Abstract

Rubus ulmifolius is a highly invasive species that negatively affects biodiversity, especially on Robinson Crusoe Island (RCI), one of the two global mini-hotspots of vascular plants. The rust Phragmidium violaceum has been used as a biological control agent for some Rubus species in mainland Chile, New Zealand and Australia. We aim to quantify P. violaceum pathogenicity over R. ulmifolius on RCI, under both field and laboratory conditions, to determine its potential as a biological control. In addition, in order to broaden our knowledge about the pathogenicity of P. violaceum from RCI, it was also tested on a highly susceptible Rubus species: R. cf. constrictus. We sampled four forest sites on RCI, measuring abiotic, micro-site, and vegetation variables that could be related with infection. In the laboratory, we inoculated P. violaceum into R. ulmifolius seedlings from RCI and R. cf. constrictus seedlings from mainland Chile. We found a low infection incidence in the field, where only 62 of 182 plots presented infected individuals. Applying a fitted generalized linear model, we predicted ten times more infection of P. violaceum on R. ulmifolius when located in southern aspect slopes and shallow soils (< 30cm). In the laboratory we determined that rust from RCI infected R. ulmifolius individuals from the island in low density, and did not infect R. cf. constrictus; this low infection rate reinforces our field results. The low infection on R. cf. constrictus may mean that P. violaceum from RCI could originate from very impoverished genetic material. We recommend testing different rust genetic origins, mainly from the Old World, to increase their genetic diversity.