Effect of feeding increasing levels of whole cottonseed on milk and milk components, milk fatty acid profile, and total tract digestibility in lactating dairy cows

Dietary fat is fed to increase energy intake and provide fatty acids (FA) to support milk fat production. Oilseeds contain unsaturated FA that increase the risk for biohydrogenation-induced milk fat depression, but FA in whole cottonseed (WCS) are expected to be slowly released in the rumen and thus have a lower risk for biohydrogenation-induced milk fat depression. Our hypothesis was that increasing dietary WCS would increase milk fat yield by providing additional dietary FA without induction of milk fat depression. Four primiparous and 8 multiparous lactating Holstein cows (136 ± 35 DIM and 127 ± 4 DIM, respectively) were arranged in a replicated 4x4 Latin square design with 21 d periods. Treatments were WCS provided at 0, 3.4, 6.8, and 9.9% of dietary DM, and WCS was substituted for cottonseed hulls and soybean meal to maintain dietary fiber and protein. Treatment did not change milk yield. There was a treatment by parity interaction ( P < 0.05) for milk fat percent and yield with a quadratic decreased in primiparous cows but no effect of WCS in multiparous cows. Cottonseed linearly increased milk fat trans -10 18:1 in primiparous cows ( P < 0.05) but not in multiparous cows. Increasing WCS increased milk preformed (18 C) FA yield and partially overcame the trans-10 18:1 inhibition of de novo FA synthesis in the primiparous cows. Apparent transfer of 18 C FA from feed to milk decreased in all cows as WCS increased ( P < 0.05), but the magnitude of the change was greater in primiparous cows. Increasing WCS decreased total-tract apparent dry matter, organic matter, and neutral detergent fiber digestibility. There was no change in total FA digestibility. However, 18 C FA digestibility tended to be decreased in both parities and 16 C FA digestibility was quadratically increased in multiparous cows but not changed in primiparous cows. Total fecal flow of intact WCS increased ( P < 0.001) as WCS level increased, but fecal flow of intact seeds as a percentage consumed was similar across treatments. Fecal flow of intact seeds was greater in multiparous cows (4.3% vs. 1.1% of consumed). Plasma concentration of glucose, nonesterified FA (NEFA), triglycerides, and insulin were not changed. However, plasma urea-N (PUN) increased with increasing WCS. Plasma gossypol increased with WCS (0.08 µg/ml to 1.15 µg/ml) but was well below expected toxic levels. In conclusion, whole cottonseed maintained milk and milk component yield when fed at up to 9.9% of the diet to multiparous cows without concerns of gossypol toxicity, but primiparous cows were more susceptible to biohydrogenation-induced milk fat depression in the current trial. This highlights the interactions of parity with diet composition when feeding rumen available unsaturated fat to dairy cows.


INTRODUCTION
Adequate intake of energy and specific nutrients, including FA, is paramount to supporting lactation in high-producing dairy cows.Increasing dietary fat increases dietary energy density and feeding whole cottonseed (WCS) is commonly a cost-effective strategy to increase dietary fat in addition to providing protein and effective fiber.Experiments investigating feeding cottonseed to lactating cows were mainly conducted in the 1980s and early 1990s and demonstrated that WCS can be safely fed at up to 15% of the diet (DePeters et al., 1985;Sullivan et al., 1993;Zinn, 1995).However, very little work has been completed since then, and the effect of cottonseed in higher-producing cows and contemporary high corn silage diets is not clear.The previous work also focused on production responses and less on fat digestibility and impact on rumen biohydrogenation as it preceded the use of modern GLC columns capable of separating trans-FA.Additionally, there are limited dose titrations testing the effect of increasing dietary fat in general above basal fat levels, as much of the recent research has focused on dry fat supplements and delivery of specific FA.
Cottonseed, like other oilseeds, contains a relatively high level of total fat (18.6 ± 2.4 mean ± SD; NASEM, 2021), and similar to conventional soybeans and corn, cottonseed is high in linoleic acid (18:2 n6; > 50% FA).Consequently, this poses an increased risk for biohydrogenation (BH) induced milk fat depression (MFD) as linoleic acid is the primary precursor for trans-10, cis-12 conjugated linoleic acid (CLA) that is implicated in the inhibition of mammary lipid metabolism (Bauman et al., 2008).However, Harrison et al. (1995) fed diets containing 12% WCS and observed a 0.25 percentage unit increase in milk fat concentration compared with a lower fat control diet.The hard hull, high fiber lint, and lack of processing are commonly thought to decrease the rate of ruminal FA availability of WCS compared with other oilseed feeds and thus reduce the risk for BH-induced MFD (Harvatine et al., 2002).
This experiment was conducted to evaluate the effect of increasing dietary fat by increasing WCS in diets of contemporary dairy cows with a specific focus on quantifying fat digestibility and BH-induced MFD.The hypothesis was that fat in WCS would increase dietary fat intake without reducing FA digestibility, and allow for greater milk fat concentration and milk fat yield without induction of BH-induced MFD.Additionally, most experiments are conducted using only multiparious cows as the dry matter intake, milk composition, and growth requirements of first lactation animals differ from multiparous cows.However, a considerable portion of herds are first lactation animals and we expected that increasing dietary fat from WCS would differentially impact intake and milk yield of primiparous cows.

