A Comparison of Eight Grades of Fat as Broiler Feed Ingredients

Transcription

1 A Comparison of Eight Grades of Fat as Broiler Feed Ingredients G. M. Pesti, 1 R. I. Bakalli, M. Qiao, and K. G. Sterling Department of Poultry Science, University of Georgia, Athens, Georgia ABSTRACT Seven samples of feed- or pet food-grade value. When the fats were incorporated into corn-andsoybean-meal-based fats (feed- and pet food-grade poultry greases, restaurant grease, white grease, animal/vegetable oil blend, palm oil, yellow grease) and one food-grade edible fat (soybean oil) were evaluated for quality and fatty acid composition. Active oxygen method (AOM) stability at 20 h ranged from 2 to 370 meq/kg; iodine value from 78 to 130 g/ 100 g; total moisture, insolubles, and unsaponifiables from 0.46 to 3.33%; initial peroxide values from 0.2 to 18.4 meq/kg; and free fatty acids from 0.08 to 21.0%. The ME of the fats ranged from 7.1 to 12.7 kcal/g and was positively correlated with AOM stability and iodine diets at 3 or 6%, no differences in live performance due to fat source were observed. Increasing fat level from 3 to 6% decreased feed conversion by 3.4 points (1.628 vs g/g). Feeding feed-grade poultry grease resulted in significantly smaller abdominal fat pads compared to the other fat sources. Only moisture, insolubles, unsaponifiables, and free fatty acids were significantly correlated with performance responses. Differences were noticed in abdominal fat pad color (lightness and redness) due to fat source. Differences in ME n were not reflected in differences in bird performance. (Key words: grease, fat, oil, metabolizable energy, broiler) 2002 Poultry Science 81: INTRODUCTION Many different grades of fat have been used as feed ingredients by the broiler industry. In recent years the composition of the different fat grades has varied more than in the past. Feed fats provide one essential nutrient (linoleic acid) and dietary ME. With corn-and-soybeanmeal-based diets, there is nearly always an excess of linoleic acid when fat is added to meet energy requirements. Therefore, the value of the various fats and oils in least-cost linear feed formulation is entirely dependent on their ME contents, and the ME content of the fats should be dependent on their digestibility and absorption. Many factors influence the absorption of fats. The chemical character of the fat itself is important. Absorption will be influenced by whether the fatty acids are free fatty acids (FFA) or triglycerides (Renner and Hill, 1961a,b; Young, 1961; Shannon, 1971; Garrett and Young, 1975). The position of the fatty acids in the triglyceride molecule affects absorption (Renner and Hill, 1961a). The digestibility of free fatty acids decreases with the increasing length of the carbon chain and saturation (Renner and Hill, 1961b). An increase in the content of unsaturated fatty acids in relation to saturated fatty acids increases the absorption of the saturated fatty acids (Young and Garrett, 1963). Oleic and linoleic acids readily form mixed micelles with bile salts in which saturated acids are then solubilized. The monoglycerides of these fatty acids appear to be more effective in this regard than the FFA. These saturated fatty acids are also more easily digested if they occupy the Number 2 position on the triglyceride, as the monoglyceride is more easily absorbed than the free fatty acid. The digestibility of fat increases as the chick ages (Fedde et al., 1960; Renner and Hill, 1960; Young, 1961). The work of Hakansson (1974) indicates that digestibility may be at its highest by 20 to 43 d of age. Age is also a factor in fat absorption in turkeys (Salmon, 1969; Whitehead and Fisher, 1975). Several other factors may affect fat absorption. They are as follows: bile supplementation of the diet (Fedde et al., 1960), calcium level of the diet (Fedde et al., 1960 and Hakansson, 1974), protein level of the diet (Biely and March, 1957), and intestinal microflora (Young et al., 1963). No differences between two strains of chickens in their ability to digest lard or tallow fatty acids were observed by Young et al. (1963). There have been few studies in which the absorption of fats and growth have been measured. Cullen et al Poultry Science Association, Inc. Received for publication April 19, Accepted for publication September 27, To whom correspondence should be addressed: gpesti@uga.edu. Abbreviation Key: A/V = animal and vegetable; AOM = active oxygen method; FCR = feed conversion ratio; FFA = free fatty acids; MIU = moisture, insolubles, unsaponifiables; NE = net energy; PUFA = polyunsaturated fatty acids. 382

