Calculations of flows and balances of plant nutrients in agricultural production systems provide some basic information for the assessment of their long-term sustainability. The objectives of this study were to assess the possible impacts of variations in element concentrations between years and of undefined sinks and sources of elements on the accuracy of balance calculations. A 3-year study was conducted on Ca, Cu, K, Mg, Mn, N, P, S, and Zn fluxes in the barns (subsystem) of a Swedish farm with separate conventional and organic milk production. Our main focus in this subproject was on barn balance calculations, the barn housing only cows. Barn balance for an element was defined as amount of that element in [feeds, heifers, bedding, water] - [milk, manure, urine, calves, culled cows]. The focus was on: (1) variations in element concentrations in the main flow carriers [feeds, milk, manure, urine]; (2) information about element dynamics and flows of dairy farming systems obtained from internal flows of elements in the barn balance compared with that obtained from the flows associated with milk production in a farm gate balance; (3) differences in element flows and concentrations between the organic and conventional farming systems on this farm. Our conclusions were: (1) the sampling methods used had low coefficients of variation and thus pooled samples can reduce the costs of element analyses. However, urine must be thoroughly mixed if less water-soluble elements are to be monitored. Magnesium differed significantly in concentrations between years in all feedstuffs; (2) year-to-year fluctuations in harvest can influence a calculation negatively if calculations are based on annual harvest and not on feed supplied. The barn balance calculation showed a source of Cu, Mn and Zn that would not have been obvious in a farm gate balance. The element content of manure and urine calculated as [inputs - milk] would have underestimated the amount of Cu, Mn and Zn in manure and overestimated the amount of K and N. The Cu analysis showed an example of conflicting goals between short-term welfare of the cows and long-term soil fertility. EU legislation regarding land for spreading of manure is not a guarantee against soil contamination by heavy metals; (3) the differences between the organic and conventional system related more to differences in forage: concentrate and home-grown: purchased ratios, which were typical for the average Swedish farm of each type, and less to differences in element concentrations of the feed ingredients.