SUMMARY
Reducing the particle size (PS) of feed ingredients is necessary and requires substantial energy expenditure. A majority of research on ingredient PS has considered the effects of corn and wheat PS on poultry performance. However, there is limited research investigating the effects of soybean meal (SBM) PS. Two experiments were conducted using expeller extruded SBM to measure the energy required to reduce soybean meal cake PS at the processor (experiment 1) and to determine the effects of various SBM PS on pullet performance from 0 to 17 weeks of age (experiment 2). The PS of SBM used in the current study was reduced using a hammermill fitted with one of three screens (2.4 mm, 5.6 mm, 7.9 mm) which represent the three treatments used in experiments 1 and 2. In experiment 1, hammermill screen size influenced motor load and power usage of the hammermill motor. The 2.4 mm screen required more energy and power to reduce the PS of SBM cake compared to all other screen sizes (P < 0.001). Experiment 2 used three identical diets that varied only in SBM PS. Overall, mash diets varying in SBM PS did not influence the measured pullet performance parameters which included FI, BWG, and FCR (P > 0.05). These data indicate that SBM may be manufactured using larger screen sizes to reduce energy usage at soybean processing facilities with no impact on ready-to-lay pullet development.
Key words
DESCRIPTION OF PROBLEM
Soybean meal (SBM) is the most common plant protein source used in poultry diets in the United States, with approximately 67% of SBM allotted for animal agriculture being used for poultry production. In 2019, approximately 20.2 million tons of SBM were used for poultry production in the United States (United Soybean Board, 2019). Poultry producers prefer SBM over other plant proteins such as sunflower meal or canola meal because of its ability to meet amino acid requirements (Ravindran et al., 2005).
Ingredient particle size (PS) is important to poultry producers due to its implications on energy usage at ingredient manufacturing facilities and overall bird performance. Reece et al. (1986) indicated that increasing hammermill screen size from 4.8 mm to 6.4 mm to grind corn reduced energy costs at the mill by more than 27%. Similarly, Wondra et al. (1993) reported that increasing hammermill screen size from 1.6 mm to 9.5 mm reduced energy consumption by 7.5 kWh/ton to grind corn. According to Deaton et al. (1989), feed PS reduction is the number one expenditure for feed manufacturing in the layer industry. When hens were fed diets varying in corn PS, Deaton et al. (1989) found no differences in BW, ADFI, or feed efficiency (P > 0.05). In a layer study conducted by Safaa et al. (2009), no differences were reported in BW or FCR (P > 0.05) but laying hens consuming coarse cereals demonstrated an increase in ADFI (P < 0.05). According to Portella et al. (1988a,b), both broilers and layers prefer coarser particles. While research has demonstrated the impact of hammermill screen size on energy usage and subsequent ingredient PS on broiler and layer performance (Deaton et al., 1989; Amerah et al., 2008; Safaa et al., 2009; Jacobs et al., 2010; Chewning et al., 2012; Pacheco et al., 2013; Rubio et al., 2020), there is limited research on ingredient PS of pullet diets, especially when considering the PS of SBM. Understanding that optimal SBM PS may influence the development of ready-to-lay pullets, the objectives of the current experiments were to determine the effects of hammermill screen selection on hammermill motor amperage and power use, soybean meal particle size, and subsequent 0 to 17-wk pullet performance.
MATERIALS AND METHODS
Experiment 1
Soybean Processing
All SBM was processed at a commercial soybean processor in Pennsylvania (Boyd Station, LLC, Danville, PA) where oil was extracted from soybeans via dry expeller extrusion. The resulting 5,443 kg allotment of SBM cake was ground using a Bliss Eliminator hammermill (Model E-22115-TF; 30-HP; 1,765 RPM) fitted with one of three screens that varied in size (2.4 mm [3/32”], 5.6 mm [7/32”], and 7.9 mm [10/32”]). Each screen generated SBM varying in PS, which represents the treatments in experiment 1. Treatments were generated in 3 replicate grinding runs per hammermill screen. A single allotment of SBM cake was used across all 9 processing runs because the SBM generated in experiment 1 was used to feed pullets in experiment 2. This ensured the nutritional value of the SBM would not be affected by the source of the soybeans. Understanding that throughput would be limited by hammermill screen size, the 2.4 mm screen was used first. The feed screw auger carrying SBM cake to the hammermill was fitted with a variable speed drive adjusted to 20 RPM so that hammermill motor amperage would not exceed 30 Amps. A replicate run of SBM cake processing was 4 min and 45 s. The 20 RPM auger speed and 4 min and 45 s processing run was used with all other hammermill screens. The motor load and power usage of the hammermill motor was recorded every second via a computer system at the processing facility. Each SBM processing replicate was augered into a labeled supersack and transported to the Penn State Poultry Education and Research Center.