2 FATS FOR BROILERS 383 (1962) observed that when fed in semipractical diets, the absorbability of choice white grease was higher than that of bleachable fancy tallow (95.6 vs. 86.7%). Differences in absorption paralleled differences found in the ME of the fats, 8.64 vs kcal ME/g, respectively. Yet, when diets containing 10% of each of the fats were fed to broilers for 8 wk, those fed the bleachable fancy tallow were significantly heavier than those fed the choice white grease. Similar observations were made by Leeson and Atteh (1995) with turkeys. They found significantly better weight gains from animal origin (tallow) than plant origin fats (corn oil, soybean oil, canola oil) or an A/V blend, despite that tallow had the lowest retention and diet ME values. Young and Artman (1961) fed chicks soybean oil or tallow at the expense of cellulose in semipurified diets. They observed consistent differences in feed conversion ratio (FCR) due to fat source and level; however, differences in growth due to feeding soybean oil or beef tallow were small or nonexistent when 24% protein was fed but larger when 28% protein was fed. These results show the importance of conducting digestibility and growth performance trials in evaluating fats as poultry feed ingredients. The experiments reported here were conducted to determine the ME of eight highly characterized fat samples [active oxygen method (AOM); FFA; moisture, insolubles, unsaponified (MIU); iodine values; peroxide values; and fatty acid profiles]. The samples were then fed to broilers during the grower phase (18 to 39 d) at 3 or 6% of the diet, to attempt to correlate ME and fat quality measurements with broiler performance and carcass composition. When variability in abdominal fat pad color was observed, fat pad color was quantified. MATERIALS AND METHODS Seven samples of feed fats were collected and shipped to us 2 [poultry grease (feed grade), poultry grease (pet food grade), feed fat from waste frying oil (restaurant grease), choice white grease (swine origin), animal/vegetable (A/V) blend, palm oil, and yellow grease]. A sample of food-grade soybean oil was purchased from a local retail outlet. Approximately 500 g of each source was sent to an independent laboratory for fat quality indicator analyses: 3 AOM at 20 h; iodine value, moisture, and volatiles by hot plate; insoluble impurities, unsaponifiable matter; peroxide value-initial, and FFA were measured by the official methods of the AOCS (1998; cd 12-57, cd Id-92, Ca 26-38, ca 3a-46, ca 6a-40, cd 8-53, and ca 5a-40, respectively). The fatty acid profiles of the 2 David S. Evans, CBP Resources, Inc., 2410 Randolph Avenue, Greensboro, NC Woodson-Tenent Laboratories, Inc., 3507 Delaware Avenue, Des Moines, IA Seaboard Farms, Athens, GA Petersime Incubator Co., Gettysburg, OH Minolta Chroma Meter CR-300, Minolta Corp., Ramsey, NJ samples were determined by gas liquid chromatography (Edwards, 1967). Day-old male broiler chicks (Ross Ross 208) obtained from a local producer 4 were used in both experiments. The birds were maintained on a 24-h light schedule. Feed and water were provided ad libitum. In Experiment 1, one-half of the chicks were randomly placed in Petersime electrically heated battery brooders 5 with raised wire mesh floors for ME n measurements at 9 to 10 d of age. The other half of the chicks were raised in floor pens and fed a corn-soy-based broiler starter (Table 1) for 35 d. At 35 d of age the birds were transferred to Petersime 5 growing batteries with raised wire mesh floors for ME n