Concentration of trans-vaccenic and rumenic acids in the milk from grazing cows supplemented with palm oil , rice bran or whole cottonseed

The changes in the concentration of trans-vaccenic (C18:1t-11) and rumenic (C18:2c-9,t-11) acids in the milk from cows grazing on Pennisetum clandestinum, fed a supplement containing palm oil, rice bran or whole cottonseed were evaluated. Three supplements were assessed: one control supplement containing palm oil (C), with a low concentration of linoleic acid mainly from palm oil, and two supplements containing rice bran (RB) or whole cottonseed (CS) as the main source of linoleic acid. Six Holstein cows (4.2±1.7 years of age, 532.5±50.7 kg BW, 125±29 days in milk and a milk yield of 21.7±5.8 kg d−1; Mean±SD) were assigned to each treatment using a double 3 × 3 × 3 Latin Square Design. Compared with treatment C, the milk fat concentrations of trans-vaccenic acid (31.1 and 23.8 g kg−1 of fatty acids for RB and C, respectively), rumenic acid (14.1 and 11.3 g kg–1 of fatty acid for RB and C, respectively) and unsaturated fatty acids (348.7 and 325.4 g kg−1 of fatty acid for RB and C, respectively) were higher for RB. Compared with C and CS treatments, the Δ9-desaturase index was higher for RB (0.37, 0.35 and 0.34 for RB, C and CS, respectively) and the thrombogenicity index was lower (3.09, 3.43 and 3.50 for RB, C and CS, respectively). The atherogenicity index was lower for RB treatment compared with C, but not compared with CS (1.85, 2.03, 1.97 for RB, C and CS, respectively). Supplementing rice bran to grazing dairy cows is a good alternative for producing a kind of milk beneficial to human health, due to its higher concentrations of trans-vaccenic and rumenic acids, unsaturated fatty acids and lower thrombogenicity and atherogenicity indexes.

The milk concentration of RU and TVA acids is mainly affected by diet (Khanal and Olson, 2004), and supplementation with lipid sources containing unsaturated FA increases such acids concentration (Dhiman et al., 2000;Khanal and Olson, 2004).The effect of oils on the concentration of TVA and RU in milk differs depending on their source (Bharathan et al., 2008;Bouattour et al., 2008;Flowers et al., 2008).Fat sources, rich in linoleic acid (LI) produce more TVA and RU than those containing linolenic acid (LN) (Bu et al., 2007).
The forage-to-concentrate ratio (Ueda et al., 2003;Loor et al., 2005b) and forage species (Mel´uchová et al., 2008;Vasta et al., 2008) are factors related to diet affecting milk fat concentration of TVA and RU.The concentration of these FA in milk is also influenced by the feeding systems (grazing or TMR) used (Kelly et al., 1998;Schroeder et al., 2003;Khanal et al., 2008).
Rice bran and whole cottonseed are frequently used in Colombia for preparing feeding supplements for dairy cows.These resources are rich in LI, 360 to 420 g kg −1 of FA for rice bran (de Campos et al., 2007) and 573 g kg −1 of FA for whole cottonseed (Dhiman et al., 1999;Khanal and Olson, 2004), and this is why they may be used to produce milk with a high level of TVA and RU.We found no reports in the literature assessing these resources as feed in grazing dairy cows to change the milk FA profile.In Colombian tropical high lands, milk production is based on grazing systems in which kikuyo (Pennisetum clandestinum) is the main grass (Correa et al., 2008), having a LN proportion in lipids between 530 to 595 g kg −1 of FA (Aguilar et al., 2009).Grazing on P. clandestinum and supplementing with lipids from a supplement (a fat intake from the supplement of near 670 g kg −1 of total fat intake) are characteristics of specialized livestock in the Bogotá Plateau (Colombia), suggesting this system has high potential to produce milk rich in TVA and RU acids.
