Synopsis A new type of artefact is described, which was observed in a wide variety of inelastic neutron scattering spectra, and which can be easily mistaken as a signature of real excitations.

Abstract Well defined peaks with energies of about 18 meV have been observe d in a variety of inelastic neutron scattering experiments on single crystals as well as on powders using either the triple-axis or the time-of-flight technique. They can easily be mistaken for signatures of real excitations. We have found that they are due to multiple scattering events involving primarily walls of the sample environment. Hence, they are particularly troublesome in experiments using very small samples as have been used with recently developed high intensity neutron spectrometers. It will be discussed what needs to be done to reduce the unwanted scattering to a minimum.