Abstract A cutting experiment was carried out in Bahia Blanca (Argentina) during 1998 and 1999 using a randomized block design in a 2 3 factorial experiment: (1) the species Digitaria eriantha and Eragrostis curvula , (2) 0 and 60 kg Nha −1 fertilizer treatment and (3) deferred harvesting on 1 June and 15 July. The existing forage was cut in mid-February, the N was distributed and the deferred treatment was initiated. Dry matter yield (DM), dead (DF) and green (GF) fractions, crude protein (CP) content, nitrogen utilization efficiency (NUE), nitrogen recovery (NR) and in vitro DM digestibility (IVDMD) for each year were determined. A treatment by year interaction in DM yields ( p p >0·05). The DM yields were substantially increased by fertilization ( p −1 for D. eriantha and 846 and 1958 kg ha −1 for E. curvula , respectively. Similarly, CP for D. eriantha was 5·2 and 6·0%, and 4·2 and 5·2% for E. curvula , respectively. The IVDMD had a small response to the fertilization ( p D. eriantha was 51% and for E. curvula 39%. The quality of the forage was not influenced by the percentage relationship between DF and GF ( p >0·05). The NUE (kg DM kg N −1 ) and the NR (%) were 24·8 and 25·4 for D. eriantha and 18·6 and 17·6 for E. curvula , respectively. Fertilization must be implemented depending on the costs and the expected increase in income due to the increase in forage production. Marked leaf lodging of D. eriantha may restrict the grazing ability of the animals.