Red spruce (Picea rubens Sarg.) has become increasingly rare across large portions of its range in eastern North America as a result of a general and widespread decline over the past century. Genetic diversity, population genetic structure, outcrossing rates in the filled seeds, and actual inbreeding levels were characterized in five small, isolated, remnant red spruce populations from the disjunct northwestern limits of its range in Ontario and five populations from the larger, more extensive Maritime populations of Nova Scotia and New Brunswick to determine genetic and reproductive status, to provide some benchmarks for monitoring genetic changes resulting from isolation and restricted population sizes, and to assist the development of restoration and conservation strategies. Thirty-seven allozyme loci coding for 15 enzymes were used for genetic diversity assessments, and six of the most polymorphic loci were used for mating system determination. On average, 29.1% (95% criterion) of the loci were polymorphic, the number of alleles per locus was 1.60, and the observed and expected heterozygosities were 0.097 and 0.100, respectively. The Ontario populations were comparable to or slightly less genetically variable than those from the Maritimes. Only 4.7% of the detected genetic variation was among stands; the remainder was among individuals within stands. The Maritime populations were genetically less differentiated from each other than those in Ontario. With the exception of three Maritime populations clustering tightly in one group, there was no clear separation of Ontario red spruce populations from Maritime red spruce populations based on genetic distance as well as canonical discriminant analyses. The average multilocus (tm) and single-locus (ts) population outcrossing rates were 0.595 and 0.558, respectively, indicating a comparatively high tolerance for inbreeding up to the filled seed stage of development in red spruce. The Ontario populations, on average, showed higher outcrossing rates (tm = 0.654, ts = 0.641) than the Maritime populations (tm = 0.535, ts = 0.475). Individual family outcrossing rates were similar to their respective population outcrossing rates and no significant differences were observed among families within populations for the multilocus estimates. When such high levels of inbreeding in filled seeds were combined with the proportions of empty (post-pollination-aborted) seeds, it appears that actual inbreeding levels may vary from 48 to 86%. The highest inbreeding levels occurred in the smallest, most isolated Ontario populations and in those populations most likely to have been affected by poorer pollination conditions. Allozyme variation indicates that in the short term, extant remnants of Ontario red spruce have maintained their genetic diversity and integrity. For artificial restoration of red spruce in Ontario, local seed sources could be used without undue concern over losses of genetic diversity. However, over the longer term, genetic drift and inbreeding may be expected to result in further losses of genetic diversity and (or) reproductive fitness if population sizes, numbers, and distribution continue to decline.Key words: Picea rubens, allozymes, gene conservation, restoration, genetic diversity, population structure, outcrossing rates, inbreeding.