MATERIALS AND METHODS
Four primiparous and 8 multiparous cows (127 ± 4 and 136 ± 35 DIM, respectively) from The Pennsylvania State University Dairy Research and Teaching Center were housed in individual stalls and milked twice per day.The experiment was conducted from July to September 2019.All treatments and procedures were pre-approved by the Penn State University Institutional Animal Care and Use Committee (#200946398).Cows were blocked by parity and randomly assigned to treat-ment sequences in a replicated 4x4 Latin square design.Treatment periods were 21 d.There was a covariate period that fed the 0% WCS diet before the trial and the diet was reformulated after 10 d and fed for an additional 14 d due to lower than expected milk fat suspected to be BH-induced MFD.Treatments were commercially sourced Upland whole cottonseed included in the diet at 0, 3.4, 6.8, and 9.9% of diet DM (Table 1).The WCS was substituted for a blend of soybean meal (SBM) and cottonseed hulls (CSH) to balance CP and NDF on a fat-free basis.
Feed was delivered to cows as a TMR and fed once daily while cows were at the milking parlor in the morning.A base diet that included the ingredients common to all treatments was mixed in a stationary mixer (Electra-Mix 1062; I.H.Rissler Manufacturing, Mohnton, PA) and then the ingredients specific to each treatment (cottonseed hulls, soybean meal, and WCS) were then added to the base mix in a mobile apron mixer (Rissler 1050 Feedcart; I.H.Rissler Manufacturing, Mohnton, PA) and immediately delivered to the cows.Cows were fed at 110% of the previous day's intake.Dry matters of forages were calculated weekly (55°C forced air oven for 48 h) and diets were adjusted accordingly.

Feed, Refusal, and Fecal Collection and Analysis
Samples of each feed ingredient, orts, and feces were collected on the last 4 d of each period (~500 g samples).Feeds were composited by period.Orts were collected using the quartering method and fecal samples were collected every 9 h over the last 3 d and composited within cow for each period.One fecal composite was used for nutrient analysis and a second was stored for wet sieving of intact seeds.Feed, orts, and fecal samples were stored at −20°C.
For nutrient analysis, feed ingredients, orts, and feces were freeze-dried (Ultra 35-XL; Virtis CO. Inc., Gardiner, NY) and all except WCS were ground using a Wiley mill (A.H.Thomas CO., Philadelphia, PA) with a 1-mm screen.The WCS was ground using a cryogenic ball grinder (Retsch GmbH) at Dairyland Labs (Arcadia, WI).Subsamples were collected and analyzed for 105°C DM (overnight in a forced air oven).Ground feed samples were sent to Cumberland Valley Analytical Services (CVAS; Waynesboro, PA) for analysis of CP (method 990.03;AOAC, 2000), amylase-treated NDF (Van Soest et al., 1991), ADF (method 973.18;AOAC, 2000), and in vitro 240 h indigestible NDF (iNDF240).Starch was analyzed by an enzymatic method [(Karkalas, 1985); Hazyme, Centerchem, Norwalk, CT] after gelatinized with sodium hydroxide.Ash was determined by complete oxidation of dried samples in a muffle Pierce et al.: Running Heading: COTTONSEED FEEDING LEVEL furnace at 600°C for 5 h.Total FA concentration and profile of feed and feces were determined using direct methylation with 10% methanolic HCL as described by Sukhija and Palmquist (1988) with internal standards as described by Rico and Harvatine (2013).Fecal samples were also analyzed for NDF, iNDF240, ash, and FA concentration and profile as described above.
To determine the amount of WCS that passed intact, ~2000 g of composited fecal samples from each cow were weighed and wet sieved with a 4.75mm sieve by hand using a low-pressure laboratory eye wash wand (Nalgene, ThermoFisher Scientific, MA USA).Intact seeds on the sieve were counted, collected with forceps into polythene bags, freeze-dried, and weighed.Seeds were cut into pieces with scissors and analyzed for FAs as described above.Cottonseed intake was calculated based on the seed density observed by weighing 100 randomly selected seeds (mean = 7.6 g per 100 seeds).Fecal seed flow as calculated based on observed intact seed per unit of fecal DM and fecal flow determined by iNDF.
Indigestible NDF was used as an internal flow marker to calculate apparent total-tract (TT) digestibility using the marker ratio method as previously described by Huhtanen et al. (1994) and total fecal flow for calculation of the flow of intact seeds.

