Several sources of ignition were considered and eliminated. At first static electricity was considered to be the obvious source but the filling rate was so low and the conductivity of the liquid so high that sufficient charge could not accumulate. A discharge from a mist of oil droplets was considered but an experiment (with water) showed that liquid falling the distance that it actually fell at the actual flow rate (2 m3/h) did not produce any significant amount of mist. The walls of the tank were examined for traces of pyrophoric or catalytically active material and the liquid was examined for traces of peroxides but nothing was found.