Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://hea-www.harvard.edu/~pgreen/Papers/precos0.pdf Äàòà èçìåíåíèÿ: Tue Dec 10 02:12:26 2002 Äàòà èíäåêñèðîâàíèÿ: Tue Oct 2 00:47:54 2012 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: ngc 1068

The Astrophysical Journal Supplement Series, 143: 257­276, 2002 December
# 2002. The American Astronomical Society. All rights reserved. Printed in U.S.A.

E

EMISSION LINE PROPERTIES OF ACTIVE GALACTIC NUCLEI FROM A PRE-COSTAR FAINT OBJECT SPECTROGRAPH HUBBLE SPACE TELESCOPE SPECTRAL ATLAS Joanna K. Kuraszkiewicz and Paul J. Green
Harvard-Smithsonian Center for Astrophysics, 60 Garden Street, Cambridge, MA 02138; jkuraszkiewicz@cfa.harvard.edu, pgreen@cfa.harvard.edu

Karl Forster
California Institute of Technology, 1200 East California Boulevard, MC 405-47, Pasadena, CA 91125; krl@srl.caltech.edu

Thomas L. Aldcroft and Ian N. Evans
Harvard-Smithsonian Center for Astrophysics, 60 Garden Street, Cambridge, MA 02138; taldcroft@cfa.harvard.edu, ievans@cfa.harvard.edu

and Anuradha Koratkar
Space Telescope Science Institute, 3700 San Martin Drive, Baltimore, MD 21218; koratkar@stsci.edu Received 2002 April 20; accepted 2002 July 11

ABSTRACT UV/optical emission lines offer some of the most detailed information obtainable about the intrinsic properties of quasars. Studies of the density, ionization and metal abundance of gas near the accreting black hole are probed through an intriguing but poorly understood complex of correlations between emission lines and overall quasar spectral energy distributions that has long suffered from a lack of large, consistently measured samples. As part of a broader effort to expand and systematize the data upon which these studies are built, we present measurements of the UV/optical emission line parameters in a sample of 158 active galactic nuclei observed with the Faint Object Spectrograph on the Hubble Space Telescope (HST ), prior to the installation of COSTAR. We use an automated technique that accounts for galactic reddening, includes iron emission blends, galactic and intrinsic absorption lines, and performs multicomponent fits to the emission line profiles. We present measured line parameters (equivalent width and FWHM) for a large number (28) of different UV/optical lines, including upper limits for undetected lines. We also study the relations between the emission line equivalent widths and luminosity (the Baldwin effect), as well as redshift (evolution). We compare results from this HST FOS sample with our previous measurements of 993 QSOs in the Large Bright Quasar Survey using the same analysis technique and sum the samples to achieve better coverage of the luminosity-redshift plane. We confirm a significant Baldwin effect for UV iron emission from Green et al. and find that evolution dominates the effect for iron and for Si iv emission. The values of the Baldwin effect slopes for all UV emission lines and the dependence of the slopes on the sample's luminosity range point to a change of the SED as the cause of the Baldwin effect in the FOS sample. Subject headings: galaxies: active -- quasars: emission lines -- quasars: general -- ultraviolet: galaxies On-line material: color figure, machine-readable tables
1. INTRODUCTION

The strong, broad emission lines observed in the spectra of active galactic nuclei (AGNs) are believed to form in a large number of small gas clouds photoionized by the central UV/X-ray continuum source (presumably a black hole with an accretion disk). Although it has been known from the late 1970s (see Netzer & Davidson 1979 for review) that photoionization models successfully predict the strengths of the strong UV lines, unresolved issues concerning the origin, evolution, geometry, and the acceleration mechanism of the emission line clouds persist (Krolik & Kallman 1988; Alexander & Netzer 1994; Perry & Dyson 1985; Murray & Chiang 1998; Koratkar & Gaskell 1991). In hope of solving these issues, studies of correlations among the emission lines and their relation to the continuum have been conducted. Among the most extensively examined is the anticorrelation between emission line equivalent width and luminosity called the Baldwin effect. Another intriguing set of correlations occurs between Fe ii 4570, [O iii] 5007 strength, and the FWHM and blue asymmetry of the H line. Principal component analysis of measurable QSO parameters yields 257

