In the first of 2 experiments, a simulated voyage was conducted to examine the effects of various treatments on bodyweight change and feeding frequency of inappetant sheep at the end of lot-feeding (non-feeders). The treatments, applied during simulated shipping, were: normal quantities of feed and length of troughs; extra trough length; extra feed. Adult Merino wethers (n = 108) were used in each treatment. A voyage to the Middle East was then conducted to establish whether shipboard mortality could be reduced by separating non-feeders (n = 305) from feeders (n = 5,620) late in the feedlot hase and housing the groups separately aboard ship. A control group of non-feeders (n = 215) mixed with feeders (n = 5,732) was used for comparison. Bars (marker bars), containing a coloured dye, were attached to feed troughs to mark sheep that fed. Most non-feeders (82%) began eating when placed in shipping pens in both studies. However, there was no significant difference in percentage of sheep that fed between non-feeders given extra trough length or extra feed compared with non-feeders given standard management at any stage of simulated shipping. There was no significant difference in mean bodyweights between treatment groups on days 1, 8 and 15 of simulated shipping. Differences in bodyweight on d 22 were probably associated with different levels of gut fill. Death rates were not significantly different in separated and control groups (1.1%, 0.9%, P = 0.6) in the voyage of 14 d to the Middle East.(ABSTRACT TRUNCATED AT 250 WORDS)