Cover cropping is a major challenge in the sustainable management of rainfed olive orchards. From a 5 year study (2009-2014), the effect of a mixture of early-maturing and self-reseeding annual legumes (Legs) grown as a cover crop in a rainfed olive orchard was compared with a cover of natural vegetation fertilized with 60 kg nitrogen (N) ha(-1) year(-1) (NV +N), and a cover of natural vegetation not fertilized (NV -N). The study took place in NE Portugal. The following were assessed: (1) the performance of the covers by measuring ground cover percentages, dry matter yields and N content in aboveground biomass; (2) the soil fertility through chemical and microbiological assays and by growing plants in pot experiments; and (3) the nutritional status of olive trees and olive yields. Legumes gave higher ground-cover percentages, produced more biomass and accumulated more N in shoots in comparison to natural vegetation, whether fertilized or not. The results showed intense biological activity [microbial carbon (C) and N, CO2-C evolved in a laboratory incubation, metabolic quotient, total culturable fungi and bacteria, and acid phosphatase activity] in the 0-10 cm soil layer of the treatments producing more biomass (Legs and NV +N). However, soil available N was greater in soil samples from the Legs plot. N recoveries by turnip (Brassica rapa var. rapa L.) and rye (Secale cereale L.) grown in pot experiments were 84.4 and 60.2 mg pot(-1) in soil samples from the Legs treatment and 29.4 and 27.1 mg pot(-1) and 14.2 and 13.6 mg pot(-1), respectively in NV +N and NV -N plots. Sown legumes appeared less effective in increasing organic C than natural vegetation. Nevertheless, in the Legs plot the increase of easily mineralizable C was proportionally higher than the increase of total organic C, which may mean that a more reactive pool of organic C is created, which may reduce the turnover of organic C and N in the soil. In the 10-20 cm soil layer, total organic C was significantly lower in Legs (14.0 g kg(-1)) than in NV -N (22.1 g kg(-1)) and N +NV (25.2 g kg(-1)) treatments, likely due to a priming effect caused by mineral N coming from the surface layer. Two years after the trial started, the N nutritional status of the olive trees was significantly higher in Legs than in natural vegetation plots even when 60 kg N ha(-1) year(-1) was applied. The cumulative olive yields in NV -N and NV +N plots were only 58.6 and 77.7 % in comparison to those found in the Legs plot, if only the last four harvests were considered, which were those influenced by the ground-cover treatments (2010-2013).
