Effect of pre-and postpartum supplementation of a pure glycerol product to dairy cows on feeding behavior, lying behavior, and reticulorumen pH

The objective of this study was to quantify the effects of supplementing a low level of dry glycerol product pre-and postpartum on the feeding behavior, lying behavior, and reticulorumen pH of dairy cows. Mul-tiparous Holstein dairy cows (n = 60) were enrolled in a 2 × 2 factorial design study. Twenty-one days before expected parturition, cows individually received a dry cow diet with (1) 250 g/d glycerol supplementation (GLY; 66% pure glycerol, United States Pharmacopeia grade), or (2) no supplementation (CON). Following parturition, cows were individually assigned to either (1) 250 g/d glycerol product (GLY; 66% pure glycerol), or (2) no supplementation (CON) to their partial mixed ration (PMR) for the first 21 d in milk (DIM). All cows were milked by an automated milking system and offered a target of 5.4 kg/d pellet (23% of target total dry matter intake [DMI]). For both treatment periods, cows were individually assigned to automated feed bins to measure PMR feeding behavior. Rumination time and lying behavior were monitored with electronic sensors for the whole study (−21 to 21 DIM). Reticuloru-men pH boluses were administered to a subset of cows (n = 40) where pH was recorded every 10 min from 21 d prepartum to 21 d postpartum. Prepartum, cows fed GLY had fewer, larger meals and spent 20.2% more time feeding than CON while consuming feed at a similar rate. Cows on the CON diet prepartum spent more time lying down in more frequent bouts in the 21 d before calving. Following parturition, cows that received GLY prepartum continued to devote more time to eating, while tending to spend less time ruminating per kilogram of DMI. Cows receiving CON postpartum had larger meals with longer intervals between meals. In the first 21 DIM, cows receiving CON prepartum tended to have shorter, but significantly more frequent, lying bouts than cows fed GLY prepartum. Glycerol supplementation pre-and postpartum resulted in less time spent lying down following parturition. Minimal differences between treatments were observed for pre-and postpartum sorting behavior or reticulorumen pH. Overall, supplementation of glycerol pre-and post-partum altered cow time budgets, with cows spending more time eating pre-and postpartum, less time lying pre-and postpartum, and having fewer, larger meals prepartum when receiving glycerol prepartum, and with cows having slower feeding rates and smaller meals following parturition with postpartum glycerol supplementation.


ABSTRACT
The objective of this study was to quantify the effects of supplementing a low level of dry glycerol product pre-and postpartum on the feeding behavior, lying behavior, and reticulorumen pH of dairy cows.Multiparous Holstein dairy cows (n = 60) were enrolled in a 2 × 2 factorial design study.Twenty-one days before expected parturition, cows individually received a dry cow diet with (1) 250 g/d glycerol supplementation (GLY; 66% pure glycerol, United States Pharmacopeia grade), or (2) no supplementation (CON).Following parturition, cows were individually assigned to either (1) 250 g/d glycerol product (GLY; 66% pure glycerol), or (2) no supplementation (CON) to their partial mixed ration (PMR) for the first 21 d in milk (DIM).All cows were milked by an automated milking system and offered a target of 5.4 kg/d pellet (23% of target total dry matter intake [DMI]).For both treatment periods, cows were individually assigned to automated feed bins to measure PMR feeding behavior.Rumination time and lying behavior were monitored with electronic sensors for the whole study (−21 to 21 DIM).Reticulorumen pH boluses were administered to a subset of cows (n = 40) where pH was recorded every 10 min from 21 d prepartum to 21 d postpartum.Prepartum, cows fed GLY had fewer, larger meals and spent 20.2% more time feeding than CON while consuming feed at a similar rate.Cows on the CON diet prepartum spent more time lying down in more frequent bouts in the 21 d before calving.Following parturition, cows that received GLY prepartum continued to devote more time to eating, while tending to spend less time ruminating per kilogram of DMI.Cows receiving CON postpartum had larger meals with longer intervals between meals.In the first 21 DIM, cows receiving CON prepartum tended to have shorter, but significantly more frequent, lying bouts than cows fed GLY prepartum.Glycerol supplementation pre-and postpartum resulted in less time spent lying down following parturition.Minimal differences between treatments were observed for preand postpartum sorting behavior or reticulorumen pH.

