The frequency of two non-metric skeletal traits, atlas bridging and clinoid bridging, were examined serially in a randomly chosen sample of 147 families who participated in the Burlington Growth Study. The sample is representative of an Ontario white Anglo-Saxon population. Partial and complete atlas bridging occurred in 29.2% of the sample, partial and complete clinoid bridging in 15.2%. Atlas bridging appears at an average age of 10.7 years, clinoid bridging at seven years, demonstrating that these characters are not simply expressions of soft tissue sclerosis in old age. Both traits show no strong associations with bone robusticity although atlas bridging is slightly more frequent in males. Both traits are more frequent in relatives of affected individuals than in the sample as a whole. Correlations between parents and offspring and between sibs are highly significant for atlas bridging, less so for clinoid bridging. These traits should fir either a single gene or quasi-continuous, polygenic model of inheritance. Several tests for polygenic inheritance such as the correlation between first and second born sibs' trait condition, the relationship between trait expression in offspring and total trait incidence in affected parents, and the correlation between trait frequency and expressivity on an intergroup basis were all positive for atlas bridging. The evidence for polygenic inheritance of clinoid bridging is weaker but suggestive. The results obtained in this study for atlas bridging are comparable to data from one earlier family study. The evidence suggests that these two traits should prove useful as genetic markers in skeletal population studies although there is still need for careful control over trait observation and description. Future research should attempt to measure non-metric traits continuously when their underlying distributions are known to be graded.