Excess nutrient loading from numerous sources (e.g., agricultural and urban runoff, treatment plant discharge, and streambank erosion) continues to adversely impact water resources, and determination of the cause(s) of accelerated nutrient enrichment has become a contentious and litigious issue in several US regions. This paper addresses one fundamental question: What are acceptable levels of nutrients in runoff from agricultural fields? It focuses on the field scale where farmers and ranchers make management decisions. Not answering this question limits the effectiveness of on-farm management and policy alternatives to address agriculture's contribution. To answer the question, some might suggest "direct comparison" with reference site data, existing criteria/standards, or measured data compilations. Alternatively, "indirect assessments" using soil test phosphorus (P) levels, P indices, field-scale models, or certainty programs might be suggested. Thus, to provide a scientific basis for policy debate and management decisions related to nutrient runoff from agricultural fields, we evaluated "direct comparisons" with measured data from case studies and evaluated "indirect assessment" alternatives. While acknowledging that scientific challenges and practical realities exist for each alternative, we concluded that certainty programs offer the most promise for ensuring acceptable nutrient runoff, and that field-scale models linked with watershed decision support tools are the most promising for assessing impacts on downstream water quality. Recognizing the reality that some nutrient loss is unavoidable from natural and anthropogenic sources, agriculture, industry, and municipalities are each encouraged to commit to implementing enhanced management where needed to minimize their sector's contribution to excess nutrients in our nation's waters.