Ground-based results on cyclic variations of the apparent solar radius are so far controversial and inconsistent. This is blamed to atmospheric noise which effects can be so severe that even in cases in which the observations are made with similar instruments, the results show strong disagreements (Li et al.: 2003, Astrophys. J. 591, 1267). Such claim concerns the results of Danjon astrolabes which during the last decades have been used widely at several sites for solar metrology. The long-term series with thousands of radius measurements made with astrolabe at Calern, France, and at Santiago, Chile, is a case in which the results of radius variations in time are strongly discrepant in spite that the observations were made simultaneously, in quite similar conditions and with almost identical instruments (NoÂ¨el: 2004, Astron. Astrophys. 413, 725). However, we show here that most of astrolabe discrepancies may be due to data analysis biased by theoretical preconceptions, by empirical results which without scientific arguments are considered as canonical references and by over interpretations of casual agreements between visual and CCD astrolabe results.