Using Norway 1948-2004 as a case, I test whether changes in variables related to social integration can explain changes in suicide rates. The method is the Box-Jenkins approach to time-series analysis. Different aspects of family integration contribute significantly to the explanation of Norwegian suicide rates in this period. The estimated effect of separations is stronger than the effect of divorces, both for men and women, probably because separations are closer in time to the "real" marital breakup. This difference has not been demonstrated in earlier time-series research. Marriages decrease the suicide rates for males. The unemployment estimate for men has a negative sign, contributing to fewer suicides. Both increasing alcohol (beer) consumption and fewer marriages seem to be implicated in the soaring suicide rate for young men since 1970.