Previous studies have observed higher levels of soluble nutrients leaving vegetative buffers than entering them, suggesting that the buffers themselves are acting as a source rather than a sink by releasing previously stored nutrients. This study used 98 atom % (15)N-labeled KNO(3) at a rate of 5 kg ha(-1) to quantify buffer efficiency for sequestering new inputs of NO(-)(3)-N in an extensively grazed irrigated pasture system. Buffer treatments consisted of an 8-m buffer, a 16-m buffer, and a nonbuffered control. Regardless of the form of runoff N (NO(-)(3), NH(+)(4), or dissolved organic nitrogen [DON]), more (15)N was lost from the nonbuffered treatments than from the buffered treatments. The majority of the N attenuation was by vegetative uptake. Over the course of the study, the 8-m buffer decreased NO(-)(3)-(15)N load by 28% and the 16-m buffer decreased load by 42%. For NH(+)(4)-(15)N, the decrease was 34 and 48%, and for DON-(15)N, the decrease was 21 and 9%. Although the buffers were effective overall, the majority of the buffer impact occurred in the first four weeks after (15)N application, with the buffered plots attenuating nearly twice as much (15)N as the nonbuffered plots. For the remainder of the study, buffer effect was not as marked; there was a steady release of (15)N, particularly NO(-)(3)- and DON-(15)N, from the buffers into the runoff. This suggests that for buffers to be sustainable for N sequestration there is a need to manage buffer vegetation to maximize N demand and retention.