A major problem in the continuous casting of aluminium killed strip steel is the incidence of black and grey lines in the strip surface caused by the alumina inclusions. One place for limiting the number of alumina inclusions is by providing favourable fluid flow conditions in the ladle so that the inclusions float out to the surface or adhere to the ladle wall. In laboratory equipment 20 kg of steel was melted in an alumina crucible of 14 cm diameter in an induction furnace. Deoxidation was carried out while the melt was constantly stirred by argon bubbling. Bubble frequencies were measured at the lance tip using a microphone technique. High speed cine photo- graphy was incorporated to measure melt top surface velocities. Experimental results indicate that the rate of removal of inclusions is a function of both the gas input and the power input. However due to unfavourable surface area to volume ratio, reoxidation of the melt took place. Some other experimental difficulties in working at high temperatures are also discussed such as lance design, melting arrangement and sampling techniques.