We evaluated the changes in the concentration of TVA and RU acids in the milk from dairy cows grazing on P. clandestinum, and supplemented with palm oil, rice bran or whole cottonseed.

Material and Methods
All the experimental procedures were approved by the Bioethics Committee of Facultad de Medicina Veterinaria y de Zootecnia (School of Veterinary Medicine and Animal Production; Act 001 of 2009) of Universidad Nacional de Colombia.This experiment was carried out between May and July 2009.Average temperature is 13 °C (with variations from 0 to 20 °C); relative humidity ranges from 80 to 85%; and average yearly precipitation is 900 mm yr −1 , with two rainy seasons (April to May and September to November).The experiment term was 84 d, divided into three periods of 28 d each (20 d of adjustment to experimental diets, and a collection period of 8 d).The treatments consisted of three supplements: a control supplement (C) with a low level of LI, mainly coming from palm oil, and two supplements with similar concentrations of LI, mainly coming from rice bran (RB) or from whole cottonseed (CS; Tables 1 and 2).Three kg d −1 of supplement, 60 g of mineral salt and 5 g of Cr 2 O 3 were offered to grazing cows during each one of two milking processes (05.00 and 14.00 h).
Pasture was cut 50 d before the experiment and N fertilization was applied after 5 d (46 kg of N ha −1 ).The animals grazed in a 2.5 ha plot of P. clandestinum under strip system, moving the cord twice a day (morning and evening).A pasture proportion of low, medium and large amounts of forage, according to its height, was established and used to calculate the forage mass.To estimate the total available biomass, three squares of 0.5 m 2 each, representative of each height, were used to determine the average of each production level, which was multiplied by the proportion of each pasture level.To determine the daily offer and the plot size, a Global Positioning System GPSMAP 76CSx (Garmin Ltda., Kansas, EUA) was used to measure the forage areas.The grazing area was adjusted to ensure a daily offer of 40 g kg −1 of BW.On day 28 in each period, the animals were weighed immediately after their morning milking (05.00 h) and body condition score (BCS) was evaluated according to Edmonson et al. (1989).
The milk resulting from each milking (morning and afternoon) was weighted and mixed proportionally until reaching a volume of 200 mL, which was divided into two aliquots of 100 mL each.Sixty milligrams of potassium dichromate were added to one aliquot and then frozen at −20 °C.The other aliquot was used to determine its concentration of fat.
On days 20 and 27, forage samples (approximately 500 g) were collected simulating animal bite and using a hand plucking system (Cook, 1964;Muir, 2002).The samples collected were vacuum-packed, preserved at −20 °C, lyophilized (Martin Christ Alpha 1-4LD Plus), and ground using a 1 mm sieve Romer grinder.Forage samples were taken on days 13, 16, 19, 22, 25 and 28, dried for 48 h at 60 °C, ground using a 1 mm sieve Romer grinder, and homogenously mixed to obtain a composite sample that was used to determine the indigestible ADF (iADF).A sample of each supplement was collected (approximately 500 g), then ground in a 1 mm sieve Romer grinder.Supplement orts were collected between days 21 and 28, weighed, and stored at −8 °C until determining their DM.
Samples of ruminal fluid (350 mL) were collected on day 25 using an oro-ruminal probe (Haumptner ® ).The first 250 mL were discarded and the remaining 100 mL were filtered using two layers of cheese cloth.An aliquot was used to measure pH by a potentiometer (Beckman).Another sample was acidified with hydrochloric acid (Huang et al., 2009), frozen at −20 °C and used to determine the ammonium content.
Between days 22 and 27, two daily samples of feces were collected (05.00 h and13.00h), conserved at −20 °C, dried (60 °C for 48 h), ground using a 1 mm sieve Romer grinder and homogenously mixed, obtaining a composite sample for each animal.
