[en] In analysing niobium in low alloy steel with an emission spectrochemical technique using Nb II 3194.977 A as an analytical line, it was found that higher experimental values were obtained in spite of the absence of known interference line. The values of 0.004% to 0.007% were, for example, found in spite of the absence of niobium in the steel samples. Studying this phenomenon, we found the following facts; (1) Among the elements such as molybdenum, iron, tungsten, vanadium and titanium which were classified as interfering elements in the JIS G 1253 (1973), the interferences of titanium and molybdenum with larger spectral intensities can be corrected by the formulae, ΔNb(%)=0.0174Ti (%)-0.0050 (Ti > 0.28%), ΔNb(%)=0.0049Mo (%)-0.0079 (Mo > 1.60%). On the other hand, vanadium do not interfere with niobium, even if vanadium of about 2% is contained in steel samples. (2) The precision (sigmasub(d)) was low for the steel sample containing much manganese. (3) It was demonstrated by this work that a spectral line of manganese is present at the wavelength position of about 3194.87 A in the vicinity of Nb II 3194.977 A. (4) The interference of Mn 3194.87 A on the Nb II 3194.977 A can be corrected by the following formula; ΔNb(%)=0.0034Mn(%)-0.0020 (Mn>0.80%). By the correction of the interference of the Mn 3194.87 A, the precision (sigmasub(d)) of 10 steel samples was much improved and became practically sufficient for the routine analysis. (auth.)