The multiplication and systemic distribution of two Xanthomonas axonopodis pv. vignicola strains from Benin differing in virulence were monitored after inoculation via leaf infiltration until 25 days post inoculation (dpi) in the susceptible cowpea genotype IT84E-124. The virulent strain multiplied fast in leaves to yield a final population density of about 109 colony forming units (cfu) cm(-2) of leaf, spread through the whole aerial part of the plant within two weeks (10(6) cfu g(-1) of stem) and caused blight, also on non-inoculated leaves at 15 dpi. In contrast, the less virulent strain multiplied slowly in leaves with a final population of about 10(2) cfu cm(-2) of leaf and did not spread through the whole plant until 25 dpi, colonising only the lower stem portion with 2.4 x 10(4) cfu g(-1) of stem. Comparing cowpea genotypes IT84E-124 (susceptible), IT86D-715 (moderately resistant) and IT86D-719 (resistant), considerable differences in bacterial numbers were obtained after infection. At 15 dpi, bacteria had spread in high densities through the whole plant in genotype IT84E-124, in leaves and occasionally in stems in genotype IT86D-715, but in genotype IT86D-719 bacteria were limited to the inoculated leaves. In genotypes IT86D-715 and IT86D-719, bacteria moved predominantly upwards through stems from the point of onset of the inoculated leaf. Contamination of seeds derived from infected plants, varied with the resistance level of cowpea genotypes. There was no relationship between contamination and disease severity. Lower bacteria densities were found in seeds from plants with low disease severity, and no bacteria were detected in seeds of genotypes with high or low disease severity. Most bacteria were concentrated on the seed surface, and only few were found inside the seed under the seed coat. Typical 'water soaked spots' appeared on 17% of seedlings emerging from contaminated seeds of genotype IT86D-472. In conclusion, movement of bacteria on or into seeds depended on genotypes, and seed contamination did not always result in seedling infection. In breeding for resistance, the contamination of seeds and infection of seeds in symptomless plants in apparently resistant genotypes should be determined and considered.