Milk fat was extracted according to Hurley et al. (1987) andDíaz-González et al. (2002).One hundred milliliters of milk were centrifuged (15 min at 3,000 rpm) and the aqueous phase was discarded.Fifteen milliliters of a detergent solution (50 g of sodium hexametaphosphate and 24 mL of Tritón X-100 dissolved in 1 L of water) were added to the creamy supernatant.The solution was shaken twice and placed into a water bath (90 °C for 10 min).The upper fraction containing milk fat was carefully removed using a micropipette, and then stored at −20 °C.An aliquot of fat (100 µL) was solubilized in a 1 mL solution of chloroform:methanol (1:1 v/v).To methylate FA, 20 µL of a fat solution were placed in a vial with conic inserts, and 20 µL of methyl esterification reagent Meth-Prep II (0.2 M m-trifluoromethylphenyl-trimethylammonium hydroxide in methanol, Alltech Associates Inc., Deerfield, IL, USA) and 160 µL of chloroform:methanol (1:1 v/v) were added.Supplement fat was extracted according to Folch et al. (1957) and the same procedure described for milk was used for the methylation of FA.
Forage fat extraction and the methylation of FA were performed using a modification of the Garces and Mancha (1993) and Yamasaki et al. (1999) procedures.Absolute methanol (2,150 µL), 990 µL toluene, 66 µL sulfuric acid 98%, 1,000 µL of N,N-dimethylformamide and 2 mL of n-hexane were added.This mixture was placed into a water bath (80 °C for 2 h), left there for 5-10 min, stirred and the hexane supernatant was then recovered.Hexane was evaporated under nitrogen, and 300 μL of dichloromethane were added to the tube.
Fatty acid methyl esters were quantified using gas chromatography with a Shimadzu GC-2014, AOC-20i injector and an AOC-20C auto sampler.Fatty acid methyl esters were separated in a capillary column (Restek Rt-2560; 100 m × 0.25 mm i.d.× 0.2 µm layer).The injector and the FID temperatures were 260 °C and 270 °C, respectively; the temperature program was: 140 °C for 5 min, increased by 4 °C min −1 , increased to 190 °C and left for 32.5 min.Split ratio was 1:100 and He was used as a gas carrier with a pressure of 40.4 psi.Retention times were compared with standards (Food Industry FAMEX Mix cat 35077).
Fresh samples of ruminal fluid were used to measure pH (Beckman pHmeter) and frozen samples were used to determine ammonium, using the distillation procedure of the Kjeldahl method (AOAC-2001.11;AOAC, 2010).
The indigestible acid detergent fiber of feces was determined by the Sunvold and Cochran (1991) procedure and Cr by atomic absorption with a Shimadzu AA-7000 (Williams et al., 1962).
Forage intake was determined using Cr 2 O 3 as an external indigestible marker (Holden et al., 1994), and iADF as an internal indigestible marker (Sunvold and Cochran, 1991).Feces production (kg d −1 ) was determined according to Holden et al. (1994) and using a Cr recovery rate of 0.794 (Correa et al., 2009) , in which FP = feces production (kg d -1 ); EMD = external marker dose (g of Cr d -1 ); R = recovery rate; [EM F ] = fecal external market concentration (g of Cr g −1 of DM).
Fatty acid intake was determined according to the following equation: FAI = (FoI × [TFA] Fo × [FA] Fo ) + (SI × [TFA] S × [FA] S ), in which FAI = FA intake (g of FA d −1 ); FoI and SI = forage and supplement intakes, respectively (kg of DM d −1 ); [TFA] Fo and [TFA] S = total FA concentrations in forage and supplement, respectively (g of total FA g −1 of DM); [FA] Fo and [FA] S = specific FA concentrations in forage and concentrate, respectively (g of specific FA g −1 of total FA).Total FA intake was estimated according to Allen (2000).