Milk Sampling and Analysis
Cows in the experiment were milked twice per day at approximately 0700 and 1830 h in a parlor and milk yield was calculated using automated milk meters [Afimilk (SAE Afikim), Kibbutz Afikin, Israel].Deviation in individual milk weights was adjusted based on a correction that accounts for the effect of milking (AM/ PM), day, cow, and stall in the parlor over the previous 7 d of each observation period (Andreen et al., 2020).Milk samples were collected at each milking on d 18 through 21 of each period, stored at 4°C in a bronopolbased preservative, and submitted for analysis for fat and protein by Fourier transform infrared spectroscopy (DairyOne Lab, Ithaca, NY).A subsample from each milking on d 20 and 21 was collected and composited based on milk weight within each period, then centrifuged at 1300 x g for 20 min at 4°C, and the fat cake was collected and stored at −20°C.Fat was extracted using hexane: isopropanol and FA were transmethylated using sodium methoxide and the resulting FAME quantified by gas chromatography with a flame ionization detector and a capillary column [SP-2560; 100 m x 0.25 mm (i.d.) with a 0.2-µm film thickness; Supelco Inc., Bellefonte, PA] as described by Baldin et al. (2018).

Blood Collection and Plasma Analysis
Blood was sampled from the coccygeal vessel before feeding at 0500 h (AM) and after feeding at 1430 h (PM) on d 20 of each period using potassium EDTA vacuum tubes (9 mL Griner Bio-One North America Inc., Monroe, NC).In addition, at the 0500-h sampling, blood was collected into 7 mL tubes containing heparin for plasma gossypol analysis.All tubes were immediately placed on ice and centrifuged within 30 min of sampling at 1300 x g for 15 min at 4°C.The plasma was extracted, transferred into microcentrifuge tubes, labeled, and stored at −20°C until further analysis.Plasma samples were analyzed for insulin, glucose, nonesterified FA (NEFA), urea nitrogen (PUN), and triglycerides (TAG) using enzyme-linked colorimetric assays as described by Niu et al., (2014).Plasma gossypol was analyzed using HPLC with simultaneous determination of total gossypol and the (+) and (−) gossypol isomers as described by Kim and Calhoun (1995).It is important to point out that the plasma gossypol levels were below the level of detection in all control cows (0% WCS) and a value of 0.03 µg/ml, which is less than 40% of the minimum reported value for WCS fed cows (0.08 µg/ml), was used to allow statistical modeling.

Statistical Analysis
Data were analyzed using the REML method of JMP Pro (version 14.0 to 16.0; SAS Institute Inc., Cary, NC) using the following model with cow as the experimental unit: Where Y ijkl was the dependent variable of interest, μ was the overall mean, C i was the random effect of cow (i = 1 to 16), P j was the random effect of period (j = 1 to 4), T k was the fixed effect of treatment (k = 1 to 4), R l was the fixed effect of parity (l = primiparous or multiparous), T k × R l was the interaction of parity and treatment, and e ijkl was the residual error.The overall treatment means and means within the primiparous and multiparous cows are reported in the tables.Fecal flow of intact WCS were log-transformed before analysis and back-transformed data are reported.
Plasma metabolites were observed twice over the day and data were analyzed using repeated measures in PROC MIXED (SAS Institute Inc., Cary, NC) using the heterogeneous autoregressive [ARH(1)] covariance structure, cow × period as the subject, and the Kenward Rogers denominator degree of freedom adjustment.
Studentized residuals outside of ± 3.0 were considered outliers and removed from the analysis.Uniformity of residuals was verified during analysis.Significance of main effects and preplanned contrasts were declared at P ≤ 0.05 and trends at P ≤ 0.10 and interactions declared at significant at P ≤ 0.05 and trends at P ≤ 0.10.Preplanned contrasts tested the linear and quadratic effects of increased WCS.

