A note on the errors involved in the calculation of elevations of the geoid