Quantitative estimates of denitrification are needed in designing artificial wetlands to optimize nitrate (NO⁻₃) removal. Acetylene blockage and ¹⁵N‐tracer methods were employed to quantify denitrification in constructed wetlands receiving agricultural tile drainage, using plastic tubes to enclose in situ mesocosms. Estimates were also made through NO⁻₃ disappearance from mesocosm water columns. The ¹⁵N and C₂H₂ methods yielded comparable rates. At 4 to 25°C, and with 9 to 20 mg NO⁻₃‐N L⁻¹ initially in the mesocosm water columns, denitrification rates by the C₂H₂ technique ranged from 2.0 to 11.8 mg N m⁻² h⁻¹. In the June–August ¹⁵N experiment, when wetland NO⁻₃ was below detection, a time series of denitrification rates followed a bell‐shaped curve after a pulse input of NO⁻₃ (∼15 mg N L⁻¹, 70 atom% ¹⁵N). The maximal denitrification rate (9.3 mg N m⁻¹ h⁻¹) was observed 5.4 d after the pulse. After 33 d, 58% of the ¹⁵NO⁻₃ had been evolved as N₂, only ∼0.1% as N₂O; 6 to 10% was recovered in plant shoots and as organic N in the upper 5 cm of sediment. From 32 to 36% of the ¹⁵NO⁻₃ spike was not recovered, and presumably seeped into the sediments. The NO₃ disappearance rates in the water column ranged from 12 to 63 mg N m⁻² h⁻¹ at 11 to 27°C. Because water infiltration carries NO⁻₃ through the anaerobic sediment/water interface for denitrification, a subsurface‐flow wetland may denitrify more NO⁻₃ than a surface‐flow wetland.