3C 273 is the second radio source identified with a stellar object (Hazard, MacKey & Shimmins 1963) and the first for which the emission lines were identified with a redshifted Hydrogen sequence (Schmidt 1963). It was immediately clear that if the redshift is a measure of the distance to the object its luminosity must be extraordinarily large. (Schmidt 1963) already recognized the presence of a faint jet-like nebulousity. Early work on quasars has been summarized in (Burbidge 1967) and 2 years later by (Schmidt 1969).
Early spectroscopic work interpreted the emission lines already identified in the frame of the then familiar gaseous nebulae theory (Greenstein & Schmidt 1964). These authors found that the line emitting gas must have a temperature of the order of 17 000 K and electron densities around 106 cm-3. It must be stressed that there was then no distinction between the narrow and broad line regions. These interpretations made it clear that the heavy element abundances must be close to the values found in galactic nebulae, hence close to those found in young stars. This was soon found to be a problem if the objects were to be at the cosmological distances suggested by their redshifts (Shklovsky 1964). Indeed in this case it was recognized that the light emitted by quasars must have originated when they were very young and it was not expected then that extensive stellar nucleosynthesis takes place in so short a time after the formation of structures in the expanding Universe. It now seems well established that star formation is closely associated with the process of accretion of matter from the galaxies to the very central regions in which nuclear activity takes place. Thus explaining why no nuclear activity involving only Hydrogen and Helium is observed.
Soon after the discovery of 3C 273 as an unusual object existing plate collections were used to obtain light curves. Using the Harvard (Smith & Hoffleit 1963) and Pulkovo (Sharov & Efremov 1963) plate collections it was then realised that the object had varied by a factor of approximately 2 in the course of the preceding decades, although no such variations had been observed with the then modern data. Radio flux variations were also soon found in one of the components of 3C 273 (Dent 1965).
It was also noted early that the continuum spectral energy distribution of 3C 273 (and Seyfert galaxies) might cover all the known bands of the observable electro-magnetic spectrum. The emission is such that roughly the same amount of energy is observed in the different parts of the spectrum. This remark followed the possible detection of X-rays from 3C 273 reported by (Friedman & Byram 1967) and the observation of an important infrared flux (Pacholczyk & Weymann 1968).
The main observational pieces of the problems posed by quasars (extreme luminosity, variability and hence compactness and emission covering most of the observable electromagnetic spectrum) were thus in place in a very short time after the first identification of point like radio sources with red shifted optical sources.
The very unusual properties of 3C 273 and a few other quasi stellar sources as compared to any of the astrophysical objects then known (stars, nebulae, supernovae etc) provoked a wide variety of possible explanations summarized in (Burbidge & Burbidge 1967). The necessity to use concepts very different from those related to nuclear processes to explain the energy requirements of Quasi Stellar Objects (QSOs), if these are at the distances implied by their redshifts, even led to alternative ideas to explain the redshifts. These interpretations became less and less plausible when the association of AGN and quasars with galaxies in which the emission and absorption line redshifts coincide and when absorption lines due to intervening matter also at cosmological distance were discovered. It became more and more evident then that the energy released by matter when it is accreted into the very deep potential well associated with supermassive black holes is at the origin of the QSO phenomena. This idea had first been suggested by (Zel'dovich & Novikov 1964) and (Salpeter 1964) and has since then become the paradigm in which the theory of Active Galactic Nuclei (AGN) develops.