All completed suicides during the 42years in the counlJ..' of Viisterbotten in northern Sweden (1}466cases) iuere analYzed by ,nultiple logistic regressions to relate suicide methods Ivith gender, age, place of residence, marital status) year of suicide) season ofsuicid«, and season of birth. The ratio ofpreferences between hanging andf irearlJls remained stable throughout 1952-1993, both to be oterualeen in preference first bj' petrol gases andthen bypoisoning in recentyears. Firearmstuereprefetred significantlY more often by males and in rural areas having a long tradition of hunting wild life. Poisoning and drowning were preferred '710re often by [essales, nonmarried (single) diuorced, widolvIwidower) andin urban Umed. Drowning iuas preferredaround stosmerand,?:y the older group. Petrolgases were prejerred byyoungerperso1ts, males, 'Harried andin urban Umed. Those born during the season January to A pril iuere 1110re like!J toprefer hanging rather thanpoisoning orpetrolgases. Preference depends both onavailability and acceptability. A ssociations between preference for the suicide methods andtheir availability were found. "V T ariations in preference were found with respect to the individual's sociodemographic and temporal circumstances. No criteria validgenerallY across the sociodemograpbic variables j'ustified a grouping· of the methods into violent and nonviolent.