INTRODUCTION
Metabolic and other diseases often arise from lactating dairy cows failing to consume adequate DMI early in lactation to offset increasing nutrient requirements of milk production, resulting in a state of negative energy balance (Herdt, 2000).Promoting consistent DMI during the transition period may reduce the risk of metabolic disease early in lactation (Goldhawk et al., 2009).Inadequate prepartum DMI, and the resulting increased risk of subclinical ketosis and metritis following parturition, have been associated with a reduction in meal frequency and time spent eating each day in the weeks leading up to parturition (Urton et al., 2005;Huzzey et al., 2007;Goldhawk et al., 2009).Those cows at risk for, and experiencing, postpartum diseases often have reduced feeding time, which is exchanged for more time spent lying down (Kaufman et al., 2016).Following parturition, greater DMI, and resultant milk production, may be related to the increased time cows spend feeding, meal frequency, and rumination time (Johnston and DeVries, 2018).
Consistency in intake across the transition period, both in terms of amount and composition of diet consumed, is important to minimize the severity of negative energy balance and maintain body condition (Havekes et al., 2020a,b,c).One factor effecting the composition of diet consumed is sorting behavior (Dann et al., 2006;Havekes et al., 2020a,b,c).Discouraging this behavior is important, as extensive sorting for rapidly fermentable carbohydrates can not only lead to cows consuming a different diet than predicted, but in early lactation, lead to a greater risk of SARA (DeVries et al., 2008;Coon et al., 2019).The time cows spend sorting may result in greater eating time (Greter and DeVries, 2011); however, it may also take away from time that could be devoted to resting, as researchers have previously observed cows that sorted less had increased lying (resting) time per day (Kronqvist et al., 2021).In addition, cows with greater lying time and time spent ruminating have been associated with a greater probability of ruminating while lying down (McWilliams et al., 2021).Increasing the overall time cows allocate to rumination has been positively associated with DMI (Johnston and DeVries, 2018;Antanaitis et al., 2019).Overall, this means that the intake and health of dairy cows at transition may be not only influenced by, but also contribute to, differences in the behavior of cows, including time spent lying, eating, and ruminating.
Feeding behavior, including meal patterning, rumination, and feed sorting, is affected by various aspects of dietary composition, including the moisture level, particle size, and feed ingredient type (Miller-Cushon and DeVries, 2017;DeVries, 2019).Some ingredients that affect feed flavor and ration energy content have the opportunity to affect eating behavior.For example, adding molasses to dry and lactating diets reduced feed sorting behavior and altered eating behavior (DeVries and Gill, 2012;Havekes et al., 2020a).Supplementation of glycerol has previously been reported to shift the timing of eating to later in the day, while reducing sorting against long forage particles during the transition period (Carvalho et al., 2012).Glycerol nutritionally has a 20% greater NE L than corn (Schröder and Südekum, 1999) and may influence rumen fermentation and feed efficiency (Rémond et al., 1993;Hippen et al., 2008).Dietary glycerol, a glucogenic precursor, is almost exclusively fermented by ruminal bacteria and absorbed as VFA, primarily propionate and butyrate; however, glycerol can also be directly absorbed in the rumen (Rémond et al., 1993;Silva et al., 2014;Werner-Omazic et al., 2015).Although elevating VFA production may result in lower rumen pH (Plaizier et al., 2008), no changes in ruminal pH concentrations with glycerol supplementation were observed in previous studies (Khalili et al., 1997;Rico et al., 2012;Boyd et al., 2013).
The literature is inconsistent when it comes to the relationship between feed intake and glycerol (a byproduct of the biodiesel industry) supplementation around parturition potentially due to differences in purity of glycerol, dose, amount of glycerol degraded in the rumen (vs.directly absorbed), diet composition, supplementation form and location, and stage of lactation.In Van Soest et al. (2023) we reported that dairy cows supplemented prepartum with glycerol (GLY), as compared with those control cows without supplementation (CON), increased their partially mixed ration (PMR) DMI both before (CON: 1.8 ± 0.05 vs. GLY: 2.0 ± 0.05% BW/d) and after calving (CON: 2.04 ± 0.05 vs. GLY: 2.23 ± 0.05% BW/d), and had reduced fat mobilization in early lactation.Those cows supplemented postpartum with glycerol, however, had lower PMR DMI (CON: 18.4 ± 0.23 vs. GLY: 17.9 ± 0.23 kg/d) for the first 21 DIM and devoted less energy to milk production, while demonstrating lesser fat mobilization, as evidenced by lower nonesterified fatty acids (NEFA) concentration and improved milk fatty acid profile.Other researchers have observed DMI to decrease as the dose of glycerol increases (Ezequiel et al., 2015;de Andrade et al., 2018).This decrease may result from the hepatic oxidation theory (Allen et al., 2009), where satiety is signaled due to the increase in propionate from rumen degradation of glycerol (Allen et al., 2009;Werner-Omazic et al., 2015).Increasing the amount of absorbed propionate has previously resulted in smaller and less frequent meals, translating to less DMI (Oba and Allen, 2003;Gualdron-Duarte and Allen, 2017).
To our knowledge, limited research has been done to assess the effects of either pre-or postpartum glycerol supplementation on feeding behavior, lying behavior, rumination time, and ruminal pH.The objective of this study was to determine whether the changes in DMI observed in Van Soest et al. (2023) were associated with changes in feeding behavior, lying behavior, time spent ruminating, and rumen pH of dairy cows during the transition period.It was hypothesized that dry glycerol supplementation both pre-and postpartum would have the greatest effect on increasing meal frequency and reducing sorting, without negatively affecting reticulorumen pH in early lactation, and that the changes in feeding behavior would alter lying behavior.

MATERIALS AND METHODS
Animal use and experimental procedures used throughout this trial complied with the guidelines of the Canadian Council on Animal Care (2009) and were approved by the University of Guelph Animal Care Committee (protocol #4493).

