The objective of this study was to determine optimum extrusion-spheronisation parameters for producing bioactive-loaded beads with maximum consistency. The Taguchi method for robust engineering design was applied, using an L-9 orthogonal array (fractional factorial design). The processing parameters analysed were spheronising speed, time and the amount of granulating liquid. The potential importance of some interactive effects was assessed by applying a response surface analysis to the data. The amount of granulating liquid, spheronising speed and the interaction between these two parameters were found to have a statistically significant influence on the size, with the amount of granulating liquid being the most dominant factor. The diameter of beads increased with the amount of granulating liquid and consequently, reduced the yield in the desired range (600-1,180 A mu m). The roundness of the beads was strongly influenced by the spheronising time. The beads became rounder with increasing spheronising time, up to a limit when the effect was observed to level off. Similar optimum conditions were indicated by a response surface method and the Taguchi method: low amount of granulating liquid (50%, w/w) and extrudates spheronised at the highest speed (1,800 rpm) for 15 min. However, the predicted response values using the Taguchi method were physically more acceptable: yield in the desired size range of 98.83%, D-50 of 662 A mu m and roundness of 0.82.