This paper examines the experimentum crucis under the light of the Duhem’s holistic thesis. This methodological instrument is not usable in physics, because physical theories are always logically connected to many assumptions. On the contrary, it is usable in psychological research oriented to perceptual laws, when these laws are, without any hypothetical term, isolated systems. An application of experimentum crucis in Experimental Phenomenology of perception is presented. In conclusion, the role of perceptual knowledge as an essential assumption in other scientific disciplines that have a high degree of theoricity is also underlined.