the principal axis of parameter variation called Eigenvector 1 (PC1; Boroson & Green 1992; Francis et al. 1992). Strong projections onto PC1 include UV spectral properties such as C iii] width, Si iii]/C iii] ratio, C iv and N v strength (Wills et al. 1999), C iv shift/asymmetry (Marziani et al. 1996), and also soft X-ray continuum properties (Boroson & Green 1992; Laor et al. 1994, 1997). Most studies of emission line properties to date have been conducted using large samples of nonuniform data sets compiled using emission line measurements taken from the literature (Zheng & Malkan 1993; Zamorani et al. 1992; Baldwin, Wampler, & Gaskell 1989). Such compilations often include spectra that differ significantly in quality and resolution, with emission lines measured using a variety of techniques. Most inconsistencies in the emission line measurements arise from differences in continuum placement and difficulty in estimating the amount of blended iron emission. Studies that have concentrated on uniform and consistent emission line measurements have tended to span small samples and ranges of QSO parameters (Boroson & Green 1992; Wills et al. 1999; Wilkes et al. 1999).


258

KURASZKIEWICZ ET AL. 3. Corrects the data for time-dependent detector sensitivity degradation; 4. Scales the data to the white dwarf model G191-B2B for photometric reference. At this stage the sample consisted of 933 data sets and 263 objects, among which 112 targets had observations with only one grating. We included spectra taken with both highresolution (G130H, G190H, G270H, G400H, G570H, G780H) and low-resolution (G160L, G650L) gratings. Those obtained with prism grating were excluded from the analysis, due to extremely low resolution that precludes from any reasonable emission line measurements. We have also excluded the broad absorption line (BAL) QSOs, where strong absorption features heavily disrupt the emission lines, and BL Lac objects, which spectra show no emission lines. Off-nuclear spectra were not analyzed. To obtain a more reliable continuum fit for each object, spectra from different observations and gratings were merged together whenever the flux levels in the overlap region differed by less than 10%. When the difference in flux was larger, spectra were analyzed separately. Merged spectra were rebinned in the overlap region, in such a way that the resolution varied linearly from that of the blue disperser at the blue end of the overlap region, to that of the red disperser at the red end of the overlap region. The flux level in each wavelength bin was calculated as a weighted mean of fluxes from the two overlapping spectra at that bin. The weights changed linearly with the bin position in the overlap region, with the blue spectrum having a 100% weight and the red spectrum having 0% weight at the blue end of the overlap region, with the reverse true at the red end (for details, see I. N. Evans & A. P. Koratkar 2002, in preparation). Data obtained at different epochs were merged together in a slightly different way by rebinning the spectra in the overlap region to the lower dispersion of the two. If one of the merged spectra was particularly noisy in the overlap region, it was ignored in that region, and only the higher signal-to-noise spectrum was adopted. Finally, we removed from the sample spectra with no emission lines, whether due to low signal-to-noise or a redshift that placed strong emission lines outside the wavelength range.

With the prospect of an increasing number of quasar spectra becoming available in the next few years (30,000 quasars from the 2dF and 100,000 from the SDSS surveys), there is hope that large numbers of emission line measurements will provide enough data for reliable statistical analysis, leading to an enhanced understanding of the differences in quasar emission line properties and their cause (orientation, black hole mass, accretion rate, evolution?). This goal is unlikely to be realized without an automated procedure to facilitate consistent measurements of the emission line properties from such large samples. We have undertaken a major study of AGN spectra using a measuring technique which is largely automated. This procedure accounts for blended optical and UV iron emission, finds absorption lines and models them together with the emission lines, and provides upper limits for undetected lines. With this automated technique, we were able to consistently measure the largest single sample of QSO spectra available at the time, the Large Bright Quasar Survey (LBQS), for which 993 spectra have been modeled by Forster et al. (2001, hereafter Paper I) with the emission line properties analyzed in Green, Forster, & Kuraszkiewicz (2001, hereafter Paper II). The LBQS is a homogeneous and ° complete sample, with fairly low resolution (6­10 A) and S/N (a median of $ 5 averaged over the entire spectrum). The tight luminosity-redshift correlation in such a uniform magnitude-limited sample actually hampers disentanglement of luminosity and evolution effects (Paper II). To broaden the power of statistical analyses of quasar properties, we have analyzed the higher S/N and resolution spectra in the archive of AGNs observed with the Faint Object Spectrograph (FOS; Keyes et al. 1995 and references therein) on board of the Hubble Space Telescope (HST ). The sample is heterogeneous and extends toward lower redshifts and luminosities than the LBQS sample. In this paper, we analyze AGNs observed prior to the installation of COSTAR and present a catalog of these spectra and their emission line measurements. Spectra observed with COSTAR/FOS will be analyzed in a future paper (J. K. Kuraszkiewicz et al. 2002, in preparation).

