The radio band is where the class of BL Lacertae objects was discovered. For many years the great majority of the BL Lacs known were found among the sources detected in large radio surveys. As a consequence, all the classical and well-studied objects have been discovered at these frequencies.
2.1.1 The 1-Jy sample
This is currently the largest complete radio sample of BL Lacs. Described in Stickel et al. (1991) and Stickel, Fried & Kühr (1993), it includes 34 objects extracted from the 1-Jy catalogue (radio flux fr 1 Jy at 5 GHz: Kühr et al. 1981a), a flux-limited catalogue which covers essentially the whole sky excluding the galactic plane (|b| < 10°) and the Magellanic Clouds, according to the following criteria: (1) flat radio spectrum between 2.7 and 5 GHz ( 0.5, f -); (2) magnitude brighter than 20; (3) emission lines in the optical spectrum absent or weak with a rest-frame equivalent width of the strongest line < 5 Å. Three more objects are included in the updated version of the 1-Jy catalogue (Stickel, Meisenheimer & Kühr 1994). Two of these have V 20, while the third one (PKS 2149+173) has V = 18.9, although during the period of the spectroscopic observations it was below the 20th magnitude limit (Stickel & Kühr 1993a). PKS 0521-365, an object classified as a BL Lac in the literature but originally excluded from the 1-Jy sample because some of its lines had W > 5 Å, has been included as an uncertain BL Lac, since its [O III] luminosity is more typical of BL Lacs than of quasars (see discussion in Urry & Padovani 1995).
Brunner et al. (1994) have reported on a ROSAT observation of S5 0454+844 (which belongs both to the 1-Jy and S5 samples), which shows the BL Lac object to be only 48 arcsec away from a source about five times brighter. They therefore suggest that previous reports of X-ray emission from this object (which are referred to in Table 1) could be due to a misidentification.
It has been suggested (Perlman et al. 1995a) that the requirement on the radio spectral index (r 0.5), imposed as a criterion for the selection of the 1-Jy BL Lacs to exclude the bulk of the radio galaxies, might have resulted in the loss of some objects, since BL Lacs with steeper radio spectral indices are known. To estimate the magnitude of this effect, we performed the following simple calculation: out of the 119 confirmed BL Lacs which, to our knowledge, have 2.7 - 5 GHz spectral index information, only 8 have r > 0.5. Out of these, one (S5 1749+701) was included nevertheless in the 1-Jy sample because the steep radio spectrum was clearly due to variability and non-simultaneous measurements (Stickel et al. 1991). As regards the remaining ones, variability cannot be the explanation for the steepness of r in the case of PKS 0548-322, MS1207.9+3945 and MS1402.3+0416 since the multifrequency radio measurements were contemporaneous (Stocke et al. 1985), while it cannot be excluded in the remaining four objects, ON 231, MS1407.9+5954, RXJ00079+4711 and RXJ16442+4546. We then estimate that, out of the 85 BL Lacs not included in the 1-Jy, S4 and S5 samples (which were selected to have flat radio spectra), between 3 (4 per cent) and 7 (8 per cent) have 2.7-5 GHz spectral indices steeper than 0.5. This suggests that the condition requiring a flat radio spectrum might cause the loss of only 1 to 3 BL Lacs in the 1-Jy sample.
2.1.2 The S4 sample
The S4 sample includes 14 objects extracted from the S4 catalogue (fr 0.5 Jy at 5 GHz, 35° 70° and |b| 10°: Pauliny-Toth et al. 1978; Stickel & Kühr 1994) by Stickel & Kühr (1994) using the same criteria as those applied to 1-Jy BL Lacs. Note that S4 1652+398 (Mrk 501) and S4 1823+568 have been mistakenly classified as a normal galaxy and a QSO respectively in Stickel & Kühr (1994): both are in fact confirmed 1-Jy BL Lacs. About 10 per cent of the S4 sources are still classified as empty fields so a small number of BL Lacs could still be unidentified.
2.1.3 The S5 sample
The S5 sample includes 13 objects extracted from the S5 catalogue (fr 0.25 Jy at 5 GHz, 70° and |b| 10°: Kühr et al. 1981b) by Kühr & Schmidt (1990). The selection criteria are slightly different from those adopted for the 1-Jy and S4 samples: they include in fact maximum optical polarization Pmax larger than 3 per cent on at least one occasion, while it is not clear what is the equivalent width limit adopted to separate BL Lacs from quasars. (Note that all but two 1-Jy BL Lacs have Pmax > 3 per cent [Stickel et al. 1994], although this was not one of the selection criteria.) We have excluded S5 1053+81 from the sample because its spectrum shows emission lines (Xu et al. 1994), and we have added a BL Lac candidate, S5 2353+81 (Stickel & Kühr 1993b). The S5 catalogue is currently being updated by Stickel & Kühr (in preparation).
We note that the dynamical range of source flux in radio surveys is small and of order 10. This demonstrates that the study of BL Lacs is still at a very early stage even in the part of the electromagnetic spectrum where these objects were first discovered more than 25 years ago.