SOFT-X-RAY OBSERVATIONS OF A COMPLETE SAMPLE OF X-RAY-SELECTED BL-LACERTAE OBJECTS

We present the results of ROSAT PSPC observations of the X-ray selecte d BL Lacertae objects (XBLs) in the complete Einstein Extended Medium Sensitivity Survey (EMMS) sample. None of the objects is resolved in t heir respective PSPC images, but all are easily detected, All BL Lac o bjects in this sample are well-fitted by single power laws. Their X-ra y spectra exhibit a variety of spectral slopes, with best-fit energy p ower-law spectral indices between alpha = 0.5-2.3. The PSPC spectra of this sample are slightly steeper than those typical of flat ratio-spe ctrum quasars. Because almost all of the individual PSPC spectral indi ces are equal to or slightly steeper than the overall optical to X-ray spectral indices for these same objects, we infer that BL Lac soft X- ray continua are dominated by steep-spectrum synchrotron radiation fro m a broad X-ray jet, rather than flat-spectrum inverse Compton radiati on linked to the narrower radio/millimeter jet. The softness of the X- ray spectra of these XBLs revives the possibility proposed by Guilbert , Fabian, & McCray (1983) that BL Lac objects are lineless because the circumnuclear gas cannot be heated sufficiently to permit two stable gas phases, the cooler of which would comprise the broad emission-line clouds. Because unified schemes predict that hard self-Compton radiat ion is beamed only into a small solid angle in BL Lac objects, the ste ep-spectrum synchrotron tail controls the temperature of the circumnuc lear gas at r less than or equal to 10(18) cm and prevents broad-line cloud formation. We use these new ROSAT data to recalculate the X-ray luminosity function and cosmological evolution of the complete EMSS sa mple by determining accurate K-corrections for the sample and estimati ng the effects of variability and the possibility of incompleteness in the sample. Our analysis confirms that XBLs are evolving ''negatively ,'' opposite in sense to quasars, with [V-e/V-a] = 0.331 +/- 0.060. Th e statistically significant difference between the [V-e/V-a] values fo r X-ray and radio-selected BL Lac objects remains a difficulty for mod els which unify these two types of objects. We have identified one add ition to the sample, so that the sample now has 23 objects. We find no evidence for a substantial number of unidentified low-luminosity BL L ac objects hidden in our sample, as had been suggested by Browne & Mar cha (1993) although a few such objects may be present.