measurements at 39 to 40 d of age. There were three pens of eight chicks each per treatment. The birds were fed the basal diet (Table 1) with 6% glucose or one of the fat sources added at the expense of the diet and 0.1% chromic oxide from 0 to 10 d or 35 to 40 d of age. Excreta were collected for 48 h each when the birds were 9 and 10 d old or 39 and 40 d old. ME was determined by the method of Hill and Anderson (1958). For Experiment 2, chicks were randomly placed in floor pens (1.22 m 3.66 m) on pine wood shavings. Birds were fed a common starter diet for 18 d (Table 1). At 18 d, birds were weighed, and three pens of 33 males each were randomly allocated to each fat source and level. The pens were in two rooms of 24 pens each. Therefore, there had to be two replicates of each treatment in one room and one replicate in the other. At 39 d of age, birds and residual feed were weighed, and three birds per pen were chosen for processing. The birds were on feed withdrawal overnight (approximately 12 h) and were slaughtered on Day 40. Abdominal fat pads were collected into plastic bags. When considerable variation in fat pad color was noticed, the fat pads were frozen and later thawed for color determination in a reflectance spectrophotometer. 6 Data from both experiments were analyzed by oneway ANOVA for randomized complete block designs. The experimental unit was the pen mean. Significant treatment effects were determined using ANOVA with mean separation by Duncan s new multiple-range test when there were significant main effects (Steel and Torrie, 1960). In addition, two-way ANOVA was performed for Experiment 2 with a block for room included in the analyses. All analyses were conducted with the general linear models procedure of SAS software (SAS Institute, 1985). RESULTS AND DISCUSSION The fatty acid profiles of the samples were consistent with literature values (Table 2). The A/V blend and yellow grease samples provided had fatty acid profiles very similar to soybean oil, suggesting they were based on soybean oil (soybean oil might have been the major component). However, there were quality indicators that distinguished the samples (Table 3). The A/V blend had much lower FFA levels and a much higher initial

3 384 PESTI ET AL. TABLE 1. Composition of the experimental diets Experiment Ingredients and composition 1 Starter Grower 3 Grower 6 (%) Corn, ground Soybean meal Poultry by-product meal Poultry grease Fat source Defluorinated phosphate Limestone Common salt Vitamin premix Mineral premix DL-Methionine Cupric citrate Bacitracin BMD Aviax Composition by calculation 5 Protein, % ME, kcal/g variable variable Met, % Met + Cys, % Lys, % Vitamin premix provided per kilogram of diet: vitamin A (as retinyl acetate), 9,920 IU; cholecalciferol, 3,300 IU; vitamin E (as dl-α-tocopheryl acetate), 19.8 IU; menadione, 1.8 mg; vitamin B 12, 16.5 µg, thiamin, 1.65 mg; riboflavin, 9.9 mg; niacin, 58 mg; pantothenic acid, 16.5 mg; folic acid, 1.06 mg; pyroxidine, 2.88 mg; biotin, 0.08 mg. 2 Mineral premix provides per kilogram of diet: Mn, 120 mg; Zn, 100 mg; Fe, 60 mg; Cu, 10 mg; I, 2.1 mg; Se, 0.1 mg and contains calcium carbonate, ferrous sulfate, magnesium oxide, manganese sulfate, zinc sulfate, cupric sulfate pentahydrate, calcium iodate, and sodium selenite. 3 Alpharma, Fort Lee, NJ Phibro Animal Health, Exton, PA Based on NRC (1994) feed composition tables. TABLE 2. Fatty acid composition of samples studied Poultry grease Restaurant White A/V 1 Palm Yellow Soybean Fatty acid C:B 2 Feed grade Pet food grease grease blend oil grease oil Caprylic acid 8: Capric acid 10: Undecanoic acid 11: Lauric acid 12: Myristic acid 14: Pentadecanoic acid 15: Peak 7? Palmitic acid 16: Palmitoleic acid 16: Heptadeconoic acid 17: Peak 11? Stearic acid 18: Oleic acid 18: Linoleic acid 18: Arachidic acid 20: Linolenic/eicosenoic acid 18:3/20: Henicosanoic acid 21: Eicosadienoic acid 20: Behenic/eicosatrienoic acid 22:0/20: Arachidonic acid 20: Peak 21? 0.01 Lignoceric acid 24: Animal and vegetable. 2 Number of carbons:number of double bonds;?=means not determined.