Particle Size Analysis
Three samples were taken from each SBM replicate supersack, totaling nine samples per treatment. Dry sieving of the SBM samples according to the ASAE method S319.4 (ASABE, 2017) was completed for particle distribution using a Ro-Tap tester and agitators (Ro-Tap tester, Model RX-29, WS Tyler company, Mentor, OH). A Ro-Tap tester equipped with 14 sieves ranging from 4,760 µm to 38 µm and a pan was used to process 100 g samples for 10 min. A dispersing agent was not included. After 10 min, individual sieves were weighed to measure particle size distribution and calculate average geometric mean diameter, which will be described herein as PS.
Experiment 2
Experimental Diets
A 5-phase feeding program followed the 2019 Hy-Line W-36 management guide and all diet phases were formulated using Hy-Line W-36 nutrient recommendations (Table 1). Nutrient analysis was conducted on the diets (Table 2) where only SBM PS varied between treatment diets. Duplicate PS analysis using the Ro-Tap tester was conducted for all dietary treatments across all feeding phases (Table 2).
Table 1. Diet formulations and calculated nutrients used to feed Hy-Line W-36 pullets from 0 to 17 wk of age.1
| Ingredients | Starter 1(0–3 wk) | Starter 2(3–6 wk) | Grower(6–12 wk) | Developer(12–15 wk) | Pre-Lay(15–17 wk) |
|---|---|---|---|---|---|
| Empty Cell | % | ||||
| Corn | 60.39 | 65.68 | 59.32 | 63.12 | 63.27 |
| Expeller extruded soybean meal | 32.42 | 27.73 | 18.20 | 17.75 | 19.02 |
| DDGS | – | – | 10.00 | – | 8.85 |
| Wheat middlings | – | – | 8.04 | 14.62 | – |
| Soybean oil | 2.59 | 1.87 | – | – | – |
| MonoCalcium phosphate | 2.30 | 2.30 | 1.86 | 2.03 | 2.07 |
| Limestone | 1.17 | 1.21 | 1.46 | 1.43 | 5.78 |
| Vit/min premix2 | 0.40 | 0.40 | 0.40 | 0.40 | 0.40 |
| Salt | 0.44 | 0.41 | 0.37 | 0.44 | 0.38 |
| L-Lysine HCl | 0.07 | 0.12 | 0.16 | 0.06 | 0.07 |
| DL-Methionine | 0.19 | 0.23 | 0.14 | 0.13 | 0.15 |
| L-Threonine | 0.03 | 0.06 | 0.05 | 0.02 | 0.02 |
| Calculated nutrients | |||||
| ME (kcal/kg) | 3,031 | 3,031 | 3,009 | 2,965 | 2,987 |
| CP (%) | 20.00 | 18.25 | 17.79 | 16.06 | 16.57 |
| Calcium (%) | 1.00 | 1.00 | 1.00 | 1.00 | 2.50 |
| Available phosphorus (%) | 0.50 | 0.49 | 0.47 | 0.45 | 0.48 |
| Sodium (%) | 0.18 | 0.17 | 0.17 | 0.18 | 0.17 |
| Dig. Lys (%) | 1.05 | 0.98 | 0.88 | 0.76 | 0.78 |
| Dig. Meth (%) | 0.47 | 0.48 | 0.40 | 0.36 | 0.39 |
| Dig. TSAA (%) | 0.74 | 0.74 | 0.66 | 0.60 | 0.62 |
| Dig. Threonine (%) | 0.69 | 0.66 | 0.60 | 0.52 | 0.55 |
- 1
-
Diets formulated using the 2019 Hy-Line W-36 recommendations.
