Mating system parameters of a northern conifer, Pinus banksiana Lamb., were estimated from allozyme polymorphisms. Seeds analyzed were obtained from serotinous cones of 30 individuals and represented four independent fertilizations in 1975, 1976, 1977 and 1978. Results indicated that a mixed mating system model, with a mean effective outcrossing rate of 88 +/- 0.047%, described the mating system of this stand. However, there was an approximately linear increase in the apparent selfing rate from the oldest (1975) to the newest (1978) crop. Two hypotheses could account for these observations. First, there may have been changes in the mating system during the 4-yr period, but linearity of the differences observed in this study may have been due to chance. These changes were, however, independent of the variability of the observed pollen pool. This indicated that they were not a result of different proportions of outcrossed zygotes directly observed. Second, there could have been a more or less constant amount of selfing, followed by a differential loss of viability of selfed and outcrossed zygotes during the period of storage in the cones. Under this hypothesis, selfed zygotes are at a selective disadvantage relative to outcrossed zygotes. No differences in the mating system could be demonstrated among the three crown strata of this stand. There was significant interlocus heterogeneity in the filial generation genotypic distributions and in the estimated outcrossing rates, reflecting the complex nature of forces that can affect single-locus estimates. There was evidence of some additional inbreeding, possibly due to family structures in the stand; however, this was a minor component of the total inbreeding.