Because the stainless steel large volume pans required for DSC of water-containing samples at temperatures >100 C have a smaller diameter at the top than at the bottom, it was impossible to remove the samples from the pans without damage to the sample. Fitting the pans with a thermally inert silicon rubber insert gave satisfactory results: the insert could be easily removed from the pan with tweezers, which allowed a trouble-free transfer of the heated starch sample to a screw-cap tube for further characterization. With insert, the sample mass at complete filling was 20–25 mg, whilst in conventional DSC without insert it was 50 mg. Preparative DSC was compared to conventional DSC in order to establish the effect of the insert and the reduced sample mass. For potato starch the M1 and M2 transitions were comparable in both types of experiment, but a small shift in the HT endotherm was observed, which was considered acceptable. The homogeneity of potato starch heated to the completion of M1 was checked by