Data were subjected to analysis of variance for a double 3 × 3 × 3 Latin Square Design using the GLM procedure of SAS (Statistical Analysis System, version 9.0) according to the following model: , in which: y ij(k)m = dependent variable; μ = overall mean; ϒ m = effect of square m; η(ϒ) im = effect of period i within square m; λ(ϒ) jm = effect of cow j within square m; τ (k) = effect of treatment k; ε ij(k)m = random error with mean 0 and variance σ 2 .All random effects were considered ~N(0,σ 2 e).Significant differences were declared at P<0.05 for main effects.Multiple comparisons among treatment means were performed by Tukey´s test.
Average fat intake from supplements and forage was 552 and 435 g d −1 , respectively.Supplements contributed with 559 g of fat kg −1 of total fat intake and 296 g of DM kg −1 of total DMI.As compared with C, LI+LN intakes were higher when cows were fed diets RB or CS (P<0.05)mainly due to supplementation because LI and LN intakes from forage were similar among the three diets (Table 4).
The production of milk from cows fed diet CS was lower than that of cows fed diets RB or C (P<0.05).Total milk solids, CP, ash and lactose concentrations were similar in the three treatments.When compared with cows fed diet RB, milk fat concentration was higher for cows fed CS, but similar when fed diet C (P<0.05; Table 5).
Compared with CS or C treatments, the production of milk total solids was higher for cows fed the RB diet (P<0.05)due to a higher lactose and ash production (P<0.05),since fat and protein production were similar among treatments.Compared with treatments C and CS, cows fed the RB diet 2 Fatty acid intake from supplement or forage g kg −1 of total FA intake.  2 C14:1 c-9 + C16:1 c-9 + C18:1 c-9 + C18:2 c-9,t-11 .
Cows receiving the RB diet had higher proportion of unsaturated FA in milk (P<0.05).Concentrations of milk long-(>C 16 ) and medium-chain (C 12 -C 16 ) FA were affected by the addition of supplements (P<0.05),but the concentration of short-chain FA (<C 12 ) was not.Comparing milk from cows fed CS, medium chain FA concentration was higher for cows fed diet C and lower for those fed diet RB (P<0.05).Milk long-chain FA concentration was lower for cows in treatment C, compared with cows in treatments RB or CS (P<0.05;Table 6).
Concentration of preformed FA (>C 17 ) in milk fat was lower for cows fed diet C compared with cows fed diets RB or CS (P<0.05), and there was a higher concentration of de novo synthesized FA (<C 16 ) in the milk from cows in the RB treatment (P<0.05;Table 6).
Milk RU/TVA, C14:1/C14:0 and C16:1/C16:0 ratios were lower in the milk from cows receiving the RB diet, compared with cows in the CS or C treatments (P<0.05).The index of Δ 9 -desaturase and the concentrations of this enzyme products were the highest for milk from cows in treatment RB, but the concentrations of their substrates were the lowest (P<0.05).Compared with the milk from cows fed diet C, milk fat of cows in treatment RB had a higher AI (P<0.05).The thrombogenicity index was the lowest for cows under the RB treatment (P<0.05;Table 6).
Live weight, BSC and ruminal fluid (pH and N-NH 3 ) were not affected by the type of supplement (Table 7).

Discussion
The effects of offering supplements containing palm oil, cottonseed or rice bran to grazing dairy cows on their milk FA profile were studied.Three supplements were prepared with palm oil, cottonseed or rice bran, having a similar nutritional composition and a similar amount of fat.The concentration of LI was similar for supplements containing CS or RB, and lower for the supplement containing palm oil.
It has been suggested that an increase in dietary fat of more than 80 to 90 g kg −1 of DM decreases DMI (Palmquist and Jenkins, 1980;Schauff and Clark, 1992).This effect is larger with a higher concentration of unsaturated FA in the supplement (Firkins and Eastridge, 1994;Bremmer et al., 1998).Other studies suggest that increasing fat level in the diet of grazing cows does not affect DMI (Bargo et al., 2003;Schroeder et al., 2004).In our experiment, the total dietary fat concentration was 58 g kg −1 of DM, in which the supplements contributed with nearly 55.8%.The intake of DM, CP, fat and fiber was similar among the diets, regardless of fat level and the concentration of unsaturated FA.Therefore, it can be suggested that the major effects of feeding supplements are mainly associated with the FA profile and source.
