Abstract
To explain the occurrence of the sharp cusp of the low-field susceptibility chi of spin glasses, some theories have been put forward which lead to a special type of phase transition at a critical temperature T0. It is shown that the straightforward application of such theories to deduce the magnetic part cm of the specific heat leads to some anomaly of this quantity at T0. More accurate measurements of cm and chi near T0 are necessary. At the moment it looks like a contradiction between theory and experiment.