ACCORDING to a note in Physical Abstracts1, R. L. Mullokandov2 finds that thermal conductivity of powders is practically due only to the convection in the surrounding gas, and that hence "a layer of smooth spherical particles in vacuum may be used as a good thermal insulator". This was discovered by M. Smoluchowski some forty years ago3. He investigated both theoretically and experimentally the thermal conductivity of powders, and concluded that filling the space between the walls of a Dewar vessel with a suitable powder would provide improved insulation by reducing the convective currents as well as the radiation and, moreover, would improve the vacuum by absorption. His application for a patent resulted in a long argument with the German Patent Office, whose examiners were reluctant to admit that any material agent could provide a better insulation than 'absolute' vacuum. At that time a large metallic Dewar vessel with its interstice filled with powdered active carbon was built at the Physical Institute of the University in Lwow, and showed the expected high insulating properties. As that work of Smoluchowski remained little known, I think that this information may be of interest.