A series of experiments was carried out in February 1999 at South Georgia (Subarea 48.3) to examine the effects of different weighting regimes on the incidental mortality of birds caught on longlines fishing for toothfish (Dissostichus spp.). Three regimes were examined, with weights of 4.25, 8.5 and 12.75 kg attached at 40 m intervals on a Spanish-rigged longline. There was a significant reduction in bird mortality when 8.5 kg was used compared to 4.25 kg, but no further significant reduction when 12.75 kg was used. Therefore to minimise seabird by-catch, line-weighting regimes of at least 8.5 kg at 40 m intervals should be used. Results suggested that seabird by-catch on lines using effective weighting regimes may be even further reduced where all lines being set in the vicinity use effective line-weighting regimes. The importance of a good design in experiments of this type is emphasised, because even though conditions may be held as constant as possible within experiments, the ability to separate treatment effects from other sources of variation (e.g. environmental factors) must be preserved.