2. THE SAMPLE

We have collected all available (UV and optical) spectrophotometric archival data for AGNs that have been observed with the FOS/HST prior to the installation of COSTAR in 1993 December. All spectra have been uniformly calibrated to account for temporal, wavelength- and aperture-dependent variations that are seen in the instrumental response, and combined into a preliminary preCOSTAR Spectral Atlas (see I. N. Evans & A. P. Koratkar 2002, in preparation, for the more recent version of the Atlas, where the most up-to-date calibrating techniques were used). The average inverse sensitivity (AIS) calibration (adopted in 1996 March) was used, generated by a spline fit to the inverse sensitivities derived from an average of many observations of a number of standard stellar spectra. The AIS method for flux calibrating FOS data: 1. Normalizes count data from all apertures to the 4 > 3 aperture; 2. Corrects wavelength dependent aperture throughput to account for changes in aperture throughput as a function of the optical telescope assembly focus;

3. THE CATALOG

The final list of objects includes 158 AGNs and 174 spec´ tra and is presented in Table 1. To avoid a melange of AGN names, including different positional and catalog designations, we use a coordinate designation based on the equinox J2000 position (in standard IAU format consisting of HHMM+/þDDMM). In addition to this, a two letter designation is used for the spectra indicating that the spectra are from a pre-COSTAR observation (r) and whether there is more than one spectrum of the same object (a­z), e.g., NGC 3031, which has a J2000 position designation of 0955+6903 has three pre-COSTAR spectra 0955+6903ra, 0955+6903rb, and 0955+6903rc (which have not been merged due to differing flux levels). A capital letter at the end of the name indicates that the object is a gravitational lens and that separate spectra of each lensed component were observed and analyzed, e.g., 1230+1223rA, 1230+1223rB.


TABLE 1 List of Objects and Spectra Designation 0005+1609 0010+1058 0027+2241 0047+0319 0053+1241 0054+2525 0103+0221 0113+2958 0115þ0127 0118+0258 0123þ5848 0125þ0005 0136+2057 0242þ0000 0256þ0126 0319+4130 0323þ4930 0336þ3607 0351þ1429 0352þ0711 0357þ4812 0405þ1308 0407þ1211 0417þ0553 0420þ5456 0441þ4313 0456þ2159 0506þ6109 0615+7102 0630+6905 0635þ7516 0713+1146 0743þ6726 0744+3753 0745+3142 0755+2542 0837+4450 0840+1312 0847+3445 0853+4349 0859+4637 0902þ1415 0906+1646 0909+4253 0919+5106 0927+3902 0955+6903 0956+4115 0957+5522 0958+3224 1003+6813 1004+2855 1010+4132 1011+1304 1013+3551 1024+1912 1041+0610 1042+1203 1048þ2509 1051þ0051 1058+1951 1104+7658 1106þ0052 1107+1628 1114+4037 ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ Name PKS 0003+15 III Zw 2 NAB 0024+22 PKS 0044+030 I Zw 1 0052+2509 0100+0205 B2 0110+29 PKS 0112þ017 0115+027 Fairall 9 PKS 0122þ00 3C 47 NGC 1068 0253þ0138 NGC 1275 RX J03233þ4931 0334þ3617 3C 95 0350þ0719 PKS 0355þ48 PKS 0403þ13 PKS 0405þ12 PKS 0414þ06 NGC 1566 PKS 0439þ433 PKS 0454þ22 0506þ6113 Mrk 3 HS 0624+6907 PKS 0637þ75 3C 175 PKS 0743þ67 3C 186 B2 0742+31 OI 287 WSTB 55W 037 3C 207 Ton 951 US 1867 0856+4649 PKS 0859þ14 3C 215 3C 216 NGC 2841 UB3 B2 0923+392 NGC 3031 PG 0953+414 4C 55.17 0955+326 0959+68W1 Ton 28 4C 41.21 PG 1008+133 CSO 251 4C 19.34 4C 06.41 3C 245.0 NGC 3393 PG 1049þ005 PKS 1055+20 3C 249.1 PKS 1103þ006 MC 1104+167 3C 254 Type
a

