A possible explanation of the experimentally observed behaviour of the cooled gravitational-wave antennas