The main objective of this work was to gather evidence as to whether the yeasts present in ripe grapes before harvest and those found in spontaneous wine fermentations came from the vineyard soil which could then be regarded as a natural reservoir for these yeasts. Two types of management system were tested in each vineyard: conventional tillage (CT), and no-tillage with natural green cover vegetation (NV), both under semi-arid rainfed conditions. Bacteria isolated from the grapes all corresponded to three genera (Pseudomonas, Stenotrophomonas, and Bacillus) that were very abundant in soil samples taken just before grape harvest. The amounts of fermentative yeasts in vineyard soil increased significantly during the dates close to harvest. Some yeasts were isolated from soils and spontaneous fermentations (Saccharomyces and Lachancea), while others were only isolated from fermentations (Hanseniaspora, Metschnikowia, and Pichia) or from soils (Torulaspora). Saccharomyces yeasts were isolated from vineyard soil only after grape harvest. The analysis of sterile-must fermentations inoculated with soil samples showed that soil was not the origin of the most abundant fermentative yeasts in spontaneous grape fermentations (Saccharomyces and Hanseniaspora). In contrast, other fermentative wine yeasts such as Lachancea and Torulaspora seemed to be permanently resident in the vineyard soil, especially in the NV vineyard. The yeasts involved in spontaneous grape fermentation (mainly Hanseniaspora and Saccharomyces) must have reached ripe grapes by a process other than the mere accumulation of wind-borne soil dust. Conversely, vineyard soil did appear to be a permanent natural reservoir for Torulaspora and Lachancea yeasts, especially in the NV-vineyard.