Animals, Housing, and Experimental Design
This study was conducted with 60 multiparous cows at the University of Guelph Elora Research Station-  (Van Soest et al., 2023) focused on feed intake, blood metabolic status markers, change in BW and BCS, milk production and composition, and energy balance.Detailed descriptions of the housing, management, and experimental design and treatments, and sample size calculation are described by Van Soest et al. (2023).Briefly, during the dry period, at any given time, 12 cows were randomly housed in each of the 2 identically managed freestall dry cow pens (12 automated feed bins and 24 freestalls with a mattress base bedded with chopped wheat straw) and trained to eat out of individually assigned automated feed bins (Insentec B.V., Marknesse, the Netherlands).Upon signs of calving (4.1 ± 0.77 d prepartum; mean ± SD), as noted by facility staff, cows were moved to an individual maternity pen where they had access to feed (their respective treatment diet) and water.For the first 3.4 ± 0.67 DIM, cows were milked using a portable milking system 2 times a day in their maternity pen before moving into the lactation cow pen.Once moved, cows were milked by an automated milking system (AMS: DeLaval, Tetra Laval Group, Tumba, Södermanland, Sweden) and assigned to an individual feeding bin for the remainder of the study (up to 21 DIM).Stocking density in the lactation pen never exceeded 30 lactating cows, which had access to 60 bedded stalls and 30 automated feed bins within the pen.
While accounting for previous milk production and parity, 21 d before expected parturition, cows were randomly assigned to either a base dry cow diet (CON, n = 30) or a glycerol-supplemented TMR containing 250 g/d per cow of glycerol product (actual = 261 ± 2.8 g/d per cow; RUMI-LAC, Probiotech International Inc., Saint-Hyacinthe, QC, Canada) on a DM basis mixed directly into CON TMR (GLY, n = 30; Table 1).The glycerol product contained 66% pure (99.9%) glycerol (United States Pharmacopeia grade) and 34% silica carrier and flavoring agents (<0.5% of supplement DM).The silica carrier and the flavoring agent were not included in either the pre-or postpartum CON rations.Offered ad libitum, both treatments were balanced to supply 100% of the energy requirements based on the NRC (2001; Table 1).Following parturition, cows were again assigned within the dry cow diet to either the base lactation PMR (CON, n = 30) or a glycerol-supplemented diet containing 250 g/cow per day of glycerol product (actual = 250.9± 2.9 g/cow per day) on a DM basis added to CON PMR (GLY, n = 30; Table 1), while balancing for previous lactation milk production and parity.All lactating cows were supplemented with concentrate through the AMS, starting at 2.7 kg of DM/d, and increasing 225 g of DM/d each day until the targeted 5.4 kg of DM/d was reached for the first 21 DIM.Rations (PMR and AMS concentrate allowance) were balanced to support the requirements of a 650-kg cow with an expected milk production of 40 kg/d, 3.80% milk fat, and 2.88% true protein with the predicted consumption of 26.8 kg/d of DMI (NRC, 2001).A refusal rate of 5 to 10% (actual: dry = 8.5 ± 6.2%, lactation = 7.1 ± 5.7%) was targeted to ensure ad libitum DMI.This feeding design resulted in a 2 × 2 factorial study design with 4 final treatment combinations: cows that received CON for both pre-and postpartum treatment periods (CC, n = 15, parity = 2.7 ± 1.01), cows that received CON TMR prepartum and GLY PMR postpartum (CG, n = 15, parity = 2.9 ± 1.21), cows that received GLY TMR prepartum and CON PMR postpartum (GC, n = 15, parity = 2.9 ± 1.13), and cows that received GLY for both pre-and postpartum treatment periods (GG, n = 15, parity = 2.5 ± 0.96).Feed was delivered to all cows once daily (dry cow TMR at 0900 h and lactating cow PMR at 0730 h).
As reported by Van Soest et al. (2023), 24 different clinical disease events were diagnosed, treated, and recorded by the research farm staff in the first 21 DIM.Ketosis was recorded for 18 cows (6 CC,4 CG,6 GC, and 2 GG; based on 2 Keto-Test [Elanco, Indiana, United States] milk strip tests conducted 1 wk apart by farm staff at 3-17 DIM), 2 cows (1 CG and 1 GC) had a retained placenta, 1 cow (CC) had mastitis, 2 cows (GC) had hypocalcemia, and 1 cow (GC) was recorded as having endometritis.Upon farm diagnosis of clinical disease, not including ketosis, cows were moved into individual pens for up to 3 d where cows were independently treated; behavioral data for that time away from the pen was omitted from the data set.

Feed Sampling and Feed Sorting Calculation
Fresh feed samples were collected 3 times per week in duplicate for each of the 4 treatment diets and all ration ingredients (including AMS pellet) were sampled monthly for the entirety of the trial.One feed sample was collected to determine chemical composition and DM (Table 1) and the other for particle size distribution (Table 2).Orts samples from each cow were collected 3 times per week following the days fresh feed sampling occurred starting 21 d before expected parturition and continued for the remainder of the trial (21 DIM).Feed samples were stored at −20°C upon collection and later thawed for 24 h before being oven-dried at 60°C for 48 h for DM analysis.Samples taken for size distribution, including dry TMR, lactation PMR, forage samples, and orts samples, were separated before drying into 4 fractions: long (>19 mm), medium (8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19) ternational, 2000, method 920.39), lignin (using ADF residue and H 2 SO 4 ), and minerals (using aquaregia digestion inductively coupled plasma atomic emission spectroscopy).Distribution of particle size, ADF, NDF, in vitro NDF digestibility after 48-h in vitro fermentation, and physically effective NDF (peNDF) content by particle size of the offered dry and lactation treatment diets are summarized in Table 2.
Sorting for each PSPS fraction was determined by dividing the actual consumption of each fraction by the predicted consumption of that respective fraction, expressed as a percent (Leonardi and Armentano, 2003).Actual consumption for each fraction of the TMR (dry period) and PMR (lactation period) was estimated by the difference between DM offered and refused from PSPS results of TMR or PMR, respectively.For each fraction, predicted consumption was the product of the total DMI of the TMR or PMR and DM of each individual fraction from the fresh TMR or PMR, respectively.Sorting values equaling 100% indicated no sorting for that particle fraction transpired, values <100% represented sorting against the fraction in question, and a value >100% meant cows sorted in favor of that particular fraction.

