AbstractThe Beckman flame spectrophotometer was studied to determine its application to the determination of magnesium in soil extracts. A high flame temperature and a narrow slit width increased the intensity of Mg (2852 Å) relative to the flame background. Centrifuging satisfactorily cleared extracts and prevented most clogging of the atomizer. Fe, Ca, and Na caused significant interference at the Mg band head of 3708 Å particularly with Mg concentrations below 2.00 ME/100 ml. At the 2852 Å line, Ca in concentrations greater than 8 ME Mg/100 ml depressed the Mg intensity; Al or PO4 in concentrations greater than 100 ppm Al or P slightly depressed the Mg intensity, whereas Na in concentrations greater than 10 ME/100 ml increased the intensity. Other elements tested (K, Mn, Fe, S, and Si) at 2852 Å had no apparent effect within concentration ranges usually found in ammonium acetate (1N) and 0.05N HCl soil extracts. Twenty‐four different soils of Virginia were analyzed for exchangeable magnesium by the photometric method, and results did not differ significantly (5% level) from the chemical method. The average recovery of Mg added to 0.05N HCl extracts was 103% and the standard deviation was 5.3%. The most favorable instrumental conditions found were: wave length, 2852 Å; slit width, 0.15 mm; propane pressure, 2.0 cm toluene; oxygen pressure, 75 inches H2O; air pressure, 18 pounds per square inch; sensitivity control, about one turn from the counterclockwise limit. The intensity scale was set on 100 when a solution containing 2.0 ME Mg per 100 ml was atomized and the milliammeter needle brought to zero with the sensitivity control.