- 2
-
Supplied per kg of diet: 0.022% iodine; 1.1023% iron; 0.1764% copper; 0.006% selenium; 2.6455% manganese; 2.2046% zinc; 2,643,172 IU vitamin A; 881,057 IU vitamin D-3; 6,608 IU vitamin E; 440 mg of thiamin; 1,762 mg of riboflavin; 13,216 mg of niacin; 3,084 mg of pantothenic acid; 133,678 mg of choline; 220 mg of folacin; 22 mg of biotin; 3,524 mcg of vitamin B12; 660 mg of menadione; and 24,948 mg of ethoxyquin.
Table 2. Analyzed nutrients and particle size1 of diets fed to Hy-Line W-36 pullets from 0 to 17 wk of age.
| Diet phase | Screen size | Gross energy (kcal/kg) | Crude protein(%) | Crude fat(%) | Crude fiber(%) | Ash(%) | Moisture(%) | Particle size2(µm) |
|---|---|---|---|---|---|---|---|---|
| Starter 1 (0–3 wk) | 2.4 mm | 4,145 | 20.13 | 7.05 | 2.40 | 6.96 | 7.96 | 566.22 ± 2.01 |
| 5.6 mm | 4,167 | 19.75 | 7.06 | 2.50 | 6.79 | 8.38 | 613.42 ± 2.10 | |
| 7.9 mm | 4,145 | 19.81 | 6.79 | 2.40 | 6.80 | 8.47 | 635.60 ± 2.27 | |
| Starter 2 (3–6 wk) | 2.4 mm | 4,012 | 18.31 | 5.59 | 2.40 | 6.30 | 8.77 | 570.69 ± 2.23 |
| 5.6 mm | 4,057 | 17.81 | 6.45 | 2.20 | 5.78 | 8.87 | 639.06 ± 2.38 | |
| 7.9 mm | 4,034 | 18.44 | 5.62 | 2.50 | 5.65 | 8.74 | 675.61 ± 2.40 | |
| Grower (6–12 wk) |
2.4 mm | 3,924 | 17.31 | 4.31 | 3.60 | 5.51 | 10.21 | 582.80 ± 2.32 |
| 5.6 mm | 3,968 | 17.06 | 4.36 | 3.30 | 6.50 | 11.08 | 655.60 ± 2.30 | |
| 7.9 mm | 3,946 | 16.88 | 4.24 | 3.20 | 6.13 | 10.22 | 672.67 ± 2.37 | |
| Developer (12–15 wk) |
2.4 mm | 3,858 | 15.94 | 3.73 | 2.90 | 5.73 | 10.19 | 551.87 ± 2.38 |
| 5.6 mm | 3,836 | 15.38 | 3.96 | 3.10 | 5.99 | 10.38 | 587.92 ± 2.41 | |
| 7.9 mm | 3,880 | 15.94 | 3.74 | 3.00 | 5.67 | 10.20 | 625.42 ± 2.48 | |
| Prelay (15–17 wk) |
2.4 mm | 3,759 | 15.75 | 3.94 | 2.60 | 9.70 | 9.46 | 558.54 ± 2.33 |
| 5.6 mm | 3,726 | 16.38 | 4.21 | 2.50 | 9.08 | 9.44 | 661.09 ± 2.41 | |
| 7.9 mm | 3,726 | 16.19 | 3.96 | 2.50 | 10.07 | 9.02 | 702.65 ± 2.57 |
- 1
-
Particle size was determined using a Ro-Tap tester, Model RX-29 (WS Tyler company, Mentor, OH).
- 2
-
Means presented are the average of duplicate samples ± SD.