Feeding, Ruminating, and Lying Behavior Monitoring
Daily feeding behavior and DMI were recorded via individually assigned automated feed bins, by continuously recording the weight of feed at the start and end of each bin visit, as validated by Chapinal et al. (2007).The TMR and PMR DMI were calculated by multiplying the as-fed intake at each visit by the weekly average DM percent of the corresponding treatment diet.Total time spent feeding (feeding time; min/d) was determined daily by taking the sum of all time elapsed of individual feed bin visits.Feeding rate (kg/ min) was calculated by dividing the DMI of each visit by the feeding time of that visit.Meal criteria (i.e., the minimum duration of time between feed bin visits to be classified as a new meal) were calculated separately for each treatment period (dry and lactation) for each cow, as described by DeVries et al. (2003).A software package (MIX 3.1.3;MacDonald and Green, 1988) was used to fit normal distribution to the frequency of log 10transformed intervals of time between feeding visits.
When the time interval between 2 feed bin visits exceeded the cow-specific meal criterion, this was deemed a new meal.The quantity of meals in a day was termed meal frequency (meals/d).Once individual meals were determined, then time between those meals was calculated and termed interval between meals (min).Total meal time (min/d) was defined as the sum of all feeding time and non-eating time included within meals (i.e., all the time intervals between bin visits up to the length of the meal criteria).Meal length (min/meal) was then calculated as the total meal time divided by meal frequency.Meal size (kg/meal) was estimated daily by dividing TMR or PMR DMI by meal frequency.These feeding behavior measures were summarized to one value per cow per day for each of the study periods: dry treatment period (−21 to −1 d relative to parturition) and lactation treatment period (3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19)(20)(21) separately.An electronic monitoring system (HR-TAG-LD, SCR Engineers Ltd., Netanya, Israel) was used to monitor the rumination activity, as validated by Schirmann et al. (2009), for the entirety of the study.At enrollment each cow was equipped with a nylon collar fixed with the rumination data logger.Data was stored in 2-h intervals and transmitted to radio frequency readers, which were strategically placed throughout the barn and continuously uploaded to the central database.Files were downloaded twice weekly and combined into a continuous record for each cow.Time spent ruminating was summarized both daily (min/d) as well as in terms of DMI (min/kg DMI).
Lying behavior was recorded by Onset Pendant G data loggers (64k, Onset Computer Corporation, Bourne, MA) via leg orientation, validated by Ledgerwood et al. (2010).Loggers were attached to the medial side of the hind leg of each cow with veterinary bandaging (Vetrap Bandaging Tape, 3M, London, ON, Canada).Data were recorded every minute starting ~21 d before parturition until the end of the trial (21 DIM).
Data loggers were removed and replaced every 7 d to allow for data extraction.When replaced, data logger placement alternated legs on each cow to prevent sores.Data was extracted using Onset HOBOware Software (Onset Computer Corporation) before exporting to Microsoft Excel (Microsoft Corporation, Redmond, WA).Microsoft Excel macros (University of British Columbia, 2013) were used to process and summarize data by day as total lying time (min/d), lying bout frequency (bouts/d), and average bout length (total lying time/ bout frequency; min/bout).

Reticulorumen pH Monitoring
Wireless indwelling transmitting boluses (SX.1, Smaxtec GmbH, Graz, Austria) were orally administered randomly (taking parity and previous 305-d milk into consideration) at enrollment using a balling gun on a subset of cows (CC, n = 10; CG, n = 10; GC, n = 10; GG, n = 10) to measure reticulorumen pH (as validated by Gasteiner et al., 2015).The boluses recorded a value every 10 min for 24 h/d throughout the entirety of the study by wirelessly transmitting to strategically located base units.Data were downloaded twice weekly and combined into a continuous data set for each cow.Average daily pH values, as well as minimum and maximum pH for each day, were calculated and recorded.Time spent below and area under the curve (AUC) for a pH threshold of 6.0 were also calculated and recorded for each cow and day.This threshold was chosen because the pH boluses would typically stay within the reticulorumen, where the pH is traditionally 0.2 to 0.3 units higher than in the rumen (Sato et al., 2012;Falk et al., 2016;Neubauer et al., 2017).The AUC values were also standardized by dividing the AUC by DMI for each cow daily (as per Havekes et al., 2020a).

