Resumen
Ptolemy (about 150 AC) modeled atmospheric refraction influencing Al Farghani (831), Alhazen (1020), Sacrobosco (1256) and Witelo (1278): the Sun was supposed bigger at horizon like a coin appears under water in a curved bottle. The correct work of Ibn Sahl (984) remained forgotten. Tycho measured the refraction on the 1572 supernova at various altitudes. Harriot, Kepler, Snell and Descartes found independently the refraction law after 1600. A modern formulation of vertical (0.5" zenithal to 35' at horizon) and horizontal (0.5" at all altitudes) differential refraction of solar diameter appears in Du Séjour (1786). Laplace's formula (1805) computes the vertical deformation of the solar disk, while the horizontal reduction of 0.5" is proportional to the chord's length. Dicke (1967) measured the solar oblateness to determine dynamical constraints to alternative theories of General Relativity. The Astrolabe of Rio de Janeiro measured in 1998-2009 the solar diameter at all heliolatitudes, by timing solar transits across fixed altitude circles: an equatorial excess larger than RHESSI (2008) and SDS (1992-2011) data remains after refraction's corrections. Meridian transits series measured at Rome Campidoglio (1877-1937) and Greenwich (1850-1940) behave as Rio data: the scatters between annual averages were larger than statistical dispersions of each value (Gething, 1955). Anomalous refractions measured with Rio Heliometer (2013) are low frequency seeing (0.01 Hz) acting to scales of the solar diameter (32'): they affect transits measurements with random perturbations hundreds times larger than the expected values calculated from the timing accuracy. These perturbations enlarge the differences between averages values binned either in time or heliolatitude: they are larger than statistical dispersions, suggesting a wider binning. The ``adiabatic" approach of Rio Heliometer with high frequency measurements ``freezes" the slow seeing image motion component.
