Abstract
Longstanding puzzles on the Fermi surfaces in CeB6 and PrB6 have been resolved in a combined theoretical and experimental study. In CeB6, even after a metamagnetic transition, no up or down spin splitting occurs, except a small split on the small pocket surface. The heavier mass branch can be seen only by acoustic de Haas van Alphen measurement. In this case, the observed effective mass is consistent with the large T-linear specific heat. In PrB6, the splitting can be seen on all the branches but in the opposite sense, that is, the larger-area branch is observed as the smaller-area branch. Combined with the crossover characteristics of the pocket band, the puzzle of a pocket which is too large compared with the neck size is solved.