Tropical-alpine treelines are characterized by discontinuous forest patches forming abrupt boundaries with the adjacent grassland vegetation. This also holds true for the tropical-Andean forests formed by Polylepis spp. Although the peculiar vegetation pattern has frequently been attributed to anthropogenic disturbances, quantitative or experimental studies on the consequences of land use and environment for the current treeline position are limited. Knowledge on regeneration processes is crucial to the understanding of treeline formation and future stand development, however, recruitment patterns with associated environmental and human factors at Polylepis treelines still remain poorly understood. The studies which comprise this thesis deal with the impact of altitude, canopy cover, litter depth, grazing and burning on the reproduction of two Polylepis species in the Páramo de Papallacta, Ecuador. In particular, I analysed flowering, fruit set, seed viability, seedling emergence and survival as well as stand structure of Polylepis incana and P. pauta. The data on natural regeneration in situ were supplemented by sowing experiments at the upper treeline and after burning in order to discriminate effects of microsite conditions from those of seed availability. The results on tree regeneration clearly indicate different regeneration patterns at the upper treeline in comparison to forest boundaries at lower altitudes. The low recruitment at the highest distribution limit characterizes this treeline as being natural due to the harsh climatic conditions. A further upslope migration by means of sexual regeneration seems unlikely, unless climatic conditions change. Consequently, the climatic growth limitations at the upper forest border of P. incana and P. pauta, which affect growth of adult trees, saplings and ramet populations, equally hampers the production of flowers and seeds as well as seedling recruitment. In contrast, forest boundaries below this altitude did not show any limitations in generative or vegetative reproduction, and the forest edge actually proved to be a zone of high natural recruitment. I conclude that these forest borders have to be attributed to human influences. As there were no negative effects of cattle grazing or woodcutting in the study area, I suggest that fire is the main factor driving boundary formation at lower altitudes. According to these findings, the patchy discontinuous vegetation pattern found at tropical-alpine treelines in Ecuador is a consequence of the burning regime.