The effect of pre-loading and in situ loading of cobalt onto a cobalt-limited granular sludge on the performance of methanol fed bioreactors was investigated. One upflow anaerobic sludge bed (UASB) reactor was inoculated with cobalt pre-loaded sludge (24h; 30degreesC; 1 mM CoCl2) and a second UASB with unloaded sludge. The UASB reactors (30degreesC; pH 7) were operated for 77 days at 8 h hydraulic retention time and organic loading rates ranging from 5 to 20 g COD . L reactor(-1) . d(-1). Cobalt pre-loading clearly stimulated the methanogenic activity of the sludge with methanol as the substrate, e.g., after 30 days of reactor operation this activity was 5.8 times higher than that of the cobalt unloaded sludge. During the experiment, part of the cobalt leached from the pre-loaded sludge, i.e., 54% of the cobalt content was lost during the 77 days of reactor operation. Sequential metal extraction showed that losses mainly occurred from the exchangeable and carbonate fraction and in the sludge remaining cobalt was mainly present in the organic/sulfide fraction of the sludge. In situ loading of cobalt in the unloaded UASB reactor on day 57 by adding 31 muM cobalt to the influent for a 24-h period (16% of the cobalt present in the loaded sludge at day 11) resulted in a 4 time increase of the methanogenic activity of the sludge with methanol as the substrate at the end of the reactor experiment, while the accumulated amount of cobalt in the sludge only amounted to 6% of the cobalt accumulated in the loaded sludge (on day 11). This study showed that both pre-loading sludge and in situ loading are adequate for achieving an increased reactor performance of methanol fed UASB reactors operating under cobalt limitation. However, the in situ dosing procedure needs substantially lower amounts of cobalt, while it also gives significantly smaller losses of cobalt with the effluent. (C) 2004 Wiley Periodicals, Inc.