Methods to determine lead and cadmium at the low mug/kg-level in milk were investigated. Methods tested were differential-pulse anodic-stripping voltammetry (DPASV) and flameless atomic-absorption spectrophotometry (FAAS). Under the circumstances in which these methods were used, the analytical procedure based on DPASV was the most sensitive, with an estimated detection limit of about 0.2 mug/1 for lead and cadmium in liquid milk. The pretreatment of the milk samples before the DPASV-analysis included: freeze-drying, ashing in a muffle furnance at 550 degree C and finally dissolution of the ash in 0.1 M-hydrochloric acid. The recoveries of known amounts of lead and cadmium added to milk were 95 and 59%, respectively. A survey of the contents of lead and cadmium in the Swedish manket milk was performed. This investigation showed that the average lead content in Swedish market milk was 2.0 mug/1, with a standard deviation of 0.5 mug/1. The cadmium content was below 0.2 mug/1 in all samples analyzed.