Pullet Husbandry and Performance
A total of 300-day-old Hy-Line W-36 pullets were purchased from a commercial hatchery (Hy-Line USA, Elizabethtown, PA) and utilized in experiment 2. All live animals used in this experiment were approved by the Pennsylvania State University Animal Care and Use Committee (IACUC #201901135). On day of placement, all birds were weighed and randomly assigned to a cage. The pullet rearing cages (51 cm x 61 cm; Chore-Time) consisted of 2 tiers of cages, each containing 2 nipple drinkers, a feed trough, and wire floors. For the first 3 d, newspaper was placed over the wire mesh flooring and supplemental waterers were placed in each cage. For the first 6 wk, all birds resided in the top tier of cages. Here, 4 replicate cages per treatment contained 25 chicks. At 6 wk of age, cages were split to allow adequate space per pullet. The top cages then contained 13 birds and the bottom cages contained 12 birds, doubling the number of replicate cages to 8 replicate cages per treatment. Treatments were arranged in a randomized complete block design with one cage of birds serving as the experimental unit. Birds were given feed and water ad libitum throughout the entire experiment. Management and lighting practices followed the 2019 Hy-Line W-36 management guide. At the end of each feeding phase, remaining feed and cages of birds were weighed to measure feed intake (FI), body weight gain (BWG), and mortality corrected FCR.
Statistical Analysis
A one-way ANOVA was used to analyze hammermill motor load and hammermill power usage in experiment 1. A single processing run for 4 min and 45 s served as the experimental unit. For experiment 2, pullet cages were arranged in a randomized complete block design with the experimental unit being a single cage of birds. A one-way ANOVA was used to analyze performance data that included BW, BWG, FI, and FCR. All data collected from both experiments were analyzed using the GLM procedure of SAS version 9.4. A post hoc Fisher’s least significant difference test was used to separate means when P ≤ 0.05.
RESULTS AND DISCUSSION
Experiment 1
Soybean Meal Cake Processing
Results from experiment 1 are shown in Table 3. Processing SBM cake with the 2.4 mm hammermill screen drew the most amperage (P < 0.001) and consumed the most power (kW) per tonne (P < 0.001). The 5.6 mm and 7.9 mm screens required less amperage to grind the SBM meal cake. Power usage was lowest for the 5.6 mm screen whereas the 7.9 mm screen was intermediate. However, the feed screw auger speed was set to 20 RPM, affecting the hammermill motor amperage when processing SBM cake with larger screens. If the feed screw auger speed would have been adjusted to optimize hammermill motor amperage, these power usage values would likely have changed. Findings in the current experiment are similar to those reported by Reece et al. (1986) who reported that approximately 27% more energy was required to grind corn when reducing the hammermill screen size from 6.4 mm to 4.8 mm. When Wondra et al. (1993) increased the hammermill screen size from 1.6 mm to 9.5 mm to grind corn, energy consumption was reduced by 7.5 kWh/ton. Fang et al. (1997) also reported that using a 1.6 mm hammermill screen required 9.54 kJ/kg more energy to grind wheat compared to a 4.8 mm screen (P < 0.05). Similarly, Kitto (2017) reported that increasing the hammermill screen size to grind corn reduced hammermill energy use and ultimately reduced corn grinding cost. In the current experiment, increasing the hammermill screen from 2.4 mm to 5.6 mm and 7.9 mm reduced power usage by 2.39 kW/tonne and 1.67 kW/tonne, respectively.
Table 3. Hammermill motor performance when fitted with various screen sizes.
| Screen size | Motor load(Amps) | Power usage(kW/tonne) |
|---|---|---|
| 2.4 mm | 26.2a | 5.95a |
| 5.6 mm | 16.0b | 3.56c |
| 7.9 mm | 16.0b | 4.28b |
| P-Value | <0.001 | <0.001 |
| LSD | 1.6 | 0.14 |
| SEM | 0.4 | 0.04 |
- a-c
-
Means within a column with different superscripts differ (P < 0.05).
Particle Size Analysis
Results from SBM PS analysis show that the 2.4 mm hammermill screen generated the smallest PS (464 µm) whereas the 7.9 mm screen generated the largest PS (892 µm; P = 0.037). The 5.6 mm screen generated intermediate SBM PS (735 µm; Table 4). These data show how hammermill screen selection impacts ingredient PS and impacts the distribution of particles across the sieves (Figure 1). Recently, Lyu et al. (2021) explained how it is possible to grind ingredients into a specified size class to realize a higher digestibility of nutrients in feed (2021).
