1 Introduction

The horizontal surface of a fluid realizes, apparently in the most accurate way, a horizontal plane (tangent to its surface at a certain point), and is used in specialized applications with astronomical instruments: astrolabes, Photo Zenith Tubes (PZT), etc. Here, mercury in a shallow plate forms a horizontal mirror - a "mercury horizon". Its reflectivity is about 60-75% (depending on the wavelength of light), but may be somewhat lower if the surface of the mercury is not perfectly clean. Chemical reaction with some gases in the atmosphere may tarnish it; additionally, dim spots may appear over time (within an observation), reducing the mercury's reflectivity and also affecting the optical quality of images.

A higher and more stable reflectivity could be achieved with a flat mirror floating on a fluid - a "floating mirror horizon". With an aluminium coating it would reflect about 25% more incident light than the surface of mercury, and the images would have a better (sharper) definition. Its weight would render it less susceptible to vibrations; and there would be no need to clean the surface of the mercury, making the particular instrument permanently ready for observations. Using this appliance with astrolabe types of instruments, both images of a star would have practically the same brightness.

To achieve an accuracy in the definition of the horizontal (or vertical) direction, that is comparable to that of the (classical) mercury horizon, is the problem to consider. Its theoretical aspects, and possible design of the relevant device, are discussed below. Results of some trials with a small experimental model will also be given.