4 FATS FOR BROILERS 385 TABLE 3. Quality indicators of fat samples studied Poultry grease Feed White A/V 2 Palm Yellow Soybean Analysis 1 Units Feed grade Pet grade fat grease blend oil grease oil AOM stability at 20 h meq/kg Iodine value g/100 g MIU % Moisture and volatiles by hot plate % Insoluble impurities % Unsaponifiable matter % Peroxide value initial meq/kg Free fatty acids % AOM = active oxygen method; MIU = moisture and volatiles by hot plate, plus insoluble impurities, plus unsaponifiable matter. 2 Animal and vegetable. peroxide value compared to soybean oil. The yellow grease had a much higher initial peroxide value than the A/V blend or soybean oil. Despite the low inclusion levels and number of replicates involved in the ME trial (Experiment 1), significant differences in the ME n of the fats were observed (Table 4). The poultry grease and palm oil samples consistently had the lowest values, and the three products apparently based on soybean oil consistently had the highest values. Very significant influences of AOM stability and iodine value on the ME of the fats were found (Table 5). These results are due to the very high AOM stability values and high iodine values for the high ME n products (A/ V blend, yellow grease, and soybean oil, Table 3). No differences in 39 d body weight or 18 to 39 d body weight gain due to fat source were observed (P > 0.05, Table 6). As expected, adding 6 vs. 3% fat to the diets resulted in significant improvements in FCR (Table 7). Despite observed differences in ME n of the fats, no differences in gains or FCR were observed due to fat source. Differences in mortality appeared to be random (Tables 6 and 7). Differences in abdominal fat pad, expressed as a percentage of body or carcass weight, were detected due to fat source (Table 8). The only consistent effect was for chicks fed poultry grease (feed-grade) to have the smallest fat pads. The fat with the lowest ME n resulted in the least amount of fat being deposited. There was considerable individual variation in the trait, and, clearly, more observations would be helpful in distinguishing between fats. It is clear that dietary fat level did not have a large effect on abdominal fat pad size (Table 9). Due to the larger number of degrees of freedom, the standard errors were much smaller for 3 vs. 6% dietary fat (than for fat source comparisons), yet the fat pad weight differences were less than the standard errors. The only fat quality indicators to significantly correlate with performance criteria are the three components of MIU (Table 10) with percentage of fat pads and carcass yield, and FFA with percentage of fat pads. Both MIU and FFA were high in the poultry grease (feedgrade) sample that produced the smallest fat pads, accounting for the indication of statistical significance. Choice white grease and palm oil produced the lightest color fat pads, whereas yellow grease and the A/V blend produced the fat pads that were the least light (Table 11). The A/V blend and poultry grease produced TABLE 4. Metabolizable energy content of different fat sources (mean ± SE), Experiment 1 1 Fat source Day 10 Day 40 Average (kcal/g) Poultry grease (feed grade) 6.94 ± 0.84 c 6.61 ± 1.63 ab 7.11 ± 0.93 b Poultry grease (pet food grade) 6.32 ± 0.82 c 5.27 ± 1.38 b 7.37 ± 0.61 b Waste frying oil 7.44 ± 0.60 c 6.52 ± 0.97 ab 8.36 ± 1.28 ab Choice white grease 8.20 ± 1.28 bc 7.26 ± 0.47 ab 9.13 ± 2.65 ab Animal/vegetable blend ± 0.61 ab 9.87 ± 0.14 a ± 1.27 ab Palm oil 6.47 ± 0.84 c 5.32 ± 1.38 b 7.62 ± 0.54 b Yellow grease ± 0.65 ab 9.66 ± 0.68 a ± 0.75 ab Soybean oil ± 1.50 a 9.56 ± 1.97 a ± 2.22 a Average 7.53 ± 0.52 x 9.35 ± 0.59 y ANOVA 2 df Pr > F Source Fat source Age Fat source age a c;x,y Values with no common superscript differ significantly (P < 0.05) when tested with Duncan s new multiple-range test. 1 Mean ± SE of triplicate determinations per treatment. 2 Pr > F = probability that observed differences occurred due to chance.

