ABSTRACT Ferruginous Rocky Outcrops have high levels of species richness and endemism, but have been threatened by several anthropic actions, especially mining. Eragrostis polytricha, a common grass species in the vegetation of these outcrops, has shown promising features for use in the recovery of mining areas. However, in order to fully understand the species’ potential for such use, its requirements for germination, seed dormancy break and seedling development must be determined. Thus, we aimed to: (1) assess the temperature conditions needed for seeds of E. polytricha to germinate; (2) evaluate the effects of KNO₃ in breaking seed dormancy; and (3) analyze the germination efficiency of seeds that are still in spikelets. The experiment included seven treatments: 15-35 ºC with KNO₃, 20-30 ºC with KNO₃, 15 ºC with KNO₃, 25 ºC with KNO₃, 35 ºC with KNO₃, 20-30 ºC with water, and 20-30 ºC with KNO₃ using spikelets. The treatments with alternating temperatures associated with KNO3 yielded the highest germination rates, suggesting that these two factors combined can break seed dormancy. Seeds inside spikelets exhibited a high germination percentage, and thus represent an interesting alternative for seedling production.

Objetivou-se, em Lavras, MG, caracterizar a estrutura e a magnitude de dependência espacial de variáveis ambientais nas nascentes da bacia hidrográfica do Ribeirão Santa Cruz, por métodos de interpolação geoestatística. Os métodos de interpolação por krigagem ordinária e inverso da potência da distância (IPD) foram utilizados para representar altitude, vazão, áreas de recarga e de vegetação de 175 nascentes perenes em 86,99 km² de área. Observou-se melhor desempenho da estimação por krigagem quando comparado ao método IPD.

The objective of this study was to characterize the structure and the spatial dependence magnitude of geobiophysical variables in the springs of the Ribeirão Santa Cruz stream watershed, Lavras, MG, using geostatistical interpolation methods. Ordinary kriging system and inverse power distance (IPD) interpolation methods were used to map the altitude, flow, areas of recharge and vegetation of 175 perennial springs in an area of 86.99 km². Better peformance was observed with estimation using the kriging method as compared to the IPD method.