PURPOSE We compared the suppression of lingering fat signals in chemical shift selective (CHESS) images by pads filled with flour and pads filled with rice in a phantom and human subjects. MATERIALS AND METHODS First, we prepared a phantom by creating an empty space in a mass of lard and filling the space with air, rice, or flour. Then, we obtained MR images of the phantoms in the center of the magnetic field and at a position 8 cm to the left (off-center) to compare lingering fat signals. MR images of the knee were obtained in 10 healthy volunteers using CHESS after placing a polyurethane sponge pillow, rice pad, or flour pad in the popliteal space under the flexed knee. We visually assessed the number of areas with lingering fat signals and the statistical differences among the groups were assessed using Tukey's test. RESULTS Similarly to rice, flour clearly decreased lingering fat signals in the phantom study. A similar effect was obtained in the off-center images. In the volunteer study, the mean number of areas with lingering fat signals was 2.5 with a sponge pillow, 0.5 with the rice pad, and 0.3 with the flour pad. Those numbers were significantly different using flour pad and rice pad compared with sponge pillow (P < 0.001). No significant differences were seen between flour pads and rice pads (P = 0.662). CONCLUSION Flour pads can suppress lingering fat signals in CHESS images.