In a situation of observation of a trainee giving a mathematics lesson, we noticed that knowledge resulting from a task developed by researchers for teaching purposes was not circulating as smoothly as expected between the two institutions, research and teaching. This is due to the emergence of certain phenomena produced by the application of this task in the class contingency. Our presentation will show the effects of the didactical transposition, trying to interpret this fact. We will show that it is due to the fact that the pupils' knowledge, in this particular case, is rapidly manifesting from all sides, creating a kind of overflow effect among the trainees, that we have called saturation of the milieu. That means too much knowledge in the milieu at the same time. We observed then, that the trainees stop the pupils' exploration to lead them to the calculation. What appears as a transparent task at the outset is, in fact, highly complex in terms of the knowledge it generates for pupils and interactions with trainees it creates. The underlying theoretical-didactical background is so opaque that it obscures both the mathematical content and the pupils' knowledge. The trainees respond to this overflow by cutting short these interactions that surpass them, then they validate in the place of the pupils and finally institutionalize knowledge, while the pupils have not completed their investigations.