In native campos of Uruguay, overseeding legumes coupled with phosphorus (P) fertilization is a technology used to increase animal production. Short-term improvements in both forage productivity and quality are repeatedly reported. However, some evidence suggests that this management may at times lead to the collapse of the native community and invasions by exotic species. Indeed, it is yet unclear to what extent overseeding legumes into native grasslands affects its long-term integrity. This study uses data from a long-term experiment to assess whether increased P fertilizer rates—typically used to encourage legume establishment and growth—are associated with reduced species diversity. In 1996 a grazed native grassland in eastern Uruguay was either left untouched (control) or overseeded with a mix of Trifolium repens and Lotus corniculatus and then fertilized at either a moderate or high rate of P (197 or 394 kg · ha^-1 over 13 years, respectively). The three treatments were arranged in a randomized block design with four replicates of 2 hectares each. In 2005 the experiment was exhaustively sampled: 11 georeferenced sampling points per replicate, each encompassing∼20 m². Extractable P was measured in the 0—5- and 5—15-cm soil layers. In 2009, species presence and cover were measured at the same points. Across treatments, wherever legumes were introduced, extractable soil P was negatively related to species richness and diversity (P < 0.01) and native grass cover was reduced. This effect became asymptotic once soil P exceeded 27 and 36 mg · kg^-1 of P (0–5 cm), respectively. Therefore the documented reduction in species richness and diversity suggests a trade-off between increased pasture production and decreased vegetation stability may be operating in response to P fertilization of overseeded grasslands. The underlying ecophysiological mechanisms, as well as grazing management options to mitigate species diversity decline, should be further studied.