested by Broderick et al. (2002).Since sugars are fermented more quickly in the rumen than starch, they can be a beneficial addition to low-starch and low-forage diets.
We determined lower starch digestibility and higher fiber digestibility in BY-based diets as anticipated, owing to the comparatively low starch levels and highly digestible NDF content of BY diets compared with the CON diet.The higher digestibility of NDF in BY-based diets can be explained by their similarity in NDF content to forage, as well as their lower levels of lignin and indigestible NDF, and smaller particle size.These types of fiber are more easily fermented in the rumen, leading to increased production of acetate and butyrate Erfani et al.: Effects of complete substitution… rather than lactate in the rumen and promoting the growth of fiber-digesting bacteria, which can improve fiber digestion and nutrient utilization (Karlsson et al., 2018, Pang et al., 2018).
Both BY-based diets replacing grains resulted in higher molar concentrations of acetate and butyrate in the ruminal fluid.However, when compared with the BY-based diets, feeding the CON diet led to an increase in propionate concentration.Van Knegsel et al. (2007) suggested that in ruminants, lipogenic nutrients are derived from dietary fat or fiber, which stimulates the production of acetate and butyrate in the rumen.The fermentation of sucrose in the rumen can also produce lipogenic precursors like acetate or butyrate, depending on the amount of sugar in the diet, and less propionate.Glucogenic nutrients come from rumen-degraded starch.Propionic acid, glucogenic amino acids, and lactic acid contribute to gluconeogenesis in ruminants in decreasing order of significance (Brockman, 2005).Furthermore, Bannink et al. (2006) analyzed 182 diets and found that the molar proportion of propionic acid increased due to fermentation of starch as compared with fermentation of cellulose or hemicellulose, which is consistent with the present findings.The increase in propionate concentration when the CON diet was fed in comparison to BY-based diets likely was due to the higher starch and lower fiber contents.
We expected to observe a decrease in milk fat concentration when cows were fed the cereal grain-based CON diet.The high starch and relatively low total NDF content of the CON diet are risk factors for milk fat depression (Bauman et al., 2011).The replacement of CG-BY and S-BY-CM with ground corn and barley in the CON group resulted in a decrease of 9% and 7% in milk fat concentration, respectively.Notably, the observed increase in milk fat for CG-BY compared with the CON group can be attributed to a dilution effect caused by a higher milk yield of 2 kg/d when comparing the CON and CG-BY diets.This finding is consistent with recent studies conducted by Zang et al. (2021) and Malekkhahi et al. (2023).Zang et al. (2021) reported an 8% decline in milk fat content when substituting soy hulls and beet pulp for ground corn (30% DM), resulting in a reduction of starch concentration from 34.4% to 12.3% of DM.Similarly, Malekkhahi et al. (2023) documented a decrease in milk fat content when substituting beet pulp for corn silage (24% DM), leading to a reduction of starch concentration from 32% to 22% of DM.In both studies, the decrease in milk fat concentration was accounted for by a dilution effect, which can be attributed to an increase in milk yield of 2.2 kg/d (Zang et al., 2021) and 1.8 kg/d (Malekkhahi et al., 2023) when comparing high-starch diets to lowstarch diets.However, the dilution effect was not ob-served between CG-BY and S-BY-CM, despite 2.3 kg/d increase in milk yield with feeding S-BY-CM compared with CG-BY.This finding may be explained by similar NDF digestibility and ruminal acetate and butyrate production.Ruminal acetate and butyrate are known precursors for de novo milk fat synthesis (Palmquist et al., 1969, Clegg et al., 2001), which occurs upon intake of fibrous BY instead of cereal grains.Another reason for the lower milk fat concentration observed by feeding the CON group is due to the increased digestibility of ruminal starch in ground corn and barley grain, coupled with a decrease in the supply of physically effective fiber, which is believed to have led to a reduction in ruminal pH.Consequently, this shift in ruminal conditions may have altered the biohydrogenation pathways, favoring the production of trans-10 18:1 and trans-10, cis-12 18:2 FA.These specific FA are well-known for their inhibitory effects on milk fat synthesis in the mammary gland, as elucidated in studies conducted by Baumgard et al. (2002) andShingfield et al. (2009).Also, a decrease (P = 0.06; Table 3) in apparent totaltract NDF digestibility observed in the CON group might indicate reduced availability of ruminal acetate for the de novo synthesis of FA.
Both ECM and 3.5% FCM increased when dairy cows were fed with the S-BY-CM diet, but a similar response was not observed when feeding CG-BY in place of cereal grains.The explanation for the increased ECM and 3.5% FCM production is that cows fed with CG-BY diet produced 2.3 kg/d less milk, despite the same amount of milk fat as cows fed the S-BY-CM diet.Therefore, the ECM and 3.5% FCM increased when dairy cows were fed with S-BY-CM diet rather than the other 2 dietary regimens due to a higher production of milk and fat.Comparing the CG-BY diet based on high-fiber high-fat inclusions, the S-BY-CM and CONfed cows showed a lower FCE for actual milk despite greater DMI and milk production.The observed lower FCE for actual milk occurred because our S-BY-CM and CON treatments had similar milk production and DMI, while our CG-BY resulted in nearly 2.3 and 2.5 kg/d reduced milk production and DMI, respectively.Similarly, earlier research indicated no difference in FCE across treatments when partial or complete replacement of BY for cereal grains resulted in a slight change in DMI, milk, or milk component production at varying forage to concentrate ratios (Voelker and Allen, 2003, Dann et al., 2014, Karlsson et al., 2018).
