Nine reference sera have been investigated in the same types of analyses in nine laboratories. Seven sera were of human origin, and two of animal origin. The analyses were repeated six times. Comparing the apparent interlaboratory variation of each reference serum, specimen associated effects were demonstrated. The effects were often considerable, and they were detected for nearly all of the analytes investigated: albumin, alkaline phosphatase, amylase, bilirubins, calcium (II), cholesterol, cobalamines, creatininium, creatine kinase, glucose, lactate dehydrogenase, magnesium (II), protein (total), and sodium ion. The effects are discussed according to their nature, viz. matrix effects, analyte simulation effects, and analyte concentration effects. These effects seemed distinct when poor methods were compared. Matrix effects were associated with lack of specificity, and analyte concentration effects with lack of linearity. The lyophilized serum from outdated bank blood showed the best agreement with fresh frozen pool of patients sera; the mixed bovine-equine lyophilized serum (AutonormTM, Nyegaard & Co. AS, Norway) showed the most frequent disagreements. Matrix and analyte simulation effects were probably the main cause.