As part of the certification campaign of three candidate reference materials for the determination of aflatoxin M1 (AfM1) in whole milk powders, homogeneity, short- and long-term stability tests of naturally contaminated milk powders have been performed. The homogeneity of two AfM1-contaminated milk powders was studied by taking samples at regular intervals of the filling sequences and analysing in triplicate for their AfM1 contents by liquid chromatography with fluorescence detection (LC-FLD) using random stratified sampling schemes. The homogeneity testing of an AfM1 'blank' milk powder material was performed by determining the nitrogen content because AfM1 levels were below the limit of detection of the most sensitive determination method. The short-term stability of AfM1-contaminated milk powders was evaluated at three different storage temperatures (4, 18 and 40 degrees C). After storage times of 0, 1, 2 and 4 weeks, samples were investigated using LC-FLD. The long-term stability study comprised of measurements after 0, 6, 12 and 18 months after storage at -20 and 4 degrees C. Analyses were done by LC-FLD. Based on the homogeneity tests, the materials were sufficiently homogenous to serve as certified reference materials. Corresponding uncertainty contributions of 0.23-0.89% were calculated for the homogeneity. The stability measurements showed no significant trends for both short- and long-term stability studies. The long-term stability uncertainties of the AfM1-contaminated milk powders were 7.4 and 6.3%, respectively, for a shelf-life of 6 years and storage at -20 degrees C. Supplementary stability monitoring schemes over a long period of several years are currently ongoing.