Completeness of the PG survey

In principle the PG survey selected all objects with U - B < -0.46 and brighter than $B \sim$ 16.2; however, the U - B colour was measured with a relatively large error (0.24$^{\rm m}$ rms) which induced an incompleteness estimated at around 12%. Moreover, in the interval 0.6 < z < 0.8, the strong MgII $\lambda$2800 emission line is in the B filter which gives a much redder U - B colour than for neighbouring redshifts, resulting into an incompleteness of 28% in this range (Schmidt & Green 1983).

The catalogue of mean UBV data on stars (Mermilliod & Mermilliod 1994) contains 283 stars in the magnitude range 12.0 < B < 16.5 and with U - B < -0.40 in the full 10714 deg² area of the PG survey; 190 are included in the PG catalogue. Twenty-four stars photoelectrically observed because they were in the PG catalogue have been ignored; the overall completeness of the PG survey is therefore 64% (166/259). 67% (162/241) of the stars brighter than B = 16.2 were found in the PG survey, while only 22% (4/18) of the weaker stars were. The completeness of the PG survey rises from about 55% at U - B = -0.60 to 80% at U - B < -1.0. For PG QSOs, the completeness should not be less than $\sim$ 70%.

An independent estimate of the completeness of the PG survey comes from the comparison with the RASS-BSC. Of the 25 PG QSOs in the FBS area, 13 are brighter than B = 16.16 and have been detected by ROSAT. Twelve of them have been found in the FBS. All FBS objects associated with a ROSAT source, located in the PG area, have been spectroscopically identified; among them, there are eleven non-PG QSOs. No more than two are suspected to have been possibly bright enough to have been detected by the PG survey. This suggests a completeness of about 87% (13/15) with however a large uncertainty due to the smallness of the sample.

Goldschmidt et al. (1992), Köhler et al.(1997) and La Franca & Cristiani (1997) have claimed that the PG survey is incomplete by a factor of 2 to 3, but their samples are quite small (5, 8 and 7 objects, respectively); in addition, we do not know if there is an offset between their magnitude scales and that of the PG survey (the mean differences between the PG and photoelectric B magnitudes for 190 stars is equal to -0.02$^{\rm m}$). Therefore, their results should be considered as tentative.