5 386 PESTI ET AL. TABLE 5. Correlation matrix of fat quality and fat ME (kcal/g), Experiment 1 1 Quality indicator Parameter 2 Fat ME (kcal/g) AOM Stability at 20 h r P Iodine value r P MIU r P Moisture and volatiles by hot plate r P Insoluble impurities r P Unsaponifiable matter r P Peroxide value initial r P Free fatty acids r P AOM = active oxygen method; MIU = moisture, insolubles, unsaponifiables. 2 r = correlation coefficient; P = probability that the correlation coefficient is significantly different from zero. fat pads with the most redness, whereas palm oil and soybean oil produced the least redness. No differences in yellowness were noted. Although the fatty acid profile of the A/V blend was very similar to the soybean oil, some added factor caused it to noticeably color fat pads. No differences in fat level or fat level by source interactions were observed for fat pad color. Apparently, carcass appearance may be influenced by the choice of fat, and this trait could be an important consideration for markets with color preferences. The lack of performance differences due to fat source despite differences in the ME of the fats observed in these experiments is consistent with the observations of Cullen et al. (1962) and Leeson and Atteh (1995). The phenomena labeled the associative dynamic effects of protein, carbohydrates, and fat by Forbes and Swift (1944) and the extra-caloric effect of fats first reported by Touchburn and Naber (1965) and confirmed by Jensen et al. (1970) may negate differences in ME n if the different fats exhibit different extra-caloric effects. If the extra-caloric effect of fats was due to an increased absorption of other nutrients, then it should have been measured by the ME assays. If the influence is not at the gut level, then it must be due to some basic TABLE 6. Effects of different sources and levels of fat on body weight (39 d), body gain (18 to 39 d), feed consumption (18 to 39 d), feed conversion ratio (FCR) (18-39 d), and mortality of male broiler chickens (39 d) fed different sources and levels of fats (mean ± standard error of three pens of 33 male broilers), Experiment 2 Fat level Body weight Body gain Feed consumed FCR 1 Mortality Source of fat (%) (kg) (kg) (kg) (g feed/g BG) (%) Poultry grease (feed grade) ± ± ± ± ± 1.65 Poultry grease (pet food grade) ± ± ± ± ± 0.01 Waste frying oil ± ± ± ± ± 0.95 Choice white grease (swine) ± ± ± ± ± Animal/vegetable blend ± ± ± ± Palm oil ± ± ± ± Yellow grease ± ± ± ± ± 0.01 Soybean oil ± ± ± ± ± 0.95 Poultry grease (feed grade) ± ± ± ± ± 0.95 Poultry grease (pet food grade) ± ± ± ± ± 1.65 Waste frying oil ± ± ± ± ± 0.95 Choice white grease (swine) ± ± ± ± Animal/vegetable blend ± ± ± ± ± 2.86 Palm oil ± ± ± ± ± 5.30 Yellow grease ± ± ± ± ± 0.95 Soybean oil ± ± ± ± ANOVA 2 df Pr > F Pr > F Pr > F Pr > F Pr > F Room Source Level Level source Error 31 1 Feed conversion ratio. 2 Pr > F = probability that the observed differences are due to chance.

6 FATS FOR BROILERS 387 TABLE 7. Main mean effects of different sources and levels of fat on broiler performance (mean ± standard error), Experiment 2 1 Body weight Body gain Feed consumed FCR Mortality (0 39 d) (18 39 d) (18 39 d) (18 39 d) (18 39 d) Source of fat n (kg) (kg) (kg) (kg/kg) ((%) Poultry grease (feed grade) ± ± ± ± ± 0.95 Poultry grease (pet food grade) ± ± ± ± ± 0.74 Waste frying oil ± ± ± ± ± 0.60 Choice white grease (swine) ± ± ± ± ± 3.19 Animal/vegetable blend ± ± ± ± ± 1.87 Palm oil ± ± ± ± ± 2.80 Yellow grease ± ± ± ± ± 0.60 Soybean oil ± ± ± ± ± 0.48 Pr > F Level (%) ± ± ± 0.03 a ± a 2.50 ± ± ± ± 0.03 b ± b 2.38 ± 0.84 Pr > F a,b Values with no common superscript differ significantly (P < 0.05) when tested with Duncan s new multiple-range test. 1 n = number of observations of 33 male broilers per pen; Pr > F = probability that the observed differences are due to chance. effect on bird metabolism. Rand et al. (1958) studied addition of fat to chick diets with different protein levels. They