The detection of a few BL Lacs at energies 100 MeV by the Energetic Gamma Ray Experiment Telescope (EGRET) on the Compton Gamma Ray Observatory (CGRO; Gehrels, Chipman & Kniffen 1993) has been reported recently (Fichtel et al. 1994). So far, only previously known BL Lac objects have been associated with -ray sources, five at a high confidence level (> 5 ), four (including the uncertain BL Lac PKS 0521-365) at a lower (between 4) and 5) confidence level, and one (PKS 2155-304) for which the confidence level is not yet available (Vestrand, Stacy & Sreekumar 1995). (One of the five sources detected with a high confidence level, S4 0954+658, has been detected only during phase 2 of the EGRET observations [Mukherjee et al. 1995] and therefore is not included in the EGRET phase 1 catalogue of Fichtel et al. 1994. This also applies to PKS 2155-304). A cross-correlation of our BL Lac list with the EGRET catalogue shows that some high-latitude marginal detections, still unidentified, might be associated with BL Lacs (some of these are also described in the notes to table 11B in Fichtel et al. 1994). These include: 3C 66A, which is 52 arcmin away from GRO J0222+42 (with which it had been previously identified), which has a 95 per cent error radius of 47 arcmin; MS1312.1-422, only 17 arcmin away from GRO J1314-42, with an error radius of 71 arcmin. This is classified as a ``possible'' identification by Fichtel et al. (1994), probably because its relatively low radio flux (18.5 mJy at 5 GHz) would make it the extragalactic object with the largest -ray-to-radio flux ratio (all the other sources, in fact, have 5-GHz radio fluxes typically larger than 1 Jy). We note, however, that within the 95 per cent error radius of GRO J1314-42 there are about 10 unclassified radio sources with fr 50 mJy in the Parkes-MIT-NRAO (PMN) Southern survey (Wright et al. 1994), three of which have fr > 100 mJy. It is therefore likely that the counterpart of the -ray detection is one of the brighter PMN sources. PKS 2032+107 is 89 arcmin away from GRO J2039+11, with an error radius of 66 arcmin, while PKS 2029+121 is somewhat more distant, at 110 arcmin. Both objects have 5-GHz radio fluxes around 1 Jy. PKS 2149+173 is 86 arcmin away from GRO J2157+18, with an error radius of 47 arcmin, while the flat-spectrum radio quasar PKS 2201+171 is offset by 99 arcmin. Again, both objects have 5-Ghz radio fluxes around 1 Jy. The BL Lac candidate 1ES1745+504, with a 5-GHz radio flux as small as 1.3 mJy, was found to be 62 arcmin away from GRO J1742+49, with an error radius of 93 arcmin. However, a much more likely identification is that of the flat-spectrum radio quasar S4 1738+499, offset from the -ray source by 33 arcmin and with a 5-GHz radio flux of 0.6 Jy. Finally, a few BL Lacs are relatively near, although still outside the 95 per cent error circle, to -ray sources mostly identified with radio quasars. In some of these cases part of the -ray emission might be due to the BL Lac.
2.5 Miscellaneous objects
Twenty-four objects have been classified as BL Lacs both by Véron-Cetty & Véron (1993a) and by Hewitt & Burbidge (1993) but do not belong to any BL Lac sample. The reality of their classification has been recently confirmed by Véron-Cetty & Véron (1993b) for six of them: PKS 0047+023, PKS 0301-243, PKS 0808+019, PKS 1604+159, PKS 1717+177 and PKS 2254-204. We have no reason to suspect that the remaining objects would not satisfy the criteria adopted in the definition of most BL Lac samples.
Five sources (PKS 0406+121, PKS 0422+004, PKS 0754+100, MC2 1307+12 and PKS 1413+135) have been reported as having radio fluxes at 5 GHz larger than 1 Jy and have |b| > 10° so they should in principle belong to the 1-Jy catalogue (and therefore to the 1-Jy sample). The fact that they do not shows that variability can have an effect even on radio samples.
2.6 Uncertain and candidate BL Lacs
Fifty objects are listed as uncertain or candidate BL Lacs. This list is quite heterogeneous. It contains BL Lac candidates from various samples and found in the literature. It also includes sources belonging to the compilation of Véron-Cetty & Véron (1993a) or of Hewitt & Burbidge (1993) but not to both. Note that the V magnitudes reported by Véron-Cetty & Véron (1993a) are actually B magnitudes if no B - V colour is given. In those cases, which include also a few miscellaneous objects, we estimate the V magnitude assuming B - V = 0.6, the mean value for BL Lacs.