Dry Matter Intake and Milk Production
Overall, DMI was linearly decreased with increasing WCS (P = 0.03) driven by a large 2.8 kg decrease in multiparous cows (P < 0.01; Table 2).There was no effect of treatment or treatment by parity interaction for milk yield.There was a treatment by parity interaction for both milk fat percent and yield (interaction P = 0.03 and 0.08, respectively).In primiparous cows, milk fat concentration was quadratically decreased (P = 0.02) and milk fat yield tended to be quadratically decreased (P = 0.06) with lowest milk fat percent and yield at 3.4% WCS.Increasing WCS linearly increased milk fat trans-10 18:1 in primiparous cows from 0.77 to 1.91% of FA and increased trans-11 18:1.However, there was no effect of WCS on milk fat concentration and yield or trans-10 18:1 concentration in multiparous cows, although the concentration of trans-11 18:1 was increased indicating slowing of the second step of the normal biohydrogenation pathway and a buildup of this intermediate.Overall, there was an increased concentration of the other trans 18:1 isomers in primiparous cows, with the exception of trans-15 18:1 (Supplemental Data).All trans-18:1 isomers were linearly increased in primiparous cattle as WCS inclusion increased.
There was a tendency for WCS to linearly decrease milk protein concentration in multiparous cows (P = 0.07), but milk protein yield was unaffected by WCS level.Lastly, MUN was not impacted by parity or WCS inclusion level, indicating adequate balancing of CP between treatments.

Milk Fatty Acid Profile
Overall, WCS decreased the concentration of de novo synthesized FA in milk (P < 0.05; Table 3).Within parity, there was a 3.8 percentage unit decrease in milk de novo FA concentration for multiparous cows but only a tendency for a 2.1 unit decrease in primiparous cows.Daily yield of de novo synthesized FA was also lower in primiparous cows and tended to decrease with WCS inclusion (P = 0.07) with no interaction with parity (Table 3).
There was a treatment by parity interaction for the concentration of milk mixed source FA (16 C; Table 3).Mixed source FA were linearly decreased by 3.3 units with increasing WCS in multiparous cows (P < 0.01), but WCS tended to have a quadratic effect in primiparous cows with maximal concentration at 3.4% WCS (P = 0.07).Yield of mixed source milk FA was lower in primiparous cows but was not impacted by increasing WCS.
There tended to be an interaction of treatment and parity on preformed FA concentration (P = 0.11; Table 3).Concentration of preformed FA were linearly increased 7.7 units in multiparous cows (P < 0.01) and tended to be increased 2.8 units in primiparous cows (P = 0.06).Overall, primiparous cows had a higher concentration of preformed FA (41.7 vs 32.7% FA when fed the control).There was a treatment by parity interaction for yield of preformed FA with a linear increase only occurring in multiparous cows (P = 0.01; Table 3).
Feeding WCS resulted in a linear decrease in the sum of odd and branched chain (OBCFA), for multiparous cows without a change in the concentration of these FA in primiparous cows (Table 3 and Supplemental Table  3).

Nutrient Digestibility
Increasing WCS linearly decreased TT digestibility of DM, OM, and NDF in both primiparous and multiparous (P < 0.01; Table 4).There was no difference between parities in the digestibility of DM and OM, but NDF digestibility was lower in primiparous cows and WCS also decreased NDF digestibility to a greater extent in primiparous cows (11.7 vs. 9.0 percentage units).
Increasing WCS inclusion had no effect on total FA digestibility (Table 4).Overall, digestibility of 16 C FA linearly increased from 68.3 to 73.3% with increasing WCS (P = 0.04), although there was no effect observed in primiparous cows and a quadratic increase in multiparous cows (P = 0.01; Table 4).There was an overall tendency for digestibility of 18 C FA to be decreased with increasing WCS (P = 0.09), with no difference in primiparous cows and a tendency for a decrease in multiparous cows (P = 0.08).There were no effects of parity or parity by treatment interactions on FA digestibility.