Table 4. Effects of hammermill screen selection on soybean meal particle size.1
| Screen size | Particle size2(µm) |
|---|---|
| 2.4 mm | 463.88 ± 2.23b |
| 5.6 mm | 735.41 ± 2.46ab |
| 7.9 mm | 891.85 ± 2.57a |
| P-Value | 0.037 |
| LSD | 292.64 |
| SEM | 74.53 |
- a-b
-
Means within a column with different superscripts differ (P < 0.05).
- 1
-
Particle size was determined using a Ro-Tap tester, Model RX-29 (WS Tyler company, Mentor, OH).
- 2
-
Means presented are the average particle size of nine samples ± the average standard deviation.
Figure 1. Average soybean meal particle distribution across nine replicate soybean meal particle size analyses.
Experiment 2
Diet Particle Size Analysis
Descriptive PS of the dietary treatments fed in experiment 2 are located in Table 2. Regardless of feeding phase, SBM produced from the 2.4 mm hammermill screen generated a diet with the smallest PS and SBM produced from the 7.9 mm screen generated a diet with the largest PS. All ground corn used to manufacture diets was purchased on the same day. A corn PS of ∼650 µm was selected based on suggestions from Kitto (2017) who reported improvements in early pullet BW and BWG compared to 900 µm and 1,500 µm corn.
Pullet Performance
Initial BW of pullet chicks did not differ between treatments (P = 0.163; data not shown). Overall, from 0 to 6 wk and from 7 to 17 weeks, hammermill screen size and resulting SBM PS did not influence FI, BWG, FCR, or BW (P > 0.05; Table 5). Mortality was 2.3% across the flock from 0 to 17 wk of age, and was not affected by SBM PS. In agreement with these results, Deaton et al. (1989) showed that when hens were fed diets varying in corn PS, performance parameters did not differ. Current study results from each of the 5 feeding phases are located in Table 6. Measured performance parameters were not affected by SBM PS, with the exception of Starter 2 period FI. Here, pullets fed the diet containing SBM manufactured with the 5.6 mm screen consumed less feed than pullets fed diets manufactured with SBM processed with 2.4 mm or 7.9 mm screens (P = 0.039). However, this Starter 2 period FI difference did not affect pullet performance in subsequent feeding phases. The BW of pullets at 17 wk did not differ between treatments (P = 0.415; Table 5) and aligned with the 2019 Hy-Line W-36 performance objectives. Nevertheless, in a study conducted by Bozkurt et al. (2019), Lohmann LSL pullets that were provided a diet containing coarsely ground cereal grains experienced a 0.080 improvement in 0 to 16-wk FCR compared to those fed a diet containing finely ground cereal grains (P < 0.05). Although FCR was improved with feeding coarse cereal, FI did not differ among birds fed diets containing fine or coarse cereals. These findings are similar to the results of the current study, in which no differences in overall FI were observed in the development of ready-to-lay pullets. Findings from past research and the current study indicate that cereal grain PS may have a greater influence on pullet performance than SBM PS.
Table 5. Effects of hammermill screen selection and soybean meal particle size on 0–6 wk and 7–17 wk Hy-Line W-36 pullet performance.
| Screen size | 0–6 wk of age1 | 7–17 wk of age2 | ||||||
|---|---|---|---|---|---|---|---|---|
| Empty Cell | FI(kg/bd) | BWG(kg/bd) | FCR3(kg:kg) | 6-wk BW(kg/bd) | FI(kg/bd) | BWG(kg/bd) | FCR3(kg:kg) | 17-wk BW(kg/bd) |
| 2.4 mm | 1.025 | 0.381 | 2.688 | 0.417 | 4.460 | 0.863 | 5.224 | 1.269 |
| 5.6 mm | 0.998 | 0.381 | 2.624 | 0.419 | 4.425 | 0.844 | 5.280 | 1.263 |
| 7.9 mm | 1.023 | 0.380 | 2.692 | 0.415 | 4.450 | 0.854 | 5.373 | 1.255 |
| P-Value | 0.149 | 0.950 | 0.183 | 0.852 | 0.790 | 0.366 | 0.280 | 0.415 |
| LSD | 0.032 | 0.014 | 0.087 | 0.017 | 0.112 | 0.027 | 0.193 | 0.023 |
| SEM | 0.009 | 0.004 | 0.025 | 0.005 | 0.037 | 0.009 | 0.064 | 0.007 |
- 1
-
Four replicate cages of pullet chicks per treatment.