demonstrated that addition of corn oil to chick diets in carefully controlled studies with equalized feed and nutrient intakes resulted in an improvement in true growth, efficiency of protein use, and energy use. They suggested three possible mechanisms: increasing use of dietary energy or protein or by an unidentified growth factor. Rand et al. (1958) clearly demonstrated that use of fats is dependent on their source and level in the diet. Polyunsaturated fatty acids (PUFA) have been demonstrated to have effects on the synthesis of liver fatty acids (Edwards and Marion, 1963). The presence of longchain PUFA inhibit conversions of intermediate-chain PUFA (such as 18:1, 18:2, and 18:3) to long-chain PUFA (such as 20:3, 20:4, 20:5, and 22:6). The vegetable-based fats that are better absorbed have higher ME n values and have higher levels of PUFA. Long-chain PUFA may be inhibiting conversion of linoleic acid to arachidonic acid (at least compared to the more saturated, lower ME n fats). Because arachidonic acid is the precursor to many prostaglandin hormones, dietary fat source may be causing very basic changes in bird energy metabolism that are responsible for the extra-caloric effect of fats. Further studies should include measures of tissue fatty acids between fat sources to test this hypothesis. Carew et al. (1963) observed the effect of fats on feed consumption. They demonstrated that the growth-stim- TABLE 8. Effects of different sources and levels of fat on processing parameters of male broiler chickens (39 d; mean ± standard error), Experiment 2 Level Chilled carcass Fat pad Fat pad Source of fat (%) (% BW) (% BW) (% carcass weight) Poultry grease (feed grade) ± ± 0.13 c 1.64 ± 0.16 c Poultry grease (pet food grade) ± ± 0.05 ab 2.53 ± 0.09 ab Feed fat from waste frying oil ± ± 0.19 bc 2.18 ± 0.26 bc Choice white grease (swine) ± ± 0.13 ab 2.54 ± 0.18 ab Animal/vegetable blend ± ± 0.12 ab 2.54 ± 0.19 ab Palm oil ± ± 0.05 ab 2.32 ± 0.07 ab Yellow grease ± ± 0.14 ab 2.60 ± 0.20 ab Soybean oil ± ± 0.19 a 2.98 ± 0.24 a Poultry grease (feed grade) ± ± 0.09 bc 2.16 ± 0.13 bc Poultry grease (pet food grade) ± ± 0.09 ab 2.73 ± 0.12 ab Waste frying oil ± ± 0.14 ab 2.54 ± 0.18 ab Choice white grease (swine) ± ± 0.09 ab 2.43 ± 0.12 ab Animal/vegetable blend ± ± 0.12 ab 2.44 ± 0.17 ab Palm oil ± ± 0.27 ab 2.85 ± 0.36 ab Yellow grease ± ± 0.09 ab 2.44 ± 0.12 ab Soybean oil ± ± 0.30 abc 2.29 ± 0.41 abc ANOVA 1 df Pr > F Pr > F Pr > F Room Source Level Level source Error 31 a c Values with no common superscript differ significantly (P < 0.05) when tested with Duncan s new multiplerange test. 1 Pr > F = probability that the observed differences are due to chance.

7 388 PESTI ET AL. TABLE 9. Main mean effects of different sources and levels of fat processing parameters of male broiler chickens (39 d); mean ± standard error, Experiment 2 Chilled carcass Abdominal Abdominal yield fat pad fat pad Source of fat n 1 (%) (% BW) (% carcass) Poultry grease (feed grade) ± ± 0.11 b 1.90 ± 0.15 b Poultry grease (pet food grade) ± ± 0.05 a 2.63 ± 0.08 a Waste frying oil ± ± 0.12 ab 2.36 ± 0.16 ab Choice white grease (swine) ± ± 0.07 a 2.49 ± 0.10 a Animal/vegetable blend ± ± 0.08 a 2.49 ± 0.12 a Palm oil ± ± 0.15 a 2.58 ± 0.20 a Yellow grease ± ± 0.08 a 2.52 ± 0.11 a Soybean oil ± ± 0.19 a 2.63 ± 0.26 a Pr > F Level (%) ± ± ± ± ± ± 0.08 Pr > F a,b Values with no common superscript differ significantly (P < 0.05) when tested with Duncan s new multiplerange test. 1 n = number of observations of three birds per pen, each. 2 Pr > F = probability that observed differences are due to chance. ulating property of vegetable oils is not just a consequence of their high-energy value. Chicks fed diets containing soybean oil or corn oil consumed more ME than chicks fed comparable diets low in fat. Fisher and Wilson (1974) summarized many experiments and clearly established the relationship between dietary ME level, broiler growth, and feed conversion efficiency. By using similar