Comparing milk from cows receiving supplements containing palm oil or rice bran, the supplement containing cottonseed resulted in a lower yield of milk, total solids and a higher milk fat concentration.A higher milk fat concentration and a lower milk production in cows consuming the supplement containing cottonseed resulted in a fat production similar to that of the other two diets.Some researchers did not find any cottonseed effect on milk fat concentration or on milk fat production (Bitman et al., 1996), while others report an increase (Belibasakis and Tsirgogianni, 1995;Harrison et al., 1995) or a decrease (Wilks et al., 1991;Smith et al., 1993) in these parameters.
Differences in the yield of milk and total solids by feeding supplements containing cottonseed or rice bran cannot be explained by the differences in LI intake, since it was similar for treatments CS and RB.Differences in milk production have been associated with energy intake (van Knegsel et al., 2007).However, in our experiment, digestible energy intake calculated using markers, and net energy intake derived from NRC equations (NRC, 2001) were not different among treatments.A lower milk production with similar energy intakes by cows consuming the CS diet implies a lower efficiency in their use of digestible energy.
Some authors suggest that cottonseed can be used in milking cows at up to 3-4 kg cow −1 d −1 or 150 g kg −1 of their diet without negative effects (Coppock et al., 1987;Arieli, 1998;Zhang et al., 2007), although gossypol concentrations are different depending on the cotton varieties (Zhang et al., 2007).Gossypol intake was not determined in our experiment, but energy efficiency utilization was reduced in cows receiving supplement CS, though cottonseed intake (1.5 kg cow −1 d −1 ) was lower than the maximum level recommended in the literature.
Most authors suggest that increasing LI intake will increase milk TVA and RU (Dewhurst et al., 2006;AbuGhazaleh et al., 2007;Bu et al., 2007;Huang et al., 2008;Bouattour et al., 2008;Hervás et al., 2008;Bharathan et al., 2008).Linoleic and LN acids intake has been associated with milk TVA and RU (Loor et al., 2005a;AbuGhazaleh et al., 2007;Bu et al., 2007;Huang et al., 2008).In our study, LI intake was similar for cows fed RB or CS and near twice as much as those on diet C. Linolenic acid was not affected by supplementation and came mainly from forage, suggesting that differences in milk concentrations of TVA and RU were mainly explained by LI intake and source.
In our study, higher intakes of LI increased TVA and RU concentration in milk lipids for RB, but not for diet CS.Differences between diets CS and RB are not easy to explain since intakes of precursors of these two FA in milk were similar.Some authors suggest that whole seeds may produce a lower concentration of TVA and RU in milk lipid than processed seed, due to a lower ruminal availability of unsaturated FA in whole seeds, where FA are protected from ruminal biohydrogenation, resulting in a higher concentration of milk unsaturated FA (Dhiman et al., 2000;Khanal et al., 2005;Paradis et al., 2008).However, CS had lower concentrations of LI and LN acids and higher concentrations of stearic acid in milk lipids than RB, suggesting a more extensive biohydrogenation in the CS diet.Therefore, a low concentration of milk TVA and RU for cows fed the CS diet compared with those under the RB diet could not be explained by a low ruminal availability of LI, but rather by a more extensive biohydrogenation of LI in the CS diet.Pires et al. (1997) found a higher ruminal biohydrogenation of LI to C18:0 and a lower milk concentration of C18:2 when unprocessed cottonseed was fed compared with heat-treated and physically-processed cottonseed.Harvatine et al. (2002) found that when cottonseed is included in the diet of dairy cows (0 to 150 g kg -1 of DM) there is a linear increase in ruminal biohydrogenation of LI and outflow of C18:0 to the intestine.In our study we observed higher milk C18:0 concentrations in cows receiving the CS diet, compared with cows ingesting diets C or RB.This was possibly due to the fact that CS has a lower passage rate, and consequently remains longer in the rumen (Pires et al., 1997).Furthermore, RB and CS have similar concentrations of LI, but CS has a lower passage rate, suggesting RB will be a better alternative to increase milk TVA and RU.