Flow of Intact WCS to Feces
Flow of intact WCS to feces was quantified as the total number and total weight of intact seeds and as a percent and mass of seeds consumed (Table 5).The number of intact seeds and weight of intact seeds in feces increased linearly as inclusion of WCS increased regardless of parity (P < 0.01).However, there was no effect of increasing WCS when expressed as a percent of intake (P > 0.50).There were no treatment by parity interactions, but there was an impact of parity with multiparous cows passing 3 fold or more intact seeds compared with primiparous cows (P < 0.001).Treatments were whole cottonseed included in the TMR at 0, 3.4, 6.8, and 9.9% of diet DM in an iso-nitrogenous and iso-NDF substitution for cottonseed hulls and soybean meal.
2 P-values: Trt = main effect of treatment; Par = main effect of parity; TxP = interaction of treatment and parity; L and Q are the preplanned contrast testing the linear and quadratic effect of increased cottonseed inclusion.
The total amount of FA passing in intact seed was also increased with increasing WCS (P < 0.01) and was greater in multiparous cows (P < 0.001; Table 5).However, when expressed as a percent of FA consumed in WCS, there was no effect of increasing WCS level (P > 0.51), although it was lower in primiparous than multiparous cows (1.10 vs 3.64% of FA consumed in WCS at 6.8% WCS).

Transfer of FA to Milk Fat
There was a linear increase in intake and absorption of 18 C with increasing WCS in both primiparous and multiparous cows (P < 0.001; Table 6).The 18 C balance, calculated as 18 C consumed minus 18 C secreted in milk, increased linearly from an average of 94 g to over 280 g with increasing WCS (P < 0.001).There was a treatment by parity interaction for the apparent 18 C transfer, calculated as concentration in milk divided by FA intake, and true transfer, calculated as concentration in milk divided by absorbed FA (P = 0.01 and 0.05; Table 6).In primiparous cows, the apparent transfer was quadratically decreased, and the true transfer tended to be quadratically decreased while there was no effect in multiparous cows.

Plasma Metabolites and Hormones
Increasing WCS tended to linearly increase plasma glucose (P = 0.09), but there was no difference between primiparous and multiparous cows or time relative to feeding in the overall glucose concentration.Clearly, NEFA, TAG, and insulin concentrations in the control (0%) was the same in cows fed WCS and WCS level, parity or their interaction have no effect on these energy precursors.Plasma NEFA was higher before feeding compared with after feeding (160 ± 16.9 µEq/L vs 136 ± 16.9 µEq/L; P = 0.04; Suppl.Figure 1 Panel B) but was not impacted by WCS intake.Overall, increasing WCS intake increased PUN, due to an increase in primiparous cows (P = 0.01).

Plasma Gossypol
Plasma concentration of total gossypol and the positive (+), and negative (-) isomers were increased with increasing WCS (Table 8).There were no parity or treatment by parity interactions as plasma gossypol was linearly increased similarly in primiparous and multiparous cows (P < 0.001).Increasing WCS reduced the proportion of (+) gossypol isomers (44% vs. 56%).Treatments were whole cottonseed included in a TMR at 3.4, 6.8, and 9.9% of diet DM in an iso-nitrogenous and iso-NDF substitution for cottonseed hulls and soybean meal.Overall, (-) isomers remained the predominant isomer while consuming WCS.