- 2
-
Eight replicate cages of pullets per treatment.
- 3
-
Mortality corrected FCR: mcFCR = FI/(BWG + Mortality Wt).
Table 6. Effects of hammermill screen selection and soybean meal particle size on performance of Hy-Line W-36 pullets across five feeding phases.
| Empty Cell | Starter 1 (0–3 wk) | Starter 2 (3–6 wk) | Grower (6–12 wk) | Developer (12–15 wk) | Prelay (15–17 wk) | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Screen size | FI(kg/bird) | BWG(kg/bird) | FCR1(kg:kg) | FI(kg/bird) | BWG(kg/bird) | FCR1(kg:kg) | FI(kg/bird) | BWG(kg/bird) | FCR1(kg:kg) | FI(kg/bird) | BWG(kg/bird) | FCR1(kg:kg) | FI(kg/bird) | BWG(kg/bird) | FCR1(kg:kg) |
| 2.4 mm | 0.356 | 0.151 | 2.369 | 0.664a | 0.229 | 2.901 | 2.244 | 0.536 | 4.198 | 1.388 | 0.225 | 6.154 | 0.828 | 0.093 | 8.957 |
| 5.6 mm | 0.349 | 0.151 | 2.312 | 0.650b | 0.230 | 2.829 | 2.228 | 0.529 | 4.267 | 1.367 | 0.225 | 6.103 | 0.830 | 0.090 | 9.130 |
| 7.9 mm | 0.352 | 0.151 | 2.324 | 0.667a | 0.227 | 2.940 | 2.253 | 0.525 | 4.285 | 1.366 | 0.209 | 6.712 | 0.831 | 0.096 | 8.751 |
| P-Value | 0.834 | 0.981 | 0.745 | 0.039 | 0.814 | 0.136 | 0.670 | 0.225 | 0.388 | 0.317 | 0.088 | 0.250 | 0.969 | 0.432 | 0.623 |
| LSD | 0.029 | 0.007 | 0.186 | 0.013 | 0.010 | 0.116 | 0.061 | 0.014 | 0.138 | 0.034 | 0.017 | 0.826 | 0.032 | 0.010 | 0.822 |
| SEM | 0.008 | 0.002 | 0.054 | 0.004 | 0.003 | 0.033 | 0.020 | 0.004 | 0.045 | 0.011 | 0.006 | 0.272 | 0.012 | 0.003 | 0.271 |
- a-b
-
Means within a column with different superscripts differ (P < 0.05).
- 1
-
Mortality corrected FCR: mcFCR = FI/(BWG + Mortality Wt).
CONCLUSIONS AND APPLICATIONS
- •
-
Increasing hammermill screen size reduced hammermill motor amperage when processing soybean meal cake.
- •
-
Power usage was highest when the 2.4 mm screen was used to process soybean meal cake.
- •
-
Soybean meal particle size did not impact the development of ready-to-lay pullets.
- •
-
Energy cost savings from proper hammermill screen selection may reduce feed costs to produce ready-to-lay pullets.
Disclosures
The authors declare no conflict of interest.
Acknowledgments
The authors acknowledge The Pennsylvania Soybean Board for their financial support in funding this project.
REFERENCES
- Amerah et al., 2008
-
A.M. Amerah, V. Ravindran, R.G. Lentle, D.G. ThomasInfluence of feed particle size on the performance, energy utilization, digestive tract development, and digesta parameters of broiler starters fed wheat- and corn-based dietsPoult. Sci., 87 (2008), pp. 2320-2328
Kitto, 2017Kitto, L. D. 2017. Effect of corn particle size milling on broiler, pullet, and layer growth, performance, and digestibility. MS Thesis. Penn State Univ. State College.
United Soybean Board 2019United Soybean Board. 2019. Market view database. Accessed June 2021. https://marketviewdb.centrec.com.
