A survey was made to ascertain the reliability of analyses for lead in blood when performed by different laboratories. Specimens of blood containing “natural” lead in the range of 30 to 80 µg/100 gm blood and “spiked” specimens were analyzed by more than 60 laboratories, each by their method of choice. Only the spectrophoto-metric and the atomic absorption techniques were used by enough laboratories to give realistic data on reliability. Statistical analysis of the data showed no significant differences by types of laboratories nor by procedures used. Approximately half of the laboratories in each of two studies and 40% of those in both reported results of acceptable precision. A proposal for a reference sample service for the analysis of biological fluids for trace substances related to industrial exposures has been made.