In addition to increases in the concentration of TVA and RU in milk, other changes occurred in the milk FA profile, due to the dietary supplements in our experiment.When compared with cows fed diets C or CS, milk lipids of cows consuming RB supplement had higher concentrations of C12:0 and C14:0, and lower concentration of C16:0.Milk C12:0 and C14:0 are synthesized in the mammary gland (Garnsworthy et al., 2006), C14:0 being the main product of this process (Barber et al., 1997).The de novo synthesis of FA (less than 16 C) is predominant when cows are in a positive energy balance and do not depend on body energy reserves (Palmquist et al., 1993), while the concentration of preformed FA (long chain) in milk fat will be lower (Garnsworthy et al., 2006).In our study, cows receiving supplement RB had higher concentration of preformed FA in milk fat, suggesting a better energy balance than cows receiving supplements C or CS, regardless of similar energy intakes among diets, and a higher milk yield for the RB diet.Moreover, intake of C12:0 and C14:0 was higher for the C diet (P<0.05,data not shown) than for diets RB or CS.Therefore, differences in intake do not explain a higher concentration of these FA in the milk from cows receiving a supplement containing RB.
Variations in the ruminal microbial population could be related to change in FA profile in milk (AbuGhazaleh et al., 2003;AlZahal et al., 2008).Lipids present in ruminal microorganisms can be generated by the de novo synthesis (Schmidely et al., 2008;Váradyová et al., 2008) and bacteria have a higher concentration of these FA than protozoa (Or-Rashid et al., 2007).An increased concentration of milk C12:0 and C14:0 for the RB diet could be due to a decrease in ruminal population of protozoa when cows were fed the RB diet.However, protozoa populations were not measured in our study, and factors such as increasing levels of dietary starch or lower ruminal pH (Goad et al., 1998;Khafipour et al., 2009;Hook et al., 2011), which have been related to ruminal population of protozoa, were similar among diets.Some researchers have reported that feeding oil decreases the population of protozoa in the rumen, depending on the type and availability of FA (Toral et al., 2009;Reveneau et al., 2012).
Milk concentration of C16:0 for cows fed the RB diet was lower than for those fed diets CS or C. Milk C16:0 can originate from diet, de novo synthesis, adipose tissue mobilization or ruminal microorganisms (Linzell and Peaker, 1971;Barber et al., 1997).Compared with treatment C, the lower milk concentration of C16:0 for cows fed diet RB could be explained by a lower C16:0 intake.However, differences in milk C16:0 concentration between cows fed diets RB or CS could not be explained by differences in the C16:0 intake, because the two diets had similar levels and intake of C16:0.Additionally, milk concentration of C16:0 was similar for diets C and CS.A lower concentration of milk C16:0 for cows fed the RB diet did not explain a decrease in the de novo synthesis, because milk concentration of C 12 and C 14 increased with diet RB.The lower milk concentration of C16:0 also could not be explained by a lower adipose tissue lipid mobilization, since preformed FA concentration in milk was higher for cows fed diet RB.A lower C16:0 concentration in rumen microorganisms could explain a decrease in this FA in milk for cows on the RB treatment since it is one of its predominant FA (Or-Rashid et al., 2007;Váradyová et al., 2008), and it is produced by the de novo synthesis by ruminal microorganisms (Williams, 1986).