Statistical Analysis
Statistical analyses were conducted using SAS 9.4 (SAS Institute Inc., Cary, NC).Statistical significance was denoted at P ≤ 0.05 and tendencies were declared as 0.05 < P ≤ 0.10 on all main effects.For treatment by time interactions in repeated measure models, P-values for those interaction are described if P ≤ 0.05.The dry treatment period (d −21 to −1 relative to parturition) and lactation treatment period (0-21 DIM) were analyzed separately given different diets being fed across those periods.Before analyses, data were screened for normality using the UNIVARIATE procedure in SAS.The assumptions for normality were met for all measures except for duration at or below and AUC of reticulorumen pH levels of 6.0, as well as the AUC/DMI for a pH of 6.0, which were transformed Van Soest et al.: GLYCEROL, TRANSITION COW BEHAVIOR, AND RUMEN pH by taking the square root.Before selecting a square root transformation for the pH data, various potential transformations were tested to identify that which best normalized the data.A value of 1 was added to all measurements before transformation to ensure the calculation was conducted on a nonzero number.These data were back-transformed to include means and 95% confidence intervals.
Reticulorumen pH, rumination, feeding behavior (including sorting), and lying behavior data were analyzed in repeated-measures mixed-effect models, using the MIXED procedure in SAS, including cow within treatment as the subject of the repeated statement.Data from the dry cow period included the fixed effect of dry period treatment, days to calving, and the interaction of dry treatment by days to calving.For the postpartum measures the fixed effects of treatment during the dry period, treatment during the lactating period, the interaction between dry and lactation period treatments, and DIM were tested.Interactions of DIM with the dry and lactation treatments were also tested.The health status of each cow (treated or not for clinical disease) was initially tested as a potential covariate; this was not significant in any of the models and was, therefore, not included in any models.Covariance structure was selected for each individual model to determine best fit, according to Schwarz's Bayesian information criterion, which included compound symmetry, heterogeneous compound symmetry, first-order autoregressive, and heterogeneous first-order autoregressive.For all models, when significant interactions were detected, the PDIFF procedure in the LSMEANS statement, with Tukey-Kramer adjustment, was used to investigate the differences.For all feeding behavior measures the day before (−1 DIM) and the first 3 d following parturition (0-2 DIM) were omitted for all cows, as well as any other days they were not on automated feed bins and located in the maternity pens due to treatment for a clinical disease (except ketosis).To test whether sorting of the diets had occurred (i.e., sorting value differed from 100%), each fraction's sorting activity was tested within treatment, within the above model, for a difference from 100 using a t-test.

Dry Period
During the prepartum treatment period, cows fed GLY spent 20% more time feeding compared with those receiving CON (Table 3).Cows receiving CON prepartum had smaller, but more frequent meals per day in the 21 d before parturition compared with GLY supplemented cows (Table 3).This resulted in all cows consuming their dry cow TMR at a rate of 0.097 ± 0.006 kg/min and having 147 ± 6.0 min intervals between meals, with no treatment differences detected (Table 3).No difference in rumination activity, expressed either as min/d or min/kg DMI, was detected among treatment groups during the dry treatment period (Table 3).
No differences were detected in the particle distribution of the fresh dry TMR diets, including the long (>19 mm; P = 0.83), medium (<19 and >8 mm; P = 0.71), short (<8 and >4 mm; P = 0.96), and fine (<4 mm; P = 0.85) particles (Table 2).No treatment differences were detected for the sorting of long, medium, and fine particles of the diet in the dry treatment period.All cows sorted against the long and fine PSPS fractions (96.9 ± 1.3% and 98.5 ± 0.6% of expected consumption, respectively; Table 3), while consuming the expected proportion of the medium-sized particles of their respective TMR.Cows receiving GLY prepartum sorted in favor of the short particle fraction of the dry period TMR, while CON did not sort for or against the short particle fraction (Table 3).
During the 21 d before parturition, cows in the CON treatment group tended to have 8.4% more lying bouts per day and spent 3.9% more time lying down compared with those supplemented GLY prepartum (Table 3).The average lying bout duration for all cows during the prepartum treatment period was 102 ± 2.6 min/ bout, with no treatment differences observed (Table 3).
Cows had a mean reticulorumen pH of 6.63, a mean minimum pH of 6.33, and a mean maximum pH of 6.96 during the prepartum period, with no differences detected between treatments (Table 4).Similarly, no treatment differences were observed for time spent below or for AUC for the 6.0 pH cut point in the 21 d before calving.

Lactation Period
Cows that were fed GLY prepartum continued to have longer daily feeding time (18%; Table 5) from 3 to 21 DIM compared with cows fed CON prepartum.A dry-period treatment by lactation-period treatment interaction was detected for feeding rate and meal frequency (Table 5).Cows on CC, CG, and GC treatments tended to consume their PMR faster (+0.021 ± 0.008 kg/min) than those on the GG in the lactation treatment period (3-21 DIM; Table 5).For meal frequency during that postpartum period, CG cows consumed more meals (+2.0 ± 0.37 meals/d) compared with CC cows (Table 5).Cows fed CON following parturition consumed larger meals of PMR (+0.26 ± 0.056 kg/meal) in the lactation treatment period than those supplemented with GLY.No treatment differences were detected for meal length during the first 21 DIM, how-ever, cows fed CON postpartum had a longer interval between meals (+13.2 ± 3.6 min) during the lactation treatment period (Table 5).All cows spent 496 ± 9.1 min/d ruminating their feed following parturition, with no treatment differences detected (Table 5).However, when expressed as a proportion of DMI, cows fed CON prepartum tended to spend +0.9 ± 0.4 min/kg of DMI than those fed GLY prepartum (Table 5).
No differences in particle distribution were detected in the fresh lactation PMR treatment diets, including Treatment groups of either control (n = 30 cows) or glycerol (n = 30 cows, control diet + 250 g of glycerol product) for d −21 to −1 relative to parturition.
2 Sqrt = square root transformation; a value of 1 was added to all values before the transformation to ensure the square root could be calculated for all values equal to zero. the long (>19 mm; P = 0.87), medium (<19 and >8 mm; P = 0.89), short (<8 and >4 mm; P = 0.96), or fine (<4 mm; P = 0.75) particle distribution (Table 2).No treatment differences were detected for sorting behavior of the long and medium particles in the lactation treatment period; all cows sorted similarly against the medium particles (Table 5).Cows on GLY prepartum tended to consume a higher proportion of the short particles in their diet than CON cows (Table 5).Last, an interaction between dry and lactation treatments was observed for the sorting of the fine particles (Table 5) in the diet for d 3 to 21 relative to parturition.Cows on CC sorted against the fine particles, while the other 3 treatment groups did not sort for or against those particles (Table 5).
Cows that received CON prepartum continued to have 8.4% more lying bouts/d in the lactation treatment period than cows fed GLY prepartum (Table 4).Lying bout length tended to be 6.1% longer per bout in the first 21 d following parturition for cows that received GLY prepartum (Table 5).Total daily lying time was greater in the lactation treatment period for cows that received CON prepartum (+28 ± 8.6 min/d) or postpartum (+23 ± 8.8 min/d; Table 5) than cows fed GLY during the dry or lactation period.
During the first 21 d following parturition, reticulorumen pH for all cows averaged 6.55, with an average minimum pH of 6.20 and a maximum pH of 6.97 (Table 6).Cows receiving GLY prepartum tended to have a higher AUC for a pH of 6.0 compared with the CON treatment group (Table 6).No other treatment differences were detected for the time spent below, AUC, or AUC/DMI for a pH of 6.0 during the lactation treatment period.

