The paper summarizes the activities of several national metrological i
nstitutes in replacing the kilogram artefact by the mass of a certain
number of silicon atoms. This atomic path requires a very precise dete
rmination of the Avogadro constant N-A with a relative uncertainty of
better than 5 x 10(-8). At present, the most limiting factors are the
variability, from sample to sample, of the isotopic abundances and the
content of impurities and vacancies, and difficulty in the realizatio
n of accurate density standards. Recently, several improvements in mol
ar mass and density determinations have been made, so a further reduct
ion of the level of relative uncertainty towards 10(-7) may be expecte
d in the near future. Beyond that, new experimental methods will be re
quired.