data, Pesti and Smith (1984) showed that ME and fat level are significant predictors of growth and feed efficiency when included in the same model with dietary protein level. The findings of Pesti and Smith (1984), with data from many experiments, demonstrate that fats have exhibited the extra-caloric effect in several experiments, and feeding fats should have more value than simply the amounts of linoleic acid and energy that they may provide. Feed formulation models should include features for estimating growth from ME and fat level. The data reported here suggest that differences in the ME of fats may be indirectly proportional to the extra-caloric effect. As a result, the true feeding value of fats may be very similar despite differences in ME. The effects that fat and also specific fatty acids in the diet may have on liver lipid metabolism and basic metabolic rates must be given more consideration in future studies on factors influencing broiler growth and performance. The previous discussion indicates that the amount and kinds of fatty acids in the diet may be of great importance in modeling experiments for maximum profit feed formulation. A possible difficulty in directly comparing the ME n results of Experiment 1 with the performance differ- TABLE 10. Correlation matrix between bird performance and measures of fat quality, Experiment 2 1 Body Body Carcass Fat pad Fat pad Parameter 2 weight gain FC FCR Mortality yield (% BW) (% carcass) AOM stability at 20 h r P Iodine value r P MIU r P Moisture and volatiles by hot plate r P Insoluble impurities r P Unsaponifiable matter r P Peroxide value initial r P Free fatty acids r P FC = feed consumption; FCR = feed conversion ratio; AOM = active oxygen method; MIU = moisture, insolubles, unsaponifiables. 2 r = correlation coefficient; P = probability that the correlation coefficient is significantly different from zero.

8 FATS FOR BROILERS 389 TABLE 11. Main mean effects of different sources of fat on fat pad color (mean ± standard error), Experiment 2 Source of fat n 1 Lightness (L*) Redness (a*) Yellowness (b*) Poultry grease (feed grade) ± 0.80 bc 3.28 ± 0.56 ab ± 0.42 Poultry grease (pet food grade) ± 0.29 ab 2.25 ± 0.45 bc ± 0.20 Feed fat from waste frying oil ± 0.43 ab 2.77 ± 0.37 abc ± 0.32 Choice white grease (swine) ± 0.23 a 2.23 ± 0.22 bc ± 0.39 Animal/vegetable blend ± 0.49 c 3.74 ± 0.28 a ± 0.28 Palm oil ± 0.38 a 1.95 ± 0.33 c ± 0.30 Yellow grease ± 0.80 bc 2.41 ± 0.07 bc ± 0.45 Soybean oil ± 0.63 bc 1.95 ± 0.42 c ± 0.17 Pr > F Level (%) ± ± ± ± ± ± 0.17 Pr > F a c Values with no common superscript differ significantly (P < 0.05) when tested with Duncan s new multiplerange test. 1 n = number of means of 30 fat pads per pen, each. 2 Pr > F = probability that observed differences are due to chance. ences due to fat source (net energy deposition) in Experiment 2 could be related to adaptation to the fats. In Experiment 1, the birds were fed fat-supplemented diets from 0 to 8 d or from 35 to 39 d, which would represent periods of adaptation. In Experiment 2 the starter (preexperimental) diet contained 5.49% poultry grease and was fed until 18 d of age when the test grower diets were introduced. If the birds require different times to adapt to best use some or all of these fats, then the results might have been influenced by adaptation times (8, 4, or 0 d) in the experiments. The ME values in Table 4 are a measure of fat disappearance from the digestive tract; however, the performance results (Table 6) indicate the productive (or net) energy of the various fat sources. Differences in ME do not translate into differences in productive energy. For instance, Scott et al. (1976) concluded that the net energy is 75% of the ME of carbohydrates, 60% of the ME of proteins, and 90% of the ME of fats. The data in Tables 4 and 6 suggest that the net energy of the several fat sources tested is fairly similar. Therefore, the percentage of net energy in the ME of the fat sources must be variable. REFERENCES American Oil Chemists Society, Official Methods and Recommended Practices of the AOCS. 5th ed. American Oil Chemists Society, Champaign, IL. Biely, J., and B. March, Fat studies in poultry. 