DISCUSSION
Cottonseed is a major source of fat, fiber, and protein, making a selection of an ideal control impossible and previous studies have investigated cottonseed in substitution for diverse ingredients (Harvatine et al., 2002;de Souza et al., 2018a).Substitution for cottonseed hulls and soybean meal in the current trial allowed dissection of the specific impact of the FA provided in WCS from the effect of cottonseed fiber.Soybean meal was used to balance CP as both it and cottonseed have low rumen undegradable protein and soybean meal would be a common substitution on dairy farms."Book-values" are most often used when balancing diets with WCS and the recent NASEM (2021) reported an average of 18.3% FA with a standard deviation of 2.4 on 1,273 samples.Thus the 16.7% FA in the current study is within one standard deviation of the mean.The variation in WCS includes both true differences due to variety, growing conditions, and processing and technical variation from subsampling, grinding, and other laboratory errors.Grinding WCS is very challenging and often results in poor grinding of lint and poor sample homogeneity.We tested multiple approaches to processing WCS before determination of FA concentration including individual hand-cracking seeds, Wiley mill grinding after acid fuming to remove lint (Hahn Laboratories, Inc., Columbus, SC), and a cryogenic ball grinder (Retsch CryoMill).The acid pretreatment requires correction for fiber loss and may result in changes in the seed coat impacting fiber analysis.The cryogenic grinder resulted in a homogenous sample and was selected for use in the current project.
The decrease in DMI in multiparous cows was likely due to the increase in energy density as dietary fat increased and there was no change in energy corrected milk.Absorption of unsaturated FA can decrease intake as demonstrated with abomasal infusion of FA (Bremmer et al., 1998), but the lack of an increase in milk 18:2n-6 in the current trial indicates high rates of ruminal biohydrogenation and little change in unsaturated FA absorption.Intake was only numerically decreased in primiparous cows and care should be taken to not over-interpret because primiparous cows experienced diet-induced MFD which changes rumen fermentation and energy requirement.A previous study that fed cottonseed to primiparous cows also reported no change  Treatments were whole cottonseed included in a TMR at 3.4, 6.8, and 9.9% of diet DM in an iso-nitrogenous and iso-NDF substitution for cottonseed hulls and soybean meal.Cows not fed cottonseed were excluded from the analysis as we do not expect cottonseeds in feces of these cows. in DMI during the first period of their complete block trial, but noted a decrease in DMI during the second period (Harrison et al., 1995).BCS of cows were not evaluated due to the short experimental periods, so it is unclear whether primiparous cows may have partitioned the increased dietary energy toward growth.
The basal diet was balanced to have a lower risk of BH-induced MFD as the goal was to investigate the ability of additional fat to increase milk fat.However, there was an interaction of parity and treatment with WCS causing clear BH-induced MFD in primiparous cows.Interactions of parity and BH-induced MFD are not well described, but the clear induction of BHinduced MFD in the current study is likely due to differences in rumen dynamics.Interestingly, fecal flow of intact cottonseed in primiparous cows was only 25% of that in multiparous cows supporting differences in rumen function and digesta passage.This may be due to differences in feeding behavior, microbial communities, omasal size, intake relative to rumen volume, or other factors.It is not clear that primiparous cows are more susceptible to rumen disruption by unsaturated FA.Baldin et al. (2019) did not observe an interaction of parity and a diet that increased risk for MFD by increasing soybean oil and starch.The interaction in the current study may be specific to cottonseed or the basal diet conditions of the current project.Primiparous cows did have higher trans-10 18:1 when fed the control and were experiencing a moderate level of BH-induced MFD based on the meta-regression of Matamoros et al. (2020) Increasing dietary FA usually increases preformed FA in milk and decreases de novo synthesized FA (Glasser et al., 2008).Overall, increasing cottonseed reduced the concentration of de novo and mixed source FA and increased preformed FA.However, the magnitude of the change differed between primiparous and multiparous cows due to either physiological differences between parity or the occurrence of BH-induced MFD in primiparous cows.Primiparous cows had lower concentrations of de novo FA and a higher concentration of preformed FA in milk than multiparous cows, but  preformed FA only tended to be increased with increasing WCS in primiparous cows.It is interesting to note that although primiparous cows had a linear increase in trans-10 18:1 with increasing WCS, indicating progressive BH-induced MFD, there was a quadratic effect on milk fat yield as the increase in dietary FA supply allowed an increase in preformed FA to compensate for the decrease in de novo FA synthesis.In mice, Robblee et al. (2020) reported that FA supplementation did not overcome the effect of trans-10,cis-12 CLA on milk fat concentration and pup growth.
The basal diet was balanced to be low in dietary fat and thus the no WCS control had a high apparent and true 18 C FA transfer to milk.Primiparous cows had a higher transfer, likely due to lower intake relative to their milk yield.This would suggest that primiparous cows were either preferentially partitioning dietary FA to milk fat, were partitioning body reserves to milk fat, or were directing 18 C FA synthesized in non-mammary tissues to milk fat.Apparent transfer efficiencies > 100% are not uncommon, especially when feeding lower-fat diets (Khiaosa-ard et al., 2015).However, the apparent transfer rate for 18 C for all cows is much higher than the ~50% reported in abomasal infusion studies (e.g., Enjalbert et al. (2000)).The quadratic decrease in apparent and true 18 C transfer efficiency in primiparous cows was at least partially due to the occurrence of BH-induced MFD, while transfer efficiency was not changed in multiparous cows that utilized the additional dietary fat to increase milk fat preformed FA yield.
The lack of effect on milk protein yield and concentration is in agreement with previous research (e.g., Harrison et al., 1995;Abel-Caines et al., 1997;Firkins et al., 2002) and the review of WCS feeding by Arieli (1998) noted that very few trials observed a change in milk protein percentage or milk protein yield.Soybean meal was used to substitute for the protein in cottonseed.Although the amino acid profile is different, they are both predominantly rumen available, although their rates of digestion and rumen retention times likely differ.Additionally, dietary protein was well above adequate, and cows were post-peak, limiting the ability to observe differences in a protein-mediated response.
The decrease in total-tract DM and OM digestibility with increasing WCS was driven by a decline in NDF digestibility.Since cottonseed hulls were used to balance NDF between the treatments, this is not due to the digestibility of cottonseed NDF and likely is an associative effect of increasing dietary FA.Weld and Armenta- no (2017) reviewed the impact of increasing dietary FA on fiber digestion.They found little effect within the range of change investigated in the current experiment, but their analysis did not specifically investigate cottonseed feeding.As discussed above, trans-10 18:1 was elevated in the control diet, indicating that the basal diet included risk factors for disrupted rumen fermentation and may have been poised to respond to additional factors, such as an increase in unsaturated FA.The multiparous cows also had an increase in trans-11 18:1, indicating a slowing of the normal BH pathway.The slower ruminal release of WCS is expected to reduce the impact of the PUFA on rumen fermentation but may have interacted with other risk factors.It is also possible that dietary FA have a differential effect on forage vs cottonseed fiber.A lag in hydration of cotton linters has been proposed to have the potential to impede digestion (Palmquist, 1995).A recently presented meta-analysis observed a slight decrease in TT NDF digestion of 0.02 percentage units per unit of WCS (dos Santos Neto et al., 2022).However, Smith et al. (1981) found no difference in NDF and DM digestibility of cows fed up to 25% WCS per DM, although a trend toward depression in cellulose digestion was found.Moreover, the diet of cows fed 9.9% WCS contained 3.7% total FA, which is comparable to the level fed by Smith et al. (1981) and supports a possible impact on cellulolytic bacteria.Fatty acid digestibility normally decreases with increasing FA intake (Palmquist, (1991) and Boerman et al. (2015)), but increasing FA intake did not affect total FA digestibility in this study.Increasing WCS tended to decrease 18 C FA digestibility in multiparous cows, likely because multiparous cows had a higher passage of intact seeds in feces.Cottonseed oil is 24% 16:0 and digestibility of 16 C FA linearly increased with increasing WCS inclusion indicating that 16:0 in WCS is more digestible than the basal diet ingredients.Rico et al. (2017) also reported that 10% WCS increased 16 C and tended to decrease 18 C digestibility with no change in total FA digestibility, while (de Souza et al., 2018b) reported a small decrease in total FA when feeding 8.6% WCS.The seed coat of WCS is expected to slow the release of FA in the rumen but fails to protect unsaturation in the rumen as indicated by no change in milk 18:2 n-6.Fatty acids released in the rumen are expected to be highly available for intestinal digestion.Disruption of the seed coat by chewing and microbial digestion is expected to release FA in the rumen and increase the availability in the small intestine.The small percentage of seeds found intact in feces supports high availability of FA for absorption.
Intact cottonseeds are easy to identify when wet sieving and sometimes create questions on digestibility in the field, although quantification of cottonseed fecal flow has not been well investigated.In the current study, the passage of intact cottonseeds as a percent of WCS intake was low (<4.5% in multiparous and < 1.3% in primiparous) and was not changed by increasing WCS inclusion (Table 5).However, since passage as a percent of intake was constant, there was a linear increase in the absolute flow of intact seeds with increased WCS inclusion.Interestingly, passage of intact seeds was ~3-fold higher in multiparous cows than primiparous cows.This may be simply due to differences in rumen passage due to differences in intake and rumen volume  Treatments were whole cottonseed included in a TMR at 0, 3.4, 6.8, and 9.9% of diet DM in an iso-nitrogenous and iso-NDF substitution for cottonseed hulls and soybean meal.but may also be due to differences in omasal orifice size and function, resulting in selective retention in the rumen.Moreira et al. (2004) similarly reported a low fecal flow of intact seeds (<3.1% of intake) and lower passage in primiparous cows (e.g., 3.1 vs. 2.0%).Sullivan et al. (1993) recovered 11% of Pima seeds and 5% of "short-staple" seeds in feces when WCS was fed at 15% of ration DM.However, when feeding 15% WCS (Coppock et al., 1985) only recovered 0.7% of WCS consumed in multiparous cows, but increased to 11.3% when acid de-linted cottonseeds were fed.This may indicate that lint is important to retaining cottonseed in the rumen and increasing rumination (Coppock et al., 1985).The conventional (Upland) WCS fed in this study likely was adequate in lint to retain cottonseed in the rumen and stratify seeds within the rumen mat for continuous rumination.The low passage of intact seeds in feces in the current study does not support further processing for digestibility and processing of cottonseed is not advisable because the increased rate of rumen unsaturated FA availability will greatly increase the risk for BH-induced MFD as seen with previous studies feeding full-fat extruded cottonseed (Dhiman et al., 1999).
The FA concentration of intact seeds was measured to determine if all the WCS FA was available for digestion.The intact seeds retrieved from the feces had a slightly higher FA concentration than that fed (17.4 ± 0.34; mean ± SD) and did not differ by parity.The slight increase is expected due to the loss of the linters and indicates that FA within the seed coat were not available for digestion.
Gossypol is a toxin and total plasma gossypol is the best indication of the availability of free and bound gossypol in the diet (Mena et al., 2004).Whole cottonseed has been fed at up to 15% of the diet without gossypol toxicity or an increase in the erythrocyte fragility index (e.g., Chandler, 1992;Coppock et al., 1985b;Coppock et al., 1987;Harrison et al., 1995).However, increasing WCS inclusion increased the concentration of gossypol in plasma and notably increased the toxic (-)isomer across parities.A possible explanation is that the rumen microbes are binding to the (+)isomer and increasing the availability of the (-)isomers in the plasma.There was no indication of gossypol toxicity among cows as plasma gossypol levels (0.08 µg/ml to 1.15 µg/ml) were well below the toxic levels of 5 µg/ml (Millard, 2002) and 2.5 µg/ml (Adams et al., 1998) for PG and (-) isomer respectively.Cows fed 3.4%, 6.8%, and 9.9% WCS per DM consumed 1.06 kg/d, 2.07 kg/d, and 3.01 kg/d WCS, respectively, on a DM basis.Gossypol may be transferred to tissues and products (Lin et al., 1992) but reports of gossypol concentration in milk of cows consuming cotton by-products are limited.We did not analyze the gossypol concentration in milk, however Zhong, (2007) fed cows up to 15% WCS with free gossypol content of 23.3 g/d for 4 mo and reported total gossypol residue of less than 0.5 mg/kg in the milk indicating that residue of gossypol in milk from cows in our study is non-existent or at insignificant levels.