Cows fed the CON diet exhibited milk fat depression compared with those fed the BY-based diets.The milk fat depression can be a sign of pH changes in the rumen, which can indicate sub-acute ruminal acidosis.High-starch, low-fiber diets are known to increase the production of glucogenic propionate and decrease the production of lipogenic acetate, as reported by van Knegsel et al. (2007).This theory is further supported by the fact that low ruminal pH can affect the production of ruminal biohydrogenation products, which can then lead to a decrease in the de novo synthesis of FA in the mammary gland, resulting in milk fat depression, as reported by Plaizier et al. (2008).The decrease in milk fat observed by feeding glucogenic diets (CON) compared with those fed with lipogenic diets (BY) indicates that the CON-fed group may had a higher risk of ruminal acidosis and would have experienced a longer period in which pH was below 5.8 than the BY-fed group.Zebeli et al. (2012) indicated that diets should contain 31.2%peNDF > 1.18mm and 18.5% peNDF > 8mm, along with approximately 20% starch to prevent sub-acute ruminal acidosis.The CON diet not only had less peNDF (26.74% for peNDF > 1.18mm and 10% for peNDF > 8mm) than the recommended range but also had higher starch (32.5%) than advised by Zebeli et al. (2012).Therefore, the CON group had a greater risk for ruminal acidosis and milk fat depression, due to higher starch content and lower peNDF in comparison to the recommended thresholds.Notably, the CON group of the present study received almost 4 and 6 kg/d more starch than the CG-BY and S-BY-CM groups, respectively.However, the ruminal pH of the present study for CON group was 6.17, which is high and inconsistent with earlier studies testing high concentrate diets (Voelker and Allen, 2003, Kahyani et al., 2019, Razzaghi et al., 2020).The primary differences between the present study and the referenced papers are the inclusion of sodium bicarbonate in the diet and the adaptation of dairy cows to prevent a decrease in ruminal pH.The addition of sodium bicarbonate may have played a significant role in regulating ruminal pH (Russell and Rychlik, 2001).It is possible that the cows in the present study fed the CON diet had developed a resistance to ruminal acidosis.This adaptation may have allowed the cows to better regulate their ruminal pH in response to changes in diet and management practices.On the other hand, BY-based diets of the present study had lower starch (19.4 and 9.5%) and peNDF > 1.18mm (10.4 to 13.1%) for CG-BY and S-BY-CM, respectively, indicating that amount of fiber in BY-based diets may not be sufficient to effectively promote proper digestion and maintain good health in animals.In contrast to this scenario, the findings for S-BY-CM indicated that supplementing diets low in starch with high-fiber, sugar-rich BY feedstuff can have beneficial effects on ruminal pH.Specifically, this approach increased ruminal acetate, increased milk fat, and increased NDF digestibility compared with feeding CON.According to the findings of our study, replac-ing starch-rich ingredients in diets with sugar-rich BY feedstuffs may help promote a more stable rumen environment by reducing the intake of starch and peNDF.This implies that diets containing lower starch, greater sugar, and highly digestible fiber may have the potential to support improvements in DMI, milk yield, and milk component yields, as well as maintain ruminal health, without requiring additional physically effective fiber.
By replacing the cereals and SBM of CON with S-BY-CM, the increase in selective ingestion of higherenergy ration components and against longer particles low in peNDF contents might be attributed to an increase in energy demand as impacted by a fluctuating metabolic status resulting from heat stress.Sorting behavior against longer particles and in favor of smaller particles of the S-BY-CM diet is not likely responsible for less of a pH drop because of following scenarios.(1) A high-sugar diet, despite its quick fermentation in the rumen like starch, yields less carbon per unit of mass for fermentation acid production compared with starch (Hall and Herejk, 2001).( 2) Bacteria may convert sucrose to glycogen as a short-term energy store (Hall and Weimer, 2007), which temporarily decreases fermentation acid production in the rumen and may contribute to a higher ruminal pH.Additionally, pectin content of the S-BY-CM diet increased since it mostly comprised pectin-rich feedstuffs such as dried citrus pulp and sugar beet pulp.Pectin yields minimal lactate and, unlike starch, is not digested rapidly and extensively in the rumen (Marounek et al., 1985, Barrios-Urdanetat et al., 2003).
Eating behavior was affected by S-BY-CM, which increased compared with the other 2 dietary treatments.The similar DMI between the CON and S-BY-CM treatments suggests that the increased eating time of S-BY-CM-fed cows is likely due to greater sorting against long particles compared with the other diets.DeVries (2011) found that cows that sorted more against the longest ration particles had slower eating rates, which is consistent with our findings.Hence, the observed differences in sorting behavior between cows of the present study could partially account for the differences in eating rates among these groups of cows.In the current study, we found that chewing and ruminating activity remained consistent across all treatments, despite the higher NDF content from the inclusion of BY.These findings align with previous studies that have replaced cereal grains with alternative feed sources, such as beet pulp and wheat middlings in a TMR with a 50% DM forage to concentrate ratio (Dann et al., 2014), or with wheat bran and sugar beet pulp as the sole BY in a high forage diet (Ertl et al., 2016).The similarity in chewing activity between treatments suggests that BY feedstuffs may lack the physical effectiveness to stimu-Erfani et al.: Effects of complete substitution… late chewing activity in heat-stressed dairy cows.This is consistent with previous research by Mertens (1997) and Zebeli et al. (2012), which indicated that the physical form of feed particles can influence chewing activity in ruminants.Despite the lower physical effectiveness of BY feedstuffs, the findings of the current study suggest that they can still serve as a viable alternative to cereal grains in dairy cow diets, without affecting chewing and ruminating activity.
To meet the energy demands of high-yielding cows, a worldwide tendency has been to feed them human edible grains, but this has impacted the long-term sustainability of the dairy industry (Wilkinson, 2011, Ertl et al., 2015b, Karlsson et al., 2018, Pang et al., 2018).High-energy grains have become more valuable because of their application in biofuels manufacturing.Therefore, dairy farming sustainability can be increased through strategies that reduce the consumption of human edible cereal grains by cattle (Naderi et al., 2022).Our research shows that replacing grains with BY reduced dietary HEP and HEE, which is consistent with earlier findings (Ertl et al., 2015b, Karlsson et al., 2018, Münnich et al., 2018).Therefore, the complete substitution of dietary BY for grains such as barley and corn in heat-stressed dairy cows could enhance the sustainability of milk production in countries with widespread heat stress.