DISCUSSION
In the current study, the supplementation of glycerol pre-and postpartum altered the time budgets of the cows, with cows spending more time eating pre-and postpartum, less time lying pre-and postpartum, and having less frequent, larger meals prepartum when receiving glycerol prepartum and with cows having slower feeding rates and smaller meals following parturition with postpartum glycerol supplementation.These findings may relate to the DMI differences observed as a a-c Rows with superscripts occur when there is an interaction between the dry and lactation diet treatments and different superscripts within row indicate significant differences (P < 0.05) between the 4 treatment groups (pairwise comparisons with Tukey adjustment). 1 Treatment groups: control diets both pre-and postpartum (CC), control diet prepartum and glycerol diet (250 g glycerol product/d) postpartum (CG), glycerol diet (250 g glycerol product/d) prepartum and control diet postpartum (GC), or glycerol diet (250 g glycerol product/d) both pre-and postpartum (GG). 2 Standard error of the highest treatment group. 3 Particle size was determined by a Penn State Particle Separator (Heinrichs, 2013) utilizing a 19-mm screen (long), 8-mm screen (medium), 4-mm screen (short), and a pan (fine).Sorting % = actual DMI of each particle size fraction/predicted DMI of each particle size fraction × 100.
The actual DMI for each fraction was calculated by subtracting the DM refused from the DM offered, as determined by the Penn State Particle Separator analyses.The amount consumed for each fraction was calculated by multiplying the actual total DMI of the diet by the DM percentage of that specific fraction in the fed TMR.A sorting % = 100 means no sorting occurred, sorting % >100 means sorting occurred in favor of that particle fraction, and sorting % <100 means sorting occurred against that particle fraction.*P < 0.05 (difference in sorting from 100%).
result of glycerol supplementation during the transition period as previously recorded in Van Soest et al. (2023).
Cows fed GLY prepartum, in the 21 d before calving, had a 20% increase in feeding time during that time period, likely representing the fact that those cows had 10% higher DMI (CON: 1.8 ± 0.05 vs. GLY: 2.0 ± 0.05% BW/d) before parturition (Van Soest et al., 2023), as greater feeding time has been associated with greater DMI both pre-and postpartum (De Mol et al., 2016;Daros et al., 2021).Despite lesser DMI, cows receiving CON prepartum had more frequent, smaller meals in the 21 d before calving compared with cows on GLY, similar to Havekes et al. (2020b), who observed fewer and smaller meals with less DMI of a dry cow diet with longer chopped straw.Limited work has been specifically focused on glycerol supplementation and its effect on feeding behavior.Previously, when liquid glycerol was fed at 11.5% of DM of the dry cow ration (compared with the current study targeting 1.7% of DM of a dry glycerol product), the feeding rate was reduced for the first 4 h following feed delivery in cows fed glycerol compared with control cows, but the glycerol-fed cows had higher feed consumption rates in the last 12 h of the day, resulting in no difference in DMI (Carvalho et al., 2012).We observed no treatment differences in feeding rate prepartum in the current study, which may be related to the lack of differences observed in sorting behavior.In a previous study, cows that had a higher level of sorting had a slower feeding rate than those with less sorting (Havekes et al., 2020a).Surprisingly, given the DMI differences, no difference in time spent ruminating were observed before parturition between treatment groups.Johnston and DeVries (2018) previously reported that DMI was positively associated with rumination time in lactating cows, although others suggest daily rumination time may be a poor indicator of DMI (Schirmann et al., 2012).
Cows fed CON prepartum tended to have more frequent lying bouts and spent more time lying down daily in the 21 d leading up to parturition.This difference in lying time roughly reflects the observed difference in feeding time.Cows fed CON prepartum consumed less DMI and spent less time feeding, thus having more free time to lie down compared with cows fed GLY, who ate more and for a longer period of time.Interestingly, increases in lying time prepartum have also been associated with poor health postpartum and increase NEFA concentrations (Piñeiro et al., 2019a); however, this is likely related to more sick cows consuming less feed.The cows receiving CON prepartum, as described in our companion article (Van Soest et al., 2023) had increased fat mobilization in early lactation, as evidenced by blood metabolite concentrations aligning with potentially more subclinical illnesses and cows experiencing suppressed DMI.In the present study, it was also interesting to observe greater lying time without an increase in rumination time, as the 2 have been associated frequently in previous studies (Beauchemin, 1991;Schirmann et al., 2012), but may be explained by the difference in DMI. 2 Sqrt = square root transformation; a value of 1 was added to all values before the transformation to ensure the square root could be calculated for all values equal to zero.Despite the differences in DMI and larger, less frequent meals prepartum, no differences in reticulorumen pH were observed before parturition.Researchers have previously recorded, when using different types of reticulorumen pH boli, no increase in acid accumulation or risk of SARA prepartum when cows were fed dry cow diets high in straw with a relatively low NFC (Havekes et al., 2020a,b).Interestingly, the various measures of reticulorumen pH in the current study were slightly higher (less acidic) than those observed by Havekes et al. (2020a,b), while also reporting values less than half for time below the pH cut point, AUC, and acidosis index (AUC/DMI), further supporting the lack of stress the current dry cow diet had on the rumen environment before parturition.Furthermore, dry cow diets with low starch and high peNDF content could expect to result in a less severe drop in pH (Khorrami et al., 2021).