CONCLUSIONS
Increasing dietary fat by feeding WCS did not increase milk fat yield in multiparous cows with high milk fat because the additional preformed FA were offset by a decrease in de novo synthesized FA.However, increasing WCS in primiparous cows with moderate MFD resulted in a quadratic effect on milk fat with greater BH-induced MFD that was partially overcome by additional preformed FA availability.The susceptibility of multiparous and primiparous cows to BH-induced MFD appears to differ and may interact with dietary factors that have not been characterized.Total-tract FA digestion was maintained when increasing dietary FA by feeding WCS and only a small amount of seed passed intact in the feces.The greater passage of intact seeds in multiparous cows highlights differences in rumen function that have not been well explored.
Pierce et al.: Running Heading: COTTONSEED FEEDING LEVEL Pierce et al.: Running Heading: COTTONSEED FEEDING LEVEL

2P
-values: Trt = main effect of treatment; Par = main effect of parity; TxP = interaction of treatment and parity; L and Q are the preplanned contrast testing the linear and quadratic effect of increased cottonseed inclusion.

2P
-values: Trt = main effect of treatment; Par = main effect of parity; TxP = interaction of treatment and parity; L and Q are the preplanned contrast testing the linear and quadratic effect of increased cottonseed inclusion.
Pierce et al.: Running Heading: COTTONSEED FEEDING LEVEL Pierce et al.: Running Heading: COTTONSEED FEEDING LEVEL

2P
-values: Trt = main effect of treatment; Par = main effect of parity; TxP = interaction of treatment and parity; L and Q are the preplanned contrast testing the linear and quadratic effect of increased cottonseed inclusion.

COTTONSEED FEEDING LEVEL Table 1. Composition of diets with increasing inclusion of whole cottonseed
Pierce et al.: Running Heading:

Table 2 .
Pierce et al.:Running Heading: COTTONSEED FEEDING LEVEL Daily milk production and milk composition of cows fed increasing levels of whole cottonseed

Table 3 .
Milk fatty acid (FA) profile by source from cows fed increasing levels of whole cottonseed

Table 4 .
Nutrient and fat digestibility in cows fed increasing levels of whole cottonseed

Table 5 .
Appearance of intact cottonseed in feces of cows fed increasing levels of whole cottonseed

Table 6 .
Fatty acid absorption and apparent transfer of 18 C fatty acids from cows fed increasing levels of whole cottonseed in primiparous (Prim) and multiparous (Mult) cows

Table 7 .
Pierce et al.:Running Heading: COTTONSEED FEEDING LEVEL Concentrations of plasma metabolites of cows fed increasing levels of whole cottonseed

Table 8 .
Plasma gossypol concentration of cows fed increasing levels of whole cottonseed