Changes in microbial populations by supplementation with different oils result in changes in the FA composition of rumen contents (Toral et al., 2012).Approximately 75 to 150 g kg −1 of lipids in the rumen digesta (Or-Rashid et al., 2007) and 750 g kg −1 of FA in rumen microorganisms (Jenkins et al., 2008) come from protozoa.Ruminal protozoa are an important source of lipids for the host animal (Or-Rashid et al., 2007), and have a higher concentration of C16:0 than bacteria (Váradyová et al., 2008).Fat decreases ruminal protozoa population depending on their source and FA composition and availability (Reveneau et al., 2012), and supplementation with sunflower oil alone or with marine algae in diets for sheep decreases the concentration of C16:0 in rumen content (Toral et al., 2009).Therefore, it is likely that RB supplementation decreased ruminal protozoa explaining the lower concentration of milk C16:0 for cows fed diet RB.
Compared with cows fed diet C, milk from cows fed diet RB has lower AI (P<0.05) and TI (P<0.01), which could be explained by a low milk concentration of C16:0 and a high milk concentration of unsaturated FA.A higher concentration of unsaturated FA can be beneficial for human health (Jacobs et al., 2011).Most of coronary diseases are the result of blood vessel obstruction by atheroma or thrombus (Ulbricht and Southgate, 1991).Therefore, the use of milk with low AI and TI could be an alternative for health markets.
The enzyme Δ 9 -desaturase is present in the mammary gland and its activity was calculated by the ratio of milk concentration of an unsaturated FA over milk concentration of a saturated FA of the same number of C. The ratio between products over substrates of Δ 9 -desaturase can be an estimate of its activity (Jacobs et al., 2011) or its concentration (Feng et al., 2004).
Although we did not determine the concentrations of sterculic and malvalic acids in the cottonseed we used, our results suggest that these FA inactivated Δ 9 -desaturase.C14:0 is mainly synthesized de novo by the mammary gland and C14:1 originates from the desaturation of C14:0 by Δ 9 -desaturase (Corl et al., 2001).In our experiment, the ratio between milk concentrations of C14:1/C14:0 was lower for cows fed CS than for cows fed the other two diets, suggesting that the inhibition of Δ 9 -desaturase was caused by cottonseed.This lower Δ 9 -desaturase activity may explain the higher concentration of C18:0, as well as lower RU in milk fat from cows which consumed the CS diet, as was the case in our trial.

Conclusions
When cows are grazing kikuyu and fed a supplement containing rice bran instead of a supplement containing palm oil or cottonseed their milk concentration of unsaturated fatty acids, trans-vaccenic and rumenic acids increase, and their milk atherogenicity and thrombogenicity indexes decrease.Rice bran is a better source of linoleic acid to modify the fatty acid profile of milk than cottonseed, since the former produces a higher concentration of transvaccenic acid, rumenic acid and unsaturated fatty acids, and presents lower atherogenicity and thrombogenicity indexes, resulting in a kind of milk beneficial for human health.

Table 2 -
Fatty acid (FA) profile of forage and supplements

Table 3 -
Nutrient intake by dairy cows grazing kikuyu and supplemented with low (palm oil) or high (rice bran or cottonseed) linoleic acid

Table 4 -
Linoleic and linolenic acids intake by dairy cows grazing kikuyu and supplemented with low (palm oil) or high (rice bran or cottonseed) linoleic acid 1 Means in the same row, followed by different letters are different according to Tukey's test (P<0.05).

Table 5 -
Milk composition and milk yield of dairy cows grazing kikuyu and supplemented with low (palm oil) or high (rice bran or cottonseed) linoleic acid

Table 6 -
Milk fatty acid (FA) profile of dairy cows grazing kikuyu and supplemented with low (palm oil) or high (rice bran or cottonseed) linoleic acid 1 Means in the same row followed by different letters are different (P<0.05).

Table 7 -
Body weight, body condition score and pH and ammonia in rumen fluid of dairy cows grazing kikuyu and supplemented with low (palm oil) or high (rice bran or cottonseed) linoleic acid