ASAE Conference Proceeding
This is not a peer-reviewed article.
Land Application of Beef Feedyard Effluent to Forage Sorghum and Winter Wheat
M. B. Rhoades, D. B. Parker, J. M. Sweeten, N. A. Cole, and M. S. Brown
Pp. 099-106 in the Animal, Agricultural and Food Processing Wastes, Proceedings of the Ninth International Symposium, 11-14 October 2003 (Raleigh, North Carolina, USA), ed. Robert Burns. ,11 October 2003 . ASAE Pub #701P1203
CAFO's feed approximately seven million head of cattle annually in the Texas High Plains. Regulations require that all runoff from precipitation be contained and disposed to land. Feedlot runoff was applied to 27 plots of winter wheat and forage sorghum over 24 months at the USDA Conservation and Production Research Laboratory located at Bushland, Texas. Winter wheat and forage sorghum were planted and grown for hay. Runoff was applied by flood irrigation onto level borders. Rates applied were as follows: 0 cm/cropping season (TRT 1, control), 25 cm (TRT 2) and 50 cm (TRT 3). Cropping rotations of sorghum-fallow (SF), wheat-fallow (WF) and sorghum-wheat (SW) (two crops/year) were used for each TRT. Plots were irrigated every two weeks after plant emergence until the appropriate amount of effluent was obtained. Above ground biomass samples were collected and allowed to air dry for three weeks, after which they were ground mixed and sent to the laboratory for analysis. Soil samples were collected before planting and after harvest and analyzed at the laboratory. Effluent samples were collected three times during each irrigation, composited, and analyzed. TRT 3 SW had the greatest total yield (28, 564 kg/ha) followed by WF (19,153 kg/ha) and SF (14,337 kg/ha). Residual soil nutrient concentrations were more dependent on initial nutrient concentrations than on effluent application or cropping rotations. Residual soil N was 70% dependent on initial N, while residual soil P was 56% dependent on initial P.KEYWORDS. land application, beef cattle, feedyard, effluent, wheat, sorghum, nutrients
Approximately seven million cattle are fed annually in confined animal feeding operations (feedyards) in the Texas High Plains (SPS, 2000). The average time of confinement for each animal is approximately 140-160 days. At any one time, approximately 3.5 million cattle are housed in feedyards. Each animal is given 17 to 21 m 2 of pen space (Rhoades, 1990), for a total pen area of 5900 to 7200 ha.
Beef cattle feedyards in the Texas High Plains are constructed as open lots with earthen surfaces. Any precipitation that falls on these surfaces is either absorbed in the manure or runs off (Miller et al., 2000). The semi-arid region receives approximately 480 mm of rainfall/year (USDA-SCS, 1993), whereas runoff amounts have been calculated to be 20 to 50% of annual precipitation rates (Clark et al., 1981; Phillips, 1991; Swanson et al., 1974a; Swanson et al 1974b; Sweeten et al., 1994).
The most popular and least expensive means of dewatering retention ponds is through evaporation. The average evaporation rate for this region is 1,830 mm per year (MWPS,1992). Land application of the effluent has also been a historically popular means of dewatering the retention ponds. However, recent environmental concerns and new regulations are limiting the amount of feedlot runoff that can be applied to soil for crops. For a crop producer, the soil nutrients of major concern are the macronutrients, nitrogen (N), phosphorus (P), and potassium (K). After water, N is generally the most limited factor in crop production followed by P. For the Texas High Plains, K is not considered a limiting factor.
Nutrient concentrations in feedlot runoff vary widely (Lorimor et al., 1974; Clark et al., 1975). Table 1 gives the average nutrient concentrations of holding ponds in the Texas High Plains. Sprinkler irrigation is generally the preferred approach to land application of feedlot runoff (Sweeten 1991, Sweeten 1998). Application rates can be controlled to within 1.2 cm/hour with sprinklers. Level borders are also a very acceptable method for land application of feedyard effluent (Sweeten, 1991; Sweeten, 1998).
Table 1: Average concentration of nutrients, salts, and other water quality variables from stored beef cattle feedlot runoff in the Texas High Plains 1,2
Sodium Absorption Ratio (SAR)
Electrical Conductivity (mmhos/cm)
1 To convert to lbs/acre-inch, multiply ppm by 0.226
2 Composition of runoff after it had been in the runoff holding ponds for several weeks. Playas typically catch runoff from areas other than the feedlot, thus, there is a greater dilution.
2 From Clark et al. (1975)
Several researchers (Sweeten, 1991; Sweeten, 1998; Wallingford et al, 1974) have indicated the need for adequate soil testing before effluent application, and Texas Natural Resource Conservation Commission (TNRCC) regulations state that soil testing should be done. Tests should be done for N, P, K, Na, Ca and Mg (Wallingford et al, 1974). Runoff applied to the soil should also be tested for the same nutrients to avoid over-application (Lorimor et al, 1974).
Large amounts of animal waste application can decrease crop yields (Wallingford et al., 1974; Mathers et al., 1974; Murphy et al., 1972), and increased soil salinity was thought to be responsible for the decreased yields. Aldrich et al. (1997) applied poultry runoff on bermudagrass at Overton, TX and found that yield increased with effluent application. Horton et al. (1974) found that yields in corn varied with animal waste application. In a greenhouse study, Miller et al. (2000) found that effluent applications containing up to 100% of crop N requirements increased forage sorghum yield, whereas applications containing 200% of the crop N requirements decreased yield.
The objectives of this study were to 1) determine optimum land application rates for feedlot runoff onto cropland and 2) determine if crop rotation affected nutrient removal after effluent application.
Materials and Methods
Beef feedlot effluent was applied in a field study to two crops and three cropping rotations over a two-year period. The crops were forage sorghum and wheat, and the cropping rotations were sorghum-fallow, sorghum-wheat, and fallow-wheat.
The experiment was located at the USDA-ARS/Conservation and Production Research Laboratory at Bushland, Texas, 12 miles west of Amarillo, Texas. The experimental plots were located east of the feedlot retention pond. Preliminary soil samples were collected before plot layout to determine available soil nutrients, and plots were designed and constructed to obtain the most homogenous soils possible. The soil type where the plots were located is comprised of Pullman B series clay loam (fine, mixed, thermic Torrertic Paleustolls)(USDA-SCS, 1974).
Beef cattle feedyard effluent was collected at the 384-head capacity USDA-ARS Research Feedyard in a runoff retention pond. A 20 cm diameter underground polyvinyl chloride pipeline was installed to transport the effluent to the research plots. An 850 L/min pump was placed on a cement pad at the retention pond. Gated 20 cm pipe was placed on the earthen borders surrounding the research plots so effluent application could be controlled. All plots were flood irrigated. Using an inline flow meter, application rates were measured to an accuracy of 0.38 m 3 .
The land was terraced to develop level borders so an even coverage of water would be obtained when the plots were flooded. The soil was tilled with an offset plow to a depth of 20 cm. A box blade with a laser-leveling device (laser-plane) was then used over the entire area to begin the leveling process. After some initial leveling and breaking up of the larger clumps of soil, a chisel-chopper plow was used over the area. After a preliminary leveling of the ground, a Graham-Hoeme chisel was run over the plots east to west to break up the plow pan. Plots were again staked and final leveling commenced. After the final leveling process, individual plots were staked and divided with a border disc plow. Individual plot sizes were 4.5 m x 16 m. The area was divided into three sections of nine plots each for twenty-seven total plots. Each section was separated by a 5.5-m wide access road. The research area for this study had never been cropped before. After leveling and plowing, it was determined that the first year would be used to equalize the soil nutrient concentrations in the plots.
Forage sorghum ( Sorghum bicolor ) was planted on July 10, 1999, and again on June 23, 2000 at the seeding rate of 11 kg/h in all plots in Sections I and II. After germination and plant emergence, feedlot runoff was applied to the appropriate plots, as determined by random selection. Irrigation rates were determined randomly, and each plot was left at that rate throughout the study regardless of the crop planted (Table 2).
Sorghum forage was harvested when an estimated 70% of the plants had reached early boot stage. Harvest of the first sorghum crop occurred on September 20, 1999, and the second sorghum crop was harvested on September 20, 2000.
Wheat ( Triticum aestivum , awnless variety ) seed was planted at the rate of 67 kg/ha in all plots in Sections II and III after sorghum harvest was completed. The first wheat crop was planted on November 2, 1999 and the second was planted on October 20, 2000. Effluent was applied after the crop had germinated and emerged.
Wheat hay harvest was done when an estimated 70% of the plants had reached early boot stage. The first year wheat crop was harvested on June 13, 2000, and the second on May 9, 2001.
Treatments were effluent application rates of: 1) 0 cm/crop, 2) 25 cm/crop, and 3) 50 cm/crop. No other source of irrigation was used. Precipitation events were measured with a rain/snow gauge located at the northeast corner of the experimental area. Precipitation for each cropping season was as follows: 1 st sorghum – 218 mm, 2 nd sorghum – 80 mm, 1 st wheat – 153 mm, and 2 nd wheat – 223 mm. Because of drought conditions and the unavailability of effluent, the 1 st wheat crop and the 2 nd sorghum crop received less effluent than was planned (Table 2). Average precipitation amounts during the wheat cropping system for this area are 256 mm while about 203 mm are expected during the sorghum cropping season. McGill et al. (1974) determined that feedyard runoff was not a dependable irrigation source, although it could be used to supplement other sources. In a typical feedyard setting, fresh water is reserved for drinking water for cattle and for use in the boiler if steam flaking of grain is used. This results in little water for use in the irrigation of crops. Therefore, if a crop is planted, it is considered a dryland crop.
Table 2 : Dry matter yields by year, cropping rotation, and effluent application rate. Only those values for year 2 were used for statistical analysis. Yields are averages of three plots.
Effluent Application Rate (cm)
Mean Dry Matter Yield (kg/ha) 1
Mean Dry Matter Yield (kg/ha) 1
Mean Dry Matter Yield (kg/ha) 1
Total Forage Produced
Total Forage Produced
Total Forage Produced
1 Yield is mean of three plots
a,b,c,d means with different letters are significantly different at a =0.10 using Tukey’s HSD test.
Crops were irrigated when the top 5-cm of the soil was dry to the touch assuming there was sufficient effluent in the holding pond. Those plots receiving 25 cm of runoff/growing season were irrigated twice and those receiving 50 cm were irrigated three times.
Soil samples were taken after all leveling and border placement had been completed and before planting of the first crop to determine baseline soil nutrient concentrations. Thereafter, soil samples were taken within 2 weeks after harvest of each subsequent crop. Sampling was done with a tractor mounted hydraulic Giddings Probe with a 10 cm diameter. Samples were taken not less than 2 m from the edge of the plots to discount any border effect.
Soil samples were taken at three locations in each plot. Subsamples were taken at depths of 0–15 cm, 15–46 cm, and 46–91 cm. The three subsamples at the same depth were composited for quality analysis. The subsamples at each sampling locations were averaged using a weighted average for all statistical analysis. Soils were analyzed for pH, EC, total N (Kjeldahl digest with a Technicon Auto Analyzer II), NO 3 – N (Lachat Instruments Quick-Chem 8000), NH 4 – N, and total P (Kjeldahl digest using ICP). Extractable P and K were measured by Inductively Coupled Plasma (ICP) on ammonium acetate-ethylenediamenetetraacetic acid (NH 4 OAc-EDTA) extract using the TAMU Method (Hons et al. 1990).
Above ground biomass samples were taken at harvest. Within each plot, the first sorghum samples were taken from two areas, each measuring 2 m x 2.5 m. Thereafter, it was determined that an area of 1 m x 1 m would suffice for sampling for all other crops. Samples were air dried for two weeks. Dry samples were weighed and ground, then sent to the Soil, Water, and Forage Testing Laboratory (SWFTL) at Texas A& M University for quality analysis. Biomass was analyzed for crude protein by Kjeldahl digest. P and K were analyzed using ICP.
Effluent samples were taken at the beginning, middle, and end of each irrigation event, composited and sent to the SWFTL for quality analysis. Samples were analyzed for N, P, K and NO 3 in the same manner as described above.
Statistical analysis on crop yields was done by one-way ANOVA. Means were separated using Tukey’s Highly Significant Difference. Statistical analysis for the soil nutrient concentration consisted of two-way factorial analysis of covariance (ANCOVA) and least significant difference comparisons (LSD) among treatments v. cropping rotations. The co-variate used for each nutrient concentration was the initial soil nutrient concentration. For all nutrient soil concentrations, the co-variate was left in the model and no interactions were found. Main effects for each model were: a) cropping rotation and b) effluent application rate. Polynomial contrast statements were conducted to test linear and quadratic effects of effluent application rate. All statistical analyses were performed using SPSS Version 9.0.
Results and Discussion
Total residual soil N concentrations did not differ ( a = 0.10) between the sorghum-fallow and sorghum-wheat rotations or between fallow-wheat and sorghum-wheat cropping rotations. However, statistical differences were found between the sorghum-fallow and fallow-wheat rotations. The sorghum-fallow rotation had lower N concentrations in the soil than the fallow-wheat. No significant differences were found among effluent application (TRT). Residual soil N concentrations were heavily dependent upon beginning soil concentrations. Initial N concentrations accounted for 70% of the variability in the residual soil concentrations. Variability was determined by analysis of co-variance (ANCOVA). This would indicate that soil testing for N would be highly appropriate before beginning any type of effluent application.
Residual soil NO 3 -N concentrations indicated differences among both cropping rotation and effluent application. The wheat-sorghum rotation resulted in less NO 3 -N ( a = 0.10) in the soil than either the fallow-wheat or the sorghum-fallow cropping rotations. No differences were found between the fallow-wheat and the sorghum-fallow rotations. Plots with effluent application rates of 25 cm/crop had a significantly greater amount of NO 3 -N remaining in the soil profile than did plots with rates of 50 cm/crop, due to either greater crop uptake in TRT 3 or to leaching from the greater water application. However, application rates of 0 cm/crop was not significantly different than rates of 25 or 50 cm/crop. Initial soil NO 3 -N accounted for 37% of the variability in the residual NO 3 -N concentrations. Although this is not as high as for total N, NO 3 -N soil concentrations still needs to be accounted for. NO 3 -N concentrations in the effluent were all less than 0.5 mg/kg. NO 3 -N can also lead to nitrate poisoning in beef cattle, so care must be taken when land applying large amounts (Marriott et al, 1974; Smith et al, 1992). Manges et al (1974) also found that increasing amounts of feedlot waste loading increased NO 3 -N amounts in the soil.
A soil nutrient of major interest at the current time is P, especially the labile form. Current regulations state that if soil concentrations exceed 200 ppm of extractable P, the landowner must establish a nutrient management plan. Runoff in feedyard holding ponds is typically high in extractable P concentrations (Clark et al., 1975). Consequently, particular attention has been focused on P content of effluent. Phosphorus concentrations in the effluent applied in the present study ranged from 1 to 10 mg/kg. Both the sorghum-fallow and wheat-sorghum tended to remove more P from the soil than the fallow-wheat rotation, although the differences were not significant. The sorghum-fallow and wheat-sorghum rotation did not differ from each other. Effluent application rates did not significantly effect P concentrations in the soil. Residual soil P concentrations were highly dependent upon beginning soil concentrations. Initial P accounted for 56% of the variability in the residual P concentrations.
Potassium, in most cases is not a limiting factor in crop production, but has been linked to soil salinity (Stewart and Meek, 1977) and is found in high amounts in feedyard runoff. Due to these factors, K is an important nutrient to monitor in land application of effluent. Neither cropping rotation nor effluent application rate influenced soil K concentrations. K concentrations in the effluent ranged from 110 to 320 mg/kg. This resulted in applications of up to 1,000 kgs of K/ha in some cases. This caused some slight increase in soil concentrations, but they were statistically not significant (p> 0.10). Potassium leaches in the soil easily so some over-application in effluent is not considered a problem for a short term. Long term application could result in some concerns, especially if adequate water for leaching is not available. Initial K concentrations accounted for 28% of the variability in the residual soil concentrations, which was considerably less than the effect of initial N and P. K should be analyzed for in a land application program due to the amounts that can be applied in the effluent.
The means and standard deviations (unadjusted by the covariate) showed that the nutrient concentrations in the soil were highly variable. This is important to producers because if the variation cannot be controlled in small-scale research, it will be much harder to control in a large-scale production setting. This will also prompt regulating agencies to make certain that adequate and representative soil sampling is done.
Dry matter yield in a land application study is important, as it determines the quantity of nutrients that will be removed from the soil. A more mature crop may have a lower N content. However, if yields are greater for a more mature crop, it is possible that more N would be removed from the soil. Yields can also be important from an economic standpoint. Most effluent must be used in cropland adjacent to a feedyard retention pond. If a hay crop can be produced, the feedyard may be able to offset some of the cost of the land application by producing forage.
