The main aspects of an absolute method for measurement of the mixing ratio
of atmospheric carbon monoxide (CO) are presented. The method is based on c
ryogenic extraction of CO from air after oxidation to CO2 followed by accur
ate volumetric determination. Gravimetric measurement is used to determine
the quantity of sample air processed. In routine operation the overall erro
r can be kept below 1%. Furthermore, the results of a laboratory intercompa
rison are analyzed. It is shown how offsets in the commonly applied analyti
cal methods can occur and how these can seriously affect results obtained a
t the low concentration end (< 100 nmole/mole).