7. Fat and nitrogen retention in chicks fed diets containing different levels of fat and protein. Poultry Sci. 36: Carew, L. B., Jr., M. C. Nesheim, and F. W. Hill, The relationship of dietary energy level and density to the growth response of chicks to fats. Poultry Sci. 42: Cullen, M. P., O. G. Rasmussen, and O. H. M. Wilder, Metabolizable energy value and utilization of different types and grades of fat by the chick. Poultry Sci. 41: Edwards, H. M., Jr., Studies of essential fatty acid deficiency of the growing domestic cock. Poultry Sci. 46: Edwards, H. M., Jr., and J. E. Marion, Influence of dietary menhaden oil on growth rate and tissue fatty acids of the chick. J. Nutr. 81: Fedde, M. R., P. E. Waibel, and R. E. Burger, Factors affecting the absorbability of certain dietary fats in the chick. J. Nutr. 70: Fisher, C., and B. J. Wilson, Response to dietary energy concentration by growing chickens. Pages in: Energy Requirements of Poultry. T. R. Morris and B. M. Freeman, ed. British Poultry Sci. Ltd., Edinburgh, UK. Forbes, E. B., and R. W. Swift, Associative dynamic effects of protein, carbohydrate and fat. J. Nutr. 27: Garrett, R. L., and R. J. Young, Effect of micelle formation on the absorption of neutral fat and fatty acids by the chicken. J. Nutr. 105: Hakansson, J., Factors affecting the digestibility of fats and fatty acids in chicks and hens. Swedish J. Agric. Res. 4: Hill, F. W., and D. L. Anderson, Comparison of metabolizable energy and productive energy determinations with growing chicks. J. Nutr. 64: Jensen, L. S., G. W. Schumaier, and J. D. Latshaw, Extra caloric effect of dietary fat for developing turkeys as influenced by calorie:protein ratio. Poultry Sci. 49: Leeson, S., and J. O. Atteh, Utilization of fats and fatty acids by turkey poults. Poultry Sci. 74: Pesti, G. M., and C. F. Smith, The response of growing broiler chickens to dietary protein, energy and added fat contents. Br. Poult. Sci. 25: Rand, N. T., H. M. Scott, and F. A. Kummerow, Dietary fat in the nutrition of the growing chick. Poultry Sci. 37: Renner, R., and F. W. Hill, The utilization of corn oil, lard and tallow by chickens of various ages. Poultry Sci. 39: Renner, R., and F. W. Hill, 1961a. Factors affecting the absorbability of saturated fatty acids in the chick. J. Nutr. 74: Renner, R., and F. W. Hill, 1961b. Utilization of fatty acids by the chicken. J. Nutr. 74: Salmon, R. E., Soyabean versus rapeseed oil in turkey starter diets. Poultry Sci. 44: SAS Institute, SAS User s Guide. Statistics. Version 5 Edition. SAS Institute, Inc., Cary, NC. Scott, M. L., M. C. Nesheim, and R. J. Young, Nutrition of the Chicken. M. L. Scott and Associates, Ithaca, NY.

9 390 PESTI ET AL. Shannon, D. W. F., The effect of level of intake and free fatty acid content on the metabolizable energy value and net absorption of tallow by the laying hen. J. Agric. Sci. 76: Steel, R. G. D., and J. H. Torrie, Principles and Procedures of Statistics. McGraw-Hill Book Co., Inc., New York, NY. Touchburn, S. P., and E. C. Naber, Further evidence of extracaloric benefit of supplementary fats in turkey growing rations. Poultry Sci. 44:1422. (Abstr.). Whitehead, C. C., and C. Fisher, The utilization of various fats by turkeys of different ages. Br. Poult. Sci. 16: Young, R. J., The energy value of fats and fatty acids for chicks. I. Metabolizable energy. Poultry Sci. 40: Young, R. J., and N. R. Artman, The energy value of fats and fatty acids for chicks. 2. Evaluated by controlled feed intake. Poultry Sci. 40: Young, R. J., and R. L. Garrett, Effect of oleic and linoleic acid on the absorption of saturated fatty acids in the chick. J. Nutr. 81: Young, R. J., R. L. Garrett, and M. Griffith, Factors affecting the absorbability of fatty acid mixtures high in saturated fatty acids. Poultry Sci. 42: