Improving management guidelines may result in more efficient use of liquid swine manure and in reduced environmental risks. Two different studies were conducted on farmers' fields using liquid swine manure on corn and soybean crops. One study used conventional research plot methodology with objectives to evaluate (a) manure P effects on early plant growth, early P plant uptake, grain yield, and P removal with grain; (b) crop response to fertilizer P applied in addition to manure P applied once for two crops; and (c) evaluate liquid swine manure P application on soil P measured with three routine soil-P tests (Bray-1 P, Olsen, and Mehlich 3 P), and two environmental tests (Iron-oxide and Water extraction). The second study used a strip-trial methodology using precision agriculture technologies and its objective was to compare three statistical methods (classical, geostatistic, and Markov random field model) to study soil-test P spatial variability after applying liquid swine manure with variable- (VR) and fixed- (FR) rate application methods. Early growth and P uptake were not directly related to manure or fertilizer P application. Grain yield responsive sites tested less <20 mg P kg-1, except for one site. Fertilizer P did not increase grain yield at any site when applied in addition to the high manure rate. Manure or fertilizer P increased soil test P (STP) in many sites, and soil tests did not differ consistently in detecting an STP increase due to manure and fertilizer P. Differences between both P sources in increasing STP were not evident when similar rates of P were compared. Routine and environmental P tests did not differ in assessing available P from liquid swine manure or fertilized soils. The results demonstrated that liquid swine manure is a valuable resource of P for crop production and provided no evidence for lower season-long availability of P in manure compared with P in fertilizer. The VR manure application method reduced STP variability, probably because more P was applied to low-testing field areas and less or no P was applied to high-testing areas. Conclusions were similar for the three statistical methods used.