Agricultural drainage is a primary source of excessive nitrogen (N) in surface waters leading to significant water quality problems in streams and estuaries in many locations around the world.  Although there have only been a few field studies of the effect of drain depth and spacing on N loss to surface waters, published field data from Indiana and North Carolina indicate that the magnitude of NO₃-N loss in drainage waters is strongly dependent on drainage intensity (DI).  DI is defined herein as the drainage rate predicted by the Hooghoudt equation when the midpoint water table is coincident with the surface.  Trends for NO₃-N loss as a function of DI were similar for soils in the two states in spite of large differences in weather and soil conditions.  Simulation model studies agree with these trends for other soils ad locations.  These results would indicate that N losses to surface waters could be significantly reduced by designing drainage systems that minimize subsurface drainage intensity as much as possible. One way of reducing drainage intensity is to manage or control the drainage outlet so as to restrict drainage rates during periods when subsurface drainage is not critical.  Field studies under a wide range of soils and climates have shown that N losses can be reduced by 40 to 60 percent by controlled drainage.  Reduced N loss is generally attributed to increased denitrification and/or reduced net mineralization caused by shallower water tables and increased soil water contents resulting from decreased DI.  However these mechanisms have rarely been measured in field studies of drainage and controlled drainage.  This paper discusses the response of N losses to drainage design and management and the need for research to quantify the mechanisms governing those losses.