Nitrogen stable isotope (15N) natural abundance is widely used to study nitrogen cycling. In grazed ecosystems, urine patches are hot-spots of nitrogen inputs, losses, and changes in ?15N. Understanding ?15N dynamics in urine-affected vegetation is therefore crucial for accurate inferences from 15N natural abundance in grasslands. We hypothesized that leaf ?15N following urine deposition varies with time and plant functional group. Specifically, we expected (i) short-term decreases in ?15N due to foliar absorption of 15N-depleted volatilized ammonia, (ii) followed by increases in ?15N due to uptake of 15N-enriched soil inorganic nitrogen, and (iii) that the magnitude of these changes is less in legumes than in grasses. The latter should be expected because ammonia absorption depends on leaf nitrogen concentration, which is higher in legumes than grasses, and because biological nitrogen fixation will modify the influence of urine-derived nitrogen on ?15N in legumes. We applied cattle urine to a mixture of Lolium perenne and Trifolium repens in a pot experiment. Nitrogen concentration and ?15N were determined for successive leaf cohorts and bulk biomass either 17 (early) or 32 (late) days after urine application. Early after urine application, leaves of L. perenne were 15N-depleted compared to control plants (?15N 0.1 vs. 5.8?, respectively), but leaves of T. repens were not (-1.1 vs. -1.1?, respectively). Later, both species increased their ?15N, but T. repens (4.5?) less so than L. perenne (5.9?). Vegetation sampled within and outside urine patches in the field further supported these results. Our findings confirm that foliar ammonia uptake can substantially decrease grass foliar ?15N, and that in both grass and legume the direction of the ?15N response to urine changes over time. Temporal dynamics of plant ?15N at urine patches therefore need to be explicitly addressed when 15N natural abundance is used to study nitrogen cycling in grazed grasslands.