We studied age- and sex-specific mortality from multiple myeloma in Spain during the period 1957-1986. The presence of birth cohort and period effects was determined using a multivariate Poisson model. From an international perspective, multiple myeloma mortality in Spain during the 1980s was in the medium range. Adjusted rates for males and females were 1.36 and 0.96 per 100,000 person-years, respectively. Multiple myeloma mortality rose exponentially during the study period, the annual increase was 11.0% in males and 10.5% in females, surpassing the increase observed in other countries. The analytical model ascribed most of this increase to a birth cohort effect. The interpretation of this phenomenon is difficult. Progressive improvements in detection and reporting, and an actual rising incidence may underlie the cohort effect.