The experiments of this project should help to clarify the existing discrepancy between, on the one side, the proven ability of the marmosets to use the mirror for searching for hidden food and, on the other side, their considerable difficulties in dealing with both the classical (colour) and the motivation-enhancing mark test (chocolate). If, in our previous experiments, the animals had understood how a mirror functions in general, then they should also have been able to solve some cognitive problems which are more or less closely linked with the standard situation to which they had been already accustomed. Hence we devised a series of increasingly more difficult tests in which we intentionally modified the frame conditions in a manner that was thought to force the animals to approach the mirror from different perspectives. The main questions which are to be addressed by these experiments are the following: 1) Do the animals possess object permanence of higher Piagetian levels (V, VI) that are commonly thought to be important cognitive prerequisites to correctly understand the virtual, that is symbolic nature of a mirror image? 2) In the standard situation for the instrumental use of mirror are the animals able to grasp for a moving object? 3) Can the animals transfer their shown ability from the standard situation to new variations of that same situation where the mirror is offered under different spatial conditions? (e.g. after rotation of 90°) 4) Do the animals know how to correctly choose a short detour to get to the hidden reward? 5) Do the animals know their own body? For instance, are they able to consciously hide their own body behind a screen to avoid being seen? 6) If they have a certain knowledge of their own body and if, at the same time, they also know how to correctly interpret a picture in a mirror, are they then able to discriminate the picture of a potential danger (water pistol aiming at them through a mirror) from a real danger at the same place (water pistol aiming directly at them)? 7) After such an intensive training with different aspects of the mirror problem, are the animals ready to better get on with the self-recognition test if, in such an additional test, all known factors (gaze aversion, motivation, experience, age) that might limit successful performance in monkeys have been accounted for? In sum, the planned experiments should allow us to explain in more detail the reasons most monkeys and, in particular, marmosets have such problems with self-recognition. This will be done by identifying step by step the extent of their ability to transfer the solution of the mirror problem from one situation to the next one, thereby revealing the true nature of their understanding of reflections.