Following parturition, cows fed GLY prepartum again had 18% greater feeding time, which reflected the 8.5% greater DMI recorded postpartum (CON: 2.04 ± 0.05 vs. GLY: 2.23 ± 0.05% BW/d) compared with cows fed CON prepartum (Van Soest et al., 2023).Researchers have previously recorded an association between feeding time and DMI for lactating dairy cows (De Mol et al., 2016;Johnston and DeVries, 2018) and no biologically relevant differences were observed for sorting of the diet.Although we recognize that AMS concentrate intake may affect feeding behavior, including sorting, Van Soest et al. (2023) detected no differences among treatments in AMS concentrate DMI in the first 21 DIM.As a result, these differences in PMR feeding behavior between treatments would not necessarily be explained by differences in AMS concentrate intake.Cows fed CON prepartum tended to ruminate longer per kilogram of DMI; however, this small difference is likely not biologically significant and results from the difference in DMI recorded.Similarly, when crude glycerol replaced corn grain in a lactating cow diet at 12.3% of DM of the ration, time spent ruminating was also unaffected (Zacaroni et al., 2022).
Cows supplemented with GLY both pre-and postpartum (GG) tended to consume their ration the slowest postpartum, as a result of longer feeding times for cows fed GLY prepartum and lower DMI for cows supplemented with GLY postpartum (CON: 18.4 ± 0.23 vs. GLY: 17.9 ± 0.23 kg/d; Van Soest et al., 2023).Interestingly, increased sorting behaviors would not have been the cause for reduced feeding rates (Greter and DeVries, 2011), as no differences in sorting were observed.However, one potential explanation for the slower feeding rate and smaller meals observed for cows receiving GLY postpartum may be the result of the potential increase in absorbed propionate from fermentation of the supplemented glycerol (Rémond et al., 1993).This potential increase in absorbed propionate causes satiety signaling to occur through increased oxidation of the tricarboxylic acid (TCA) cycle, terminating meals sooner or potentially slowing down the feeding rate (Oba and Allen, 2003;Gualdron-Duarte and Allen, 2017).Oba and Allen (2003) observed smaller, less frequent meals with increased levels of propionate absorption.Thus, this previous work supports the smaller meals and slower feeding rate observed in the current study with glycerol supplementation.However, it also contradicts the greater meal frequency observed with glycerol supplementation.Reasons for, and effects of, that greater meal frequency are unknown, especially given that no associated changes in rumen pH with postpartum glycerol supplementation were observed, as would be predicted (DeVries, 2019).Further, the additional 2 and 1.3 meals/d for CG and GG cows, respectively, compared with CC cows did not translate into additional DMI.Johnston and DeVries (2018) reported that each additional meal per day in lactating cows was associated with a 0.2 kg/d greater DMI.Thus, further research is needed to understand the mechanisms influencing meal-related behaviors, as well as the biological effects of those, associated with glycerol supplementation in early lactation.
An alternative explanation for control cows having less feeding time pre-and postpartum along with the larger, less frequent meals postpartum may result from health status because cows that are sick or experiencing pain can have altered feeding behaviors.For example, lame cows ate 0.77 g/min more fresh matter at 0.35 fewer meals per day compared with sound cows, and cows with ketosis had lower DMI (−10.4 kg of fresh matter), feeding time (−45.5 min/d), and feeding rate (−25.3g of fresh matter/min) compared with healthy cows (González et al., 2008).Van Soest et al. (2023) reported a reduced metabolic status for cows not receiving GLY pre-or postpartum, as defined by increased plasma NEFA and BHB concentrations, tended to increase preformed fatty acids in milk, and increased BW loss.Those cows with excessive energy mobilization may be experiencing suppressed feeding behaviors and DMI due to the higher levels of circulating ketones (Duffield et al., 2009).However, as we reported in Van Soest et al. (2023), cows on the current study experiencing higher levels of energy mobilization and ketones also had greater odds of increased haptoglobin.Because increased inflammation and reductions in glucose may also cause hypophagia (Horst et al., 2021), this may also explain the altered feeding behaviors we observed following parturition.
Comparable to prepartum lying behaviors, cows fed CON prepartum had more frequent, shorter lying bouts in the 21 d following parturition, while cows that received CON pre-or postpartum spent more time lying down per day than those supplemented with GLY.This difference in lying behavior relates back to the difference in postpartum DMI.We speculate those cows receiving GLY prepartum continued prioritizing feeding time over lying time, resulting in greater DMI postpartum.Previous researchers have demonstrated that cows consuming greater DMI before parturition continue consuming greater DMI postpartum (Akhtar et al., 2021).Cows on GG tended to consume feed at the slowest rate, which left less time for lying down.Although all cows, across treatments, had between 9 and 13 h/d of lying time, increased lying time (from 8 to 15 h/d) has been associated with increased risk of diseases, lameness, poor reproductive health, and culling (Piñeiro et al., 2019a,b).
The only difference detected from the postpartum reticulorumen pH data corresponded to prepartum GLY supplemented cows tending to have a higher AUC for a pH level of 6.0; however, this difference is not biologically significant.One-third of the cows on the trial experienced a pH value lower than 6.0 in the first 21 DIM, accounting for only 12.1% of the lactating cow days on the study (152 cow days experienced a below 6.0 out of 1,260 cow days on trial).The lack of depressed pH can be explained similarly to the dry cow ration, as the PMR and AMS concentrate fed both had a relatively low starch content, coupled with a high peNDF content of the PMR, limiting the chances of cows experiencing SARA (Khorrami et al., 2021).Differences in reticulorumen pH may have been expected, because researchers have previously observed that higher DMI (Dado and Allen, 1993), higher VFA concentrations (not measured in the current study; Burrin and Britton, 1986), and more NFC in the GLY supplemented diets (Allen, 1997) result in altered ruminal pH.Alternatively, when other researchers supplemented glycerol to steers and lactating dairy cows the proportions of propionate, butyrate, and valerate increased, while acetate decreased, resulting in either a decrease (Mach et al., 2009;Wang et al., 2009) or no change in ruminal pH (Khalili et al., 1997;Rico et al., 2012;Boyd et al., 2013).

