OBJECTIVE To determine the blood lead levels in children and to identify risk factors for elevated levels. METHODS Cross-sectional study. METHODS Vancouver. METHODS Random sample of children aged 24 to 36 months, born and still resident in Vancouver. The sample was stratified proportionally by the median annual family income in the census tract where each family resided. METHODS Blood lead levels and risk factors for elevated blood lead levels, determined from a questionnaire administered to parents. RESULTS Of the children in the sample, 42% (178/422) were ineligible or could not be located. Of the remaining children, 73% (177/244) participated and adequate blood specimens were obtained from 172. The mean blood lead level was 0.29 mumol/L (standard deviation 0.13 mumol/L). (A blood lead level of 1 mumol/L is equivalent to 20.7 micrograms/dL.) The lowest level was 0.06 mumol/L, and the highest was 0.85 mumol/L. Of children with adequate samples, 8.1% (14/172) had blood lead levels of 0.48 mumol/L or higher, and 0.6% (1/172) had a level higher than 0.72 mumol/L. The logarithms of the levels were normally distributed, with a geometric mean (GM) of 0.26 mumol/L (geometric standard deviation 1.56). Of approximately 70 possible predictors of blood lead levels analysed, those that showed a statistically significant association (p < 0.05) with increased blood lead levels were soldering performed in the home as part of an electronics hobby (GM blood lead level 0.34 mumol/L, 95% confidence interval [CI] 0.27 to 0.39 mumol/L), aboriginal heritage (GM blood lead level 0.33 mumol/L, 95% CI 0.28 to 0.39 mumol/L), dwelling built before 1921 (GM blood lead level 0.32 mumol/L, 95% CI 0.28 to 0.37 mumol/L), age of water service connection to dwelling (predicted blood lead level 0.00087 mumol/L [95% CI 0.00005 to 0.00169 mumol/L] higher per year since service connection) and decreased stature (predicted blood lead level 0.018 mumol/L [95% CI 0.0353 to 0.0015 mumol/L] higher for every standard deviation below the age-specific mean height). CONCLUSIONS This study found much lower blood lead levels in children than those found in previous Canadian studies. The authors believe that this result is not an artefact due to differences in population sampling or methods of collection of blood specimens. The study showed no clear risk factors for elevated blood lead levels: although a few factors had a statistically significant association with increased blood lead levels, the differences in levels were small and unimportant.