It is difficult to adjust gob weights from a stirred, molten glass delivery system. The difficulty decreased when rotation of the stirrer stopped. However, this was not a satisfactory solution as it left striations in the glass produced. Small temperature fluctuations and spontaneous changes in the stirrer's position and stroke were found but they were not the major cause of the weight fluctuations. Cyclic variations in weight with an amplitude of about 1% and a period of around 11mins were due to the stirrer's rotation. It became clear that close but approximate matching of the gob and rotational rates was inadequate and that synchronization of these had to be perfect to eliminate this cyclic weight variation. The inevitable conclusion was that run-out associated with stirring was the cause of this problem.