Models of nutrient dynamics in streams have been influenced by two approaches, terrestrial ecosystem modeling and engineering transport models.  The use of terrestrial point models was problematic because attempts to calculate efficiencies were reach length dependent.  While engineering models allowed for transport, they did not fully incorporate dynamic interactions between benthic and water column components.  The calculation of spiralling metrics has provided a bridge between these approaches.  Our current spiraling models have improved, but they are still inadequate is several ways.  Using a synthesis of nitrogen data from LINX1, LINX2, NPARS, and other studies at Coweeta, I will illustrate some of our problems in modeling nitrogen dynamics in streams.