CONCLUSIONS
Our findings indicate that replacing grains with BY altered the nutrient profile from glucogenic to lipogenic.Both diets containing BY improved nutrient recovery, but the outcomes for dairy cows varied depending on the type of BY energy source.In particular, feeding S-BY-CM led to a higher DMI, increased ECM, and reduced signs of ruminal acidosis, such as milk fat depression.Conversely, CG-BY resulted in decreased DMI, lower milk yield, and reduced DM digestibility.The highly digestible NDF, sugar, pectin, and CM present in S-BY-CM offered numerous benefits as a substitute for barley and corn grains in mid-lactation cows experiencing heat stress conditions.Given its positive impact on feed intake, ECM, and digestibility of nutrients, our findings suggest that feeding S-BY-CM could be advantageous in low-forage-based diets for heat-stressed dairy cows without negative effects on ruminal pH.
Erfani et al.: Effects of complete substitution… 2 Erfani et al.: Effects of complete substitution… the CON diet increased the chewing time per kilogram of NDF and decreased that per kilogram of starch compared with both BY diets.The total chewing time per day was not affected by dietary treatment, similar to the results of total eating and ruminating time.

Figure 1 .
Figure 1.Temporal pattern of the mean, minimum, and maximum temperature-humidity index (THI) during the experimental period.The study was conducted between July and September during the summer of 2020 using multiparous Holstein dairy cows.
Erfani et al.: Effects of complete substitution… Erfani et al.: Effects of complete substitution… greater NDF, sugar, and fat but lesser starch and NFC by inclusion of dietary BY-based concentrate reflect differences in the component concentrations of BY and CON diets (Table

