The first plates for the great enterprise, the Carte du Ciel or the Astrographic Catalogue, were taken in 1891. At that time no fundamental system existed which was usable as a basis for reference stars over the whole sky and was commonly accepted by the astronomical community. Therefore reference stars were taken from available catalogues, sometimes updated to the epoch of the plates by application of proper motions. Other observatories used their own reference stars specially observed on transit circles. For the last-mentioned method the work at the Cape Observatory is an instructive example which was described by Sir David Gill, an experienced observer, in a catalogue of Astrographic Standard Stars (Gill, 1906). The methods of observation are fully described in the volumes of the Cape Meridian Observations. For the derivation of proper motions southern catalogues dating from 1835 to 1900 were used, applying Newcomb's values for precession. The determination and elimination of serious systematic errors played an important role. “Eye-and-Ear” observers had their personal errors which even changed with the years for the same person, occasionally even during tiresome long watches in a single night. Further sources of errors include: None-reversible instruments, zenith discontinuities, magnitude equations, possible changes of flexure, pivot errors and their variations, and unsufficient knowledge of the polar motion in old catalogues. Many of these questions were investigated and described in extended introductions to observational catalogues but in some cases such problems were only incompletely explained. Reference stars for the Astrographic Catalogues were observed on various “fundamental system” if fundamental at all.