Value Added Opportunities from the Manufacture and Feeding of Silages Produced from Liquid Cheese Whey and Other By-products to Growing and Finishing Cattle and Beef Cows

Whey silage was produced in six separate growing (studies 1-4) and finishing (studies 7-8) and two beef cow trials (studies 5 and 6). Feedstuffs and proportions for the growing and finishing trials and beef cow trials are shown in Table 1. The liquid whey used for each study varied in dry matter percent and nutrient content (Table 1) and came from three separate manufacturing plants. Liquid cheese whey used in studies 5 and 6 were from the manufacture of yogurt, which is deemed "acidic." The liquid whey used in studies 1 through 4 and 7 and 8 were from the manufacture of various cheeses and were "sweet." All analyses reported in these studies were conducted at a commercial laboratory using AOAC (2000) standard procedures.
The ingredients were combined in a feed mixer for studies 1, 2 and 4 and packed in an open bunk silo. For studies 3, 5, 6, 7 and 8 the individual feedstuffs were combined in a feed mixer and placed in a silage bag. The whey silages were sampled 3-4 weeks later for nutrients, fermentation analysis and particle size (Tables 2, 3 and 4). Representative samples of each feedstuff were collected prior to and periodically during each of the studies for nutrient analysis.
Table 5 is a summary of the feedstuffs percentages (DMB) used for each treatment within each study. Cost of rations ($/ton DM) were: Study 1, C-$86.40, T-$52.49; Study 2, C-$72.31, T-$52.22; Study 3, C-$114.85, T-$84.87; Study 4, C-$118.35, T-$82.85; Study 5, C-$89.00, T-$46.00; Study 6, C-$55.96, T-$46.00; Study 7, C-$115.65, T-$106.77; Study 8, C-$122.56, T-$112.54. The whey silage costs do not include labor costs to make the silage.
Animal and pen numbers and initial weights for C and T groups are shown in Table 6. All studies, except trial 1 utilized British-based crossbred steer calves or cows (Table 7). Study l used growing Holstein heifers. The cows used in studies 5 and 6 were pregnant and fed under maintenance conditions to gain .23 kg/day. In the growing and finishing trials all calves had been processed similarly prior to trial initiation by receiving a Brucellosis vaccination, parasite treatment (Dectomax, Pfizer Animal Health, Exton, PA), 8-Way Clostridial vaccine (Pfizer Animal Health, Exton, PA) and intranasal respiratory product (BoviShield, Pfizer Animal Health, Exton, PA). Calves or cows were placed in pens and received rations as shown (Tables 1, 2 and 3). For all studies the C and T diets were isocaloric and isonitrogenous. Calves and cows were fed at 08:00 h such that there were no refusals. Individual calf and cow weights were recorded initially, every 28 days and at trial termination. Feed intake was recorded daily. Body condition score (BCS) and backfat (BF) measurements were taken initially and at trial termination for the beef cow studies (5 and 6). Live animal ultrasound using proprietary image analysis software was used for BF measurements and in trials 7 and 8 for BF, longisimus dorsi, and marbling (Brethour, 1991). Steers in trials 7 and 8 were slaughtered at the E.A. Miller Ltd (Hyrum, UT) facility and carcasses graded after a 24h chill.

Digestibility Trials
The C and T diets that were used in studies 3 , 4, 5 and 6 were fed to four ruminally cannulated beef cows in digestibility trials using replicated 2 ¡Ñ 2 Latin square designs. Results from study 6 were not available for this report due to laboratory delays. Cows were individually housed in open front 4 m X 10 m pens with concrete floors. All feedstuffs were fed once daily at 08:00 h for a 21 d adaptation period followed by a 6-d collection period. Rations were fed at amounts that were totally consumed daily. During the collection periods, fecal grab samples (300 g) were obtained at 08:00 h from each heifer. Samples of the total mixed ration (TMR), feces and individual feedstuff samples were also obtained daily throughout the collection period. Feed and fecal samples were dried at 60ƒµ C for 72 h and ground in a Wiley mill to pass a 1 mm screen. They were then analyzed for DM (AOAC, 2000; 934.01), for total N using a LECO CHN-1000 Combustion Analyzer (Sweeeney, 1989; Yeomans and Bremmer, 1991), ADF and NDF using an Ankom Fiber Analyzer (Ankom Technology, Fairport, NY). The NDF was assayed without sodium sulfite, with alpha amylase, and without residual ash. Acid insoluble ash (AIA) (Van Keulen and Young, 1977) was used as an internal marker to estimate apparent nutrient digestibility. Feed samples were dried at 60 C for 72 h and ground to pass through a 1 mm screen and proportionately composited by cow for each of the two collection periods. Analysis of feed samples followed the same procedures and methodologies as those used for the fecal samples.
On day 6, ruminal fluid was obtained from the ventral sac of the rumen via the rumen cannula of each heifer at 0.1, 2, 4, 6, 8, 10 and 12 h after feeding and immediately analyzed for pH using a combination electrode. Rumen fluid was strained through eight layers of cheese-cloth and 2 ml of the fluid was acidified with 18 ml of 6 N HCL. Volatile fatty acid (VFA) concentrations were measured in acidified samples using gas chromatography (Hewlett Packard 5890, Avondale, PA) with a 1.83 m X 2 mm ID glass column packed with GP 10% SP-1200/1% H3PO4 on 80/100 mesh Chromosorb W-AW.
Statistical analyses were performed using the MIXED procedure of SAS (SAS Institute, Cary, NC). Performance data from the four trials, including ADG, DMI, and FE, were analyzed using a completely randomized design with a repeated measures treatment structure. Pen was the experimental unit with monthly sampling periods as repeated measures of feed treatments. Feed treatment, sampling period, and their interaction were fixed effects and pen was a random effect. Dry matter and NDF digestibility were analyzed as a replicated 2 X 2 Latin square design by using animals as the experimental units with periods of the Latin square incorporated as repeated measures of feed treatments. Treatment and period were fixed effects and animal was a random effect. Volatile fatty acid and pH data were analyzed using the same model, except hour of ruminal sampling was incorporated as an additional repeated measure. Sampling hour and its interaction with feed treatment were considered fixed effects. In all statistical models, the Kenward-Roger option was used to estimate denominator degrees of freedom. The variance-covariance matrix was chosen for each statistical model in an iterative process wherein best fit was chosen based on the Schwarz¡¦s Bayesian Criterion. Least squares means were estimated and separated using the pdiff option when fixed effects were significant.