Table 1 .
Wilkinson (2011)ffects of complete substitution… Ingredients, chemical composition of the experimental diets on a DM basis and estimated proportion of human edibles according toWilkinson (2011)

Table 2 .
Erfani et al.: Effects of complete substitution… Nutrient intake and apparent total-tract nutrient digestibility as influenced by substituting byproduct (BY) sources for cereal grains 1Experimental diets were different in concentrate fraction: CON = concentrate containing cereal grain and soybean meal; CG-BY = concentrate containing cereal grain by-products; S-BY-CM = concentrate containing sugar by-products and canola meal.

Table 3 .
Erfani et al.: Effects of complete substitution… Physical characteristics of experimental diets and sorting index (measured using the Penn State Particle Separator 1 ), and chewing variables as influenced by substituting by-product (BY) sources for cereal grains 2Experimental diets were different in concentrate fraction: CON = concentrate containing cereal grain and soybean meal; CG-BY = concentrate containing cereal grain by-products; S-BY-CM = concentrate containing sugar by-products and canola meal.

Table 4 .
Body weight (BW), body condition score (BSC), rectal temperature, and respiration rate as influenced by substituting by-product (BY) sources for cereal grains Erfani et al.: Effects of complete substitution… a,b Means within a row with different superscripts differ (P ≤ 0.05). 1 Experimental diets were different in concentrate fraction: CON = concentrate containing cereal grain and soybean meal; CG-BY = concentrate containing cereal grain by-products; S-BY-CM = concentrate containing sugar by-products and canola meal.

Table 5 .
Ruminal fermentation characteristics as influenced by substituting by-product (BY) sources for cereal grains 1 Experimental diets were different in concentrate fraction: CON = concentrate containing cereal grain and soybean meal; CG-BY = concentrate containing cereal grain by-products; S-BY-CM = concentrate containing sugar by-products and canola meal.

Table 6 .
Lactation performance as influenced by substituting by-product (BY) sources for cereal grains