The sorghum-wheat rotation produced more total forage across all effluent application rates for the two-year period than did either the sorghum-fallow or the fallow-wheat rotation. The fallow-wheat rotation produced more total forage across all effluent application rates than did the sorghum-fallow (Table 4).
Total soil N concentrations showed very little difference between initial and residual amounts. Although not statistically different, total N appeared to increase slightly in all cropping rotations and in most effluent application rates. Overall, even though trends showed increasing amounts of N, they were not significantly different for effluent applications and rotations. Residual soil N concentrations were very dependent on initial soil N concentrations.
Soil NO 3 -N results were variable and did not indicate any trends. Increasing concentrations of NO 3 -N were more than likely due to mineralization of organic N. The sorghum-fallow rotation, which has a long fallow period, tended to have a numerically greater soil NO 3 -N concentration than did the sorghum-wheat or fallow-wheat rotation. The sorghum-wheat rotation showed a lower concentration, which was expected due to the high plant availability of NO 3 -N.
Potassium is found in large amounts in feedyard runoff. K concentrations in the soil increased by as much as a 100 ppm.
The sorghum-wheat cropping rotation had the greatest total forage yield for the length of the study. The double cropping system plots also received twice the amounts of effluent. The sorghum-fallow rotation had the lowest overall dry matter yield. This cropping rotation also had the longest fallow period of the study. This could suggest that nutrients were lost from the bare soil surface due to volatilization or leaching from the root zone. It would appear that the sorghum-wheat rotation at the 50-cm/cropping season produced the highest total dry matter yields. However, only two years of data are represented and longer-term studies are needed.
The results of this study also show that nutrient variability in the soil requires very careful decision making on the part of both the feedlot operator and the regulating agency. If an unusual number is returned from the laboratory, it would probably be wise to resample at that particular site and have that sample re-tested. This also suggests that adequate and proper sampling techniques should be utilized in the field to ensure correct results. This will help to reduce the variability in the field. This study also shows that great variability can occur in a small area even with highly controlled parameters. This would indicate that variability is even more likely in a larger area where less control can be exerted. A land application program should be looked at as a long term program. Year to year nutrient levels should not be overlooked, but should be secondary to long term results.
Aldrich, L.J., C.L. Munster, V.A. Haby, and J.M. Sweeten. 1997. “Land Application of Poultry Lagoon Effluent”. Transactions of the ASAE. 40(6). Pp. 1607-1615.
Clark, R.N., A.D. Schneider and B.A. Stewart. 1975. “Analysis of Runoff from Southern Great Plains Feedlots”. Transactions of the ASAE, 15(2): 319-322.
Clark, R.N., C.B. Gilbertson, and H.R. Duke. 1981. “Quantity and Quality of Beef Feedyard Runoff in the Great Plains”. In: Managing Livestock Wastes , Proceedings of the Third International Symposium on Animal Wastes, American Society of Agricultural Engineers, St. Joseph, Michigan. Pp. 429-431.
Hons, F.M., L.A. Larson-Vollmer, and M.A. Locke. 1990. NH 4 OAc-EDTA Extractable Phosphorus as a Soil Test Procedure. Soil Science 149(5):249.
Horton, M.L., J.L. Halbeisen, J.L. Wiersma, A.C. Dittman, and R.M. Luther. 1974. “Land Disposal of Beef Wastes: Climate, Rates, Salinity, and Soil”. In: Managing Livestock Wastes, Proceedings of the Third International Symposium on Animal Wastes, American Society of Agricultural Engineers, St. Joseph, Michigan. Pp. 258-260.
Lorimor, J.C., S.W. Melvin, and B.M. Leu. 1974. “Nutrient Characteristics of Wastes from Deep Pits and Anaerobic Lagoons”. In: Managing Livestock Wastes , Proceedings of the Third International Symposium on Animal Wastes, American Society of Agricultural Engineers, St. Joseph, Michigan. Pp. 306-308.
MWPS Committee. 1993. Livestock Waste Facilities Handbook, 3 rd ed. Iowa State University. Ames, Iowa. PP 11-4 – 11-12.
