If there is one subject for which the brightness of 3C 273 is a problem rather than a help it is the study of the underlying galaxy. Subtraction of the quasar light distribution on images has proved a very delicate issue and has been possible only after the introduction of CCD cameras. [Kristian 1973] followed the idea that the quasar phenomenon is similar in nature to that observed in the nuclei of Seyfert galaxies but considerably more powerful and studied whether galaxies can be seen around at least some quasars in direct photographic plates. He concluded that in some cases galaxies are seen while for the brighter objects (including 3C 273) this is not the case, a conclusion expected from the quality of the data.
First (to my knowledge) CCD images of 3C 273 are presented by [Tyson Baum & Kreidl 1982]. These authors conclude that there is an underlying galaxy and that it is similar to the giant elliptical galaxy NGC 4889 in the Coma cluster. The integrated V magnitude of the nebulosity is about 16.05, 3 magnitudes less than the quasar. Based on some previous data on the field of 3C 273 by [Stockton 1980] they also concluded that the 3C 273 host galaxy may be a member of a poor cluster of galaxies.
Near infrared data may be of interest to study the galaxies around quasars as the ratio of galaxy to QSO fluxes may be in general larger than at visual wavelengths (quasars are blue objects). [Veron-Cetty & Woltjer 1990] show thus that for luminous quasars there is no correletion between the luminosity of the galaxy and that of the QSO. [McLeod & Rieke 1994] present a sample of high luminosity quasars including 3C 273. They show that there is in their sample some correlation between the QSO luminosity and that of the underlying galaxies, 3C 273 having the most luminous underlying galaxy of their sample and the smallest fraction of galaxy to (QSO+galaxy) luminosity. The correlation shows, however, a large dispersion. Using far infrared luminosities, [McLeod & Rieke 1994] also investigate the possible existence of starburst activity in the galaxies surrounding quasars. They (preliminarily) conclude that this may be excluded in several of the galaxies, not in all and in particular not in 3C 273.
HST data have recently been published by [Bahcall et al. 1997] for a sample of nearby luminous quasars. The main advantage of the (repaired) HST is, expectedly, the sharpness of its point spread function that allows a cleaner subtraction of the quasar light from the image. This still requires, however, a good knowledge of the point spread function at some distance from the central peak, where the galaxies contribute most of their light. It is interesting to see that the HST data is in contradiction with the popular idea that radio loud quasars lie in bright elliptical galaxies while the radio quiet quasars, like Seyfert nuclei, lie in spiral galaxies. We can conclude from this that it is not the obvious shape of the galaxy that determines the characteristics of the quasar.
In the case of 3C 273, the inferences drawn from the HST data confirm those obtained from the ground. The galaxy may be classified as an E4 galaxy, its luminosity is roughly 3 magnitudes fainter than the quasar in the visible. This indicates that the galaxy is somewhat brighter than the most luminous galaxy of a rich cluster. There is no conspicuous companion or signs of recent violent interaction with another galaxy. The galaxy major axis is about 30", corresponding to some 100kpc with the cosmological parameters used here.
Whereas the study of a single object cannot have universal value, it is still interesting to note that 3C 273, the brightest nearby quasar, is not imbedded in a distorted or peculiar galaxy. This shows that while galaxy interactions may play an important role in bringing material from the galaxy to its nucleus, some other mechanisms must also be at work and even have the dominant role in at least some very bright cases.