The 50 laboratories of the Netherlands Federation of Thrombosis Services have participated since 1974 in a voluntary external and internal quality control program. The external program comprises a monthly distribution to the member laboratories of a series of artificially prepared control blood samples, two of which are identical. The overall variation of the coagulation times found were 10% (CV) in 1974 and 8% (CV) in 1975 and 1976. Performance improved rather abruptly at the beginning of 1975, after the application of a tight methodological standardization and improvement by the manufacturer of the thromboplastin preparation (Thrombotest) used by the great majority of the laboratories involved. The main source of variation wasfound to be random error in the Thrombotest determination, approximating 6%. interbatch variation of Thrombotest and inter-aliquot variation of control blood samples both do amount to approximately 3%.