The report is devoted to the discussion of the time of reflection of neutrons from the boundary of a homogeneous medium. For a long time, the authors believed that the correct estimate of this value is the so-called group delay time (GDT), introduced by Bohm and Wigner. It is known that the use of the concept of GDT in neutron optics is very productive. The GDT correctly estimates the time relations between the initial and final state wave packets, and when multiplied by the velocity parallel to the surface of the substance, it determines the Goos-Hanchen shift at total external reflection (TER) of the neutron wave. Moreover, the depth of attenuation of the neutron wave during total reflection also coincides with the magnitude of the product of the GDT by the velocity component normal to the surface of the substance.

It is value of the GDT that is measured by the so-called Larmor clock. All this seemed to give no reason to doubt the validity of the GDT representations as a measure of reflection time, which also allows us to assess the depth of formation of the reflected wave.

However, at one time it was realized that when neutrons are reflected from a medium TER threshold, the group delay time turns out to be so small that it contradicts the idea of its connection with the true time of formation of the reflected wave. During such a time, the incident wave simply cannot penetrate into the medium to a depth comparable to interatomic distances, while the origin of the reflected wave is always due to the appearance of waves scattered on the nuclei of the medium.

The report will present a point of view that seems to provide an explanation for the apparent paradoxes. An estimate of the neutron reflection time above the air defense threshold, which seems realistic, will also be given, and an approach to experimental verification of this result will be proposed.