Manges, H.L, R.I. Lipper, L.S. Murphy and W.L. Powers. 1974. “Disposal of Beef Feedlot Wastes onto Land’. In: Managing Livestock Wastes , Proceedings of the Third International Symposium on Animal Wastes, American Society of Agricultural Engineers, St. Joseph, Michigan. pp 261-268.
Marriot, L.F. and H.D. Bartlett. 1974. “Animal Waste Contribution to Nitrate Nitrogen in Soil”. In: Managing Livestock Wastes , Proceedings of the Third International Symposium on Animal Wastes, American Society of Agricultural Engineers, St. Joseph, Michigan. pp 296-298.
Mathers, A.C., B.A. Stewart, and J.D. Thomas. 1974. “Residual and Annual Rate Effects of Manure on Grain Sorghum Yields”. In: Managing Livestock Wastes , Proceedings of the Third International Symposium on Animal Wastes, American Society of Agricultural Engineers, St. Joseph, Michigan. Pp. 252-254.
McGill, H.N. and G.C. Vittetoe. 1974. “Management of Runoff Water in Relation to Feedlot Operations”. In: Managing Livestock Wastes . Pp. 88 – 92.
Miller, B., D.B. Parker, J.M. Sweeten, and C. Robinson. 2000. Response of Seven Crops and Two Soils to Application of Beef Cattle Feedyard Effluent.
Murphy, L.S., G.W. Wallingford, W.L. Powers, and H.L. Manges. 1972. Effects of Solid Beef Feedlot Wastes on Soil Conditions and Plant Growth. Proc. Cornell Agric Waste Manange Conf. p. 449-464.
Phillips, R.L. 1991. “Maps of Runoff Volumes from Feedlots in the United States”. In: Livestock Waste: A Renewable Resource . Pp. 274 - 277.
Rhoades, J.L. 1990. Feedyard Management Principles. In (Albin and Thompson eds.) Cattle Feeding: A Guide to Management. pp. 21-25.
S.P.S. 2000. Cattle-Feeding Capital of the World, 2000 Fed Cattle Survey
Smith, S.J., B.A. Stewart, A.N. Sharpley, J.W. Naney, T. McDonald, M.G. Hickey, and J.M. Sweeten. 1992. “Nitrate and Other Nutrients Associated with Playa Storage of Feedlot Wastes”. TAES B-5081.
Stewart, B.A. and B.D. Meek. 1977. “Soluble Salt Considerations with Waste Applications”. Chap 9. In: Soils for Management of Organic Wastes and Wastewaters (L.F. Elliot and F.S. Stevenson, eds.) Soil Science Society of America, Madison, Wisconsin. pp 210-232.
Swanson, N.P., C.L. Linderman, and L.N. Mielke. 1974a. “Direct Land Disposal of Feedlot Runoff”. In: Managing Livestock Wastes , Proceedings of the Third International Symposium on Animal Wastes, American Society of Agricultural Engineers, St. Joseph, Michigan. Pp. 255-257.
Swanson, N.P., L.N. Mielke, and C.L. Linderman. 1974b. “Control, Collection, and Disposal of Feedlot Runoff”. In: Managing Livestock Wastes , Proceedings of the Third International Symposium on Animal Wastes, American Society of Agricultural Engineers, St. Joseph, Michigan. Pp. 85-87.
Sweeten, J.M. 1991 “Feedlot Runoff Characteristics for Land Application”.. In: Agricultural and Food Processing Wastes , Proceedings of the Sixth International Symposium on Agricultural and Food Processing Wastes. Chicago, Illinois.
Sweeten, J.M. 1998. “Cattle Feedlot Manure and Wastewater Management Practices”. In: Animal Waste Utilization: Effective Use of Manure as a Soil Resource . Pp. 125 – 155.
Sweeten, J.M., G.L. Sokora, R.M. Seymour, M.G. Hickey, and S.M. Young. 1994. “Irrigation of Cattle Feedlot Runoff on Winter Wheat”. Great Plains Animal Waste Conference on Confined Animal Production and Water Quality.
USDA-SCS. 1992. Agricultural Waste Management Field Handbook .
Wallingford, G.W., W.L. Powers, and L.S. Murphy. 1974. “Present Knowledge on the Effects of Land Application of Animal Waste”. In: Managing Livestock Waste , Proceedings of the Third International Symposium on Animal Wastes, American Society of Agricultural Engineers, St. Joseph, Michigan. Pp. 580-586.