Fillers have been reported to raise, have no effect upon, or to lower the glass transition temperature Tgof polymers. In those studies, comparisons have been made between filled and unfilled polymers having equal thermal histories. In the work report here, however, the thermal history (cooling rate) was also varied. Two systems, polystyrene-silica and epoxy-rubber were studied. The glass transitions were measured by using a differential scanning calorimeter and a method of analysis first described by Ellerstein and later elaborated upon by Flynn. The data could be represented by a linear plot of the log of cooling rate versus Tgor Tg−1. For the polystyrene system, the filled material had a larger negative slope for the plot of log cooling rate versus Tg−1so that after fast cooling the filled material had a lower Tgthan the unfilled material. However, when the cooling rate was lowered, the Tgof the filled and unfilled materials approached each other, and for very slow cooling (annealed samples) the filled material had a higher Tgthan the unfilled one. For the epoxy-rubber system, the rubber lowered the Tgby an amount which increased with cooling rate. Thus the extent and even the direction of the change in Tgwith filler is shown to be dependent upon thermal history. It is shown that the slope of the plot of log cooling rate versus Tg−1or Tgis related to an activation energy. A new simplified method of correcting Tgfor DSC instrument thermal lag is described.