Bank filtration offers a cost effective and low maintenance technique for t
he removal of cyanobacterial hepatotoxins from drinking water. For bank fil
tration to be effective, the toxins must be degraded. The broad aim of this
research was to determine whether the hepatotoxins, nodularin and microcys
tin-LR, could be completely removed from the soil/water matrix of three soi
ls by microbial degradation. The results indicated that complete toxin remo
val was possible within 10-16 d in 2/3 soils that were incubated in the dar
k at 20 degreesC. The soils with the highest organic carbon content (2.9%)
and the highest clay content (16.1 %) were the most effective at removing t
he toxins in batch experiments. However, the sandy soil (98.5% sand) was in
capable of degrading either toxin. The half-lives of toxin losses due to ad
sorption, desorption and degradation were calculated and for all soils. The
degradation process had the highest half-life for both toxins. This sugges
ted that degradation was likely to be the rate-limiting step of complete to
xin removal. It was concluded that when a bank filtration site was being ch
osen, the degradation potential and the textural properties of the riverban
k soil would be important when considering complete removal of cyanobacteri
al hepatotoxins.