CONCLUSIONS
Prepartum supplementation of a low level of dry United States Pharmacopeia grade glycerol product resulted in dry cows spending more time feeding, in fewer and larger meals, with less time spent lying down during the 21 d before calving.Following parturition, cows that received glycerol prepartum continued to feed for longer times and spent less time lying down per day.Postpartum supplementation of glycerol resulted in cows consuming smaller, more frequent meals, with shorter intervals between meals, and having less lying time in the first 21 DIM.These changes in behaviors are likely linked to the previously reported improved DMI and metabolic status observed in these cows.No negative effects of glycerol supplementation were observed during the transition period on time spent ruminating, reticulorumen pH, or ration sorting behavior.

3 AUC
= area under the curve.4 AUC/DMI = the area under the curve for the respective pH value (pH × min/d) divided by DMI (kg/d).5Thesedata were back transformed supplying the mean and 95% CI.

3 AUC
= area under the curve.4 AUC/DMI = the area under the curve for the respective pH value (pH × min/d) divided by DMI (kg/d).5Thesedata were back transformed supplying the mean and 95% CI.

Table 1 .
mm), shortVan Soest et al.: GLYCEROL, TRANSITION COW BEHAVIOR, AND RUMEN pH Ingredient and chemical composition (mean ± SD) of the dry and lactation treatments and the automated milking system (AMS) concentrate fed during lactation Heinrichs, 2013) fine (<4 mm) using a Penn State Particle Separator (PSPS;Heinrichs, 2013).Dried TMR, PMR, and dietary components were ground to pass through a 1-mm screen (Model 4 Wiley Laboratory Mill, Thomas Scientific, Swedesboro, NJ).Ground samples were pooled monthly for treatment rations and bimonthly for ration ingredients and sent to A & L Laboratory Services Inc. (London, ON, Canada) for analysis of ash (550°C; AOAC International, 2000, method 942.05),ADF (AOAC International, 2000, method 973.18),NDF with heat-stable α-amylase

Table 2 .
Van Soest et al.: GLYCEROL, TRANSITION COW BEHAVIOR, AND RUMEN pH Particle size distribution (mean ± SD) and nutrient content (mean ± SD) by particle size of the delivered treatment diets 1 PMR = partially mixed ration.2 Particle size was determined by a Penn State Particle Separator (PSPS; Heinrichs, 2013) using a 19-mm screen (long), 8-mm screen (medium), 4-mm screen (short), and a pan (fine).

Table 3 .
Van Soest et al.: GLYCEROL, TRANSITION COW BEHAVIOR, AND RUMEN pH Effects of glycerol supplementation on feeding, sorting, and lying behavior during the dry period (d −21 to −2 relative to parturition) (Heinrichs, 2013)f the highest treatment group.3Particlesizewasdetermined by a Penn State Particle Separator(Heinrichs, 2013)using a 19-mm screen (long), 8-mm screen (medium), 4-mm screen (short), and a pan (fine). Asorting % = 100 means no sorting occurred, sorting % >100 means sorting occurred in favor of that particle fraction, and sorting % <100 means sorting occurred against that particle fraction.*P < 0.05 (difference in sorting from 100%).

Table 4 .
Effects of glycerol supplementation throughout the dry period on measures of reticulorumen pH through the dry period (d −21 to −1 relative to parturition; mean ± SE)

Table 5 .
Van Soest et al.: GLYCEROL, TRANSITION COW BEHAVIOR, AND RUMEN pH Effects of glycerol supplementation on feeding, sorting, and lying behavior during the lactation (Lact) treatment period (d 3-21 relative to parturition)

Table 6 .
Van Soest et al.: GLYCEROL, TRANSITION COW BEHAVIOR, AND RUMEN pH Effect of glycerol supplementation on pH, area under the curve, and time spent below cut points in the reticulorumen in the lactating treatment period (d 1-21 relative to parturition; mean ± SE)