Ammonia (NH3) volatilization is an undesirable mechanism for the removal of nitrogen (N) from wastewater treatment wetlands. To minimize the potential for NH3 volatilization, it is important to determine how wetland design affects NH3 volatilization. The objective of this research was to determine how the presence of a pond section affects NH3 volatilization from constructed wetlands treating wastewater from a confined swine operation. Wastewater was added at different N loads to six constructed wetlands of the marsh-pond-marsh design that were located in Greensboro, North Carolina, USA. A large enclosure was used to measure NH3 volatilization from the marsh and pond sections of each wetland in July and August of 2001. Ammonia volatilized from marsh and pond sections at rates ranging from 5 to 102 mg NH3-N m(-2) h(-1). Pond sections exhibited a significantly greater increase in the rate of NH3 volatilization (p < 0.0001) than did either marsh section as N load increased. At N loads greater than 15 kg ha(-1) d(-1), NH3 volatilization accounted for 23 to 36% of the N load. Furthermore, NH3 volatilization was the dominant (54-79%) N removal mechanism at N loads greater than 15 kg ha(-1) d(-1). Without the pond sections, NH3 volatilization would have been a minor contributor (less than 12%) to the N balance of these wetlands. To minimize NH3 volatilization, continuous marsh systems should be preferred over marsh-pond-marsh systems for the treatment of wastewater from confined animal operations.