AbstractThe interlaboratory variations in the fumigation extraction method and the analytical procedures for measuring C and N in the soil microbial biomass were tested with one soil sample, and two soil extracts (non‐fumigated and fumigated) sent to 25 different laboratories. Four groups of analytical procedures for organic C, i.e. (1) oven oxidation/ IR detection, (2) UV‐persulfate oxidation/lR detection, (3) UV‐persulfate oxidation/colorimetric detection and (4) dichromate oxidation/ titration, and three groups for total N, i.e. (1) Kjedahl reduction to NH4+, (2) UV‐persulfate and (3) persulfate‐borate oxidation to NO3−were used by the different laboratories.The coefficient of variation for C and N measurements between different laboratories and analytical procedures varied between 15 and 34% in non‐fumigated samples, between 13 and 20% in fumigated samples, and between 12 and 24% in the differences Ecand EN. The average coefficients of variation between the replicate measurements within one laboratory were much smaller, i.e. they varied between 3.0 and 9.2% in non‐fumigated samples, between 2.4 and 5.5% in fumigated samples, and between 4.5 and 12.8% in the differences Ecand En. Extraction and fumigation were not the major source of the variations observed. They were mainly a result of differences in the analytical procedures used to measure the low concentrations of C and N in the extracts. However, all of these analytical procedures should be able to measure correct values if they are properly calibrate