The bacterial isolates studied, I112, II232 and U410, originated from cabbage roots and were previously selected in a screening for activity as a seed treatment against Pythium ultimum on cucumber and kale. Isolates I112 and II232 were identified as Pseudomonas chlororaphis and P. putida biotype A, respectively, by fatty acid analysis. Isolate U410 could not be identified with this method. Based on partial sequencing of the 16SrDNA, U410 was assigned to the complex Xanthomonas/Stenotrophomanas/Xylella within the gamma-subgroup of proteobacteria. On cucumber, seed treatment with isolate U410 was more efficacious against damping-off caused by P. ultimum than with isolates I112 or II232. A brief dip of cucumber seeds into a suspension of isolate U410 resulted in a similar level of control of P. ultimum than treatment for 5, 30 or 120 min. As a seed treatment, none of the three isolates gave statistically significant central of Pythium aphanidermatum on cucumber. Seed treatment of peas with I112 or U410 provided some control of Rhizoctonia solani (efficacy 25/30%). When applied as a soil drench, isolates I112, II232 and U410 protected cucumber against P. ultimum (efficacy: 24-44%), and cotton against P. ultimum (efficacy: 30-55%) and R. solani (efficacy: 23-44%). A root dip into suspensions of I112 or U410 protected greenhouse and field-grown strawberry against Phytophthora fragariae var. fragariae. The bacteria showed no activity against Phytophthora nicotianae on tomato. In non-sterile bur pathogen-free potting media, seed treatment with isolate II232 tended to increase the shoot dry weight of maize and pea, isolate I112 caused a reduction of the dry weight of maize. Isolate U410, and Bacillus subtilis isolated from the commercial product KODIAK(R) tended to promote the growth of cucumber and kale. The effects of seed bacterization in the absence of pathogens were, however, difficult to interpret overall, as the above mentioned results on growth promotion or inhibition were either not seen in all experiments or statistically not significant.