Sea lice are ectoparasites of salmonids that have been associated with the recent decline in sea trout numbers in north-west Europe. Observed patterns of louse load distribution between sea trout in the seas surrounding the UK, Ireland and Norway and a simple model have been used to analyse the epidemiology of lice. Loads are aggregated and deviate strongly from the Poisson distribution, although less than is observed with many other parasites. The louse numbers on fish from offshore sites are slightly less variable than for fish from coastal sites with comparable mean loads. Analysis of louse development stages and sexes shows that selection between hosts by sea lice plays a limited role. If host selection is absent, then associated poor condition would be caused by, not the cause of, high louse burdens; however the absence of such selection is not proved. Scenarios with infection that is patchy in space and time best generate the aggregated load patterns observed; these patches accord with observed swarms of copepodids. Prevalence patterns may indicate the movement of trout between environments. Control of copepodids in infection 'hot spots', either directly or through control of louse egg production in their catchment, may reduce louse loads on wild sea trout and, in particular, extreme and damaging loads.