Redshift 0.450 0.089 1.119 0.623 0.061 0.155 0.393 0.363 1.365 0.672 0.047 1.070 0.425 0.004 0.879 0.018 0.071 1.100 0.616 0.962 1.005 0.571 0.573 0.775 0.005 0.593 0.533 1.093 0.014 0.370 0.653 0.770 1.510 1.063 0.461 0.446 0.208 0.681 0.064 0.514 0.924 1.327 0.412 0.670 0.556 0.695 þ0.0001 0.234 0.909 0.531 0.773 0.330 0.612 1.287 0.070 0.828 1.270 1.029 0.012 0.360 1.110 0.312 0.423 0.632 0.734

NHb 3.87 5.90 3.73 2.81 5.07 4.38 2.40 6.04 5.37 3.51 3.19 3.25 5.12 2.95 5.60 13.93 1.73 1.40 4.20 5.78 1.16 4.24 3.66 4.34 1.35 1.30 2.77 2.53 8.76 6.38 9.22 11.51 11.91 4.87 5.12 4.54 3.00 5.40 3.32 2.62 2.30 5.92 3.75 1.39 1.29 1.61 4.34 1.12 0.88 1.77 3.75 1.85 1.13 3.72 1.24 2.42 2.82 2.70 5.89 4.21 1.89 2.87 4.20 1.42 1.75 0005+1609ra 0010+1058ra 0027+2241ra 0047+0319ra 0053+1241ra 0054+2525ra 0103+0221ra 0113+2958ra 0115þ0127ra 0118+0258ra 0123þ5848ra 0125þ0005ra 0136+2057ra 0242þ0000ra, 0256þ0126ra 0319+4130ra 0323þ4930ra 0336þ3607ra 0351þ1429ra 0352þ0711ra 0357þ4812ra 0405þ1308ra 0407þ1211ra 0417þ0553ra 0420þ5456ra 0441þ4313ra 0456þ2159ra 0506þ6109ra 0615+7102ra 0630+6905ra, 0635þ7516ra 0713+1146ra 0743þ6726ra 0744+3753ra 0745+3142ra 0755+2542ra 0837+4450ra 0840+1312ra 0847+3445ra 0853+4349ra 0859+4637ra 0902þ1415ra 0906+1646ra 0909+4253ra 0919+5106ra 0927+3902ra, 0955+6903ra, 0956+4115ra 0957+5522ra 0958+3224ra 1003+6813ra 1004+2855ra 1010+4132ra 1011+1304ra 1013+3551ra 1024+1912ra 1041+0610ra 1042+1203ra 1048þ2509ra 1051þ0051ra 1058+1951ra 1104+7658ra 1106þ0052ra 1107+1628ra 1114+4037ra

Spectra

Q S1 Q Q Q S1 Q Q Q Q S1 Q Q S2 Q S2 Q Q Q Q Q Q Q Q S1 Q Q Q S2 Q Q Q Q Q Q Q S2 Q Q Q Q Q Q Q Q Q S1.8 Q Q Q Q Q Q Q Q Q Q Q S2 Q Q Q Q S1 Q

0242þ0000rb

0630+6905rb

0927+3902rb 0955+6903rb, 0955+6903rc


TABLE 1--Continued Designation 1119+2119 1125+5910 1130þ1449 1132+1023 1139+6547 1139þ1350 1139þ3744 1151+5437 1158+6254 1204+2754 1208+4540 1210+0954 1210+3924 1214+1403 1219+0638 1221+7518 1225+3332 1229+0203 1230+1223 1231þ0224 1233+0931 1244+1721 1247+3209 1252+5634 1254+1141 1256þ0547 1259+3423 1301+5902 1305þ1033 1308+3005 1309+0819 1319+2728 1319+5148 1323+2910 1331+3030 1336+1725 1341+4123 1342+6021 1343+2844 1349+5341 1354+0052 1357+1919 1358+5752 1359þ4152 1404+0937 1405+2555 1409+2618 1417+2508 1417+4456 1418+1703 1427+1949 1427þ1203 1436+6336 1437þ0147 1442+3526 1446+4035 1454þ3747 1514+3650 1524+0958 1539+4735 1547+2052 1557+3304 1613+3412 1614+2604 1620+1736 1633+3924 ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ Name PG 1116+215 1123+5926 PKS 1127þ14 1130+106Y 3C 263 PKS 1136þ13 NGC 3783 PG 1148+549 1156+6311 GQ Comae PG 1206+459 1208+101A, B NGC 4151 PG 1211+143 PG 1216+069 Mrk 205 NGC 4395 PG 1226+023 NGC 4486A, B PKS 1229þ02 Q1230+0947 PG 1241+176 B2 1244+32B 3C 277.1 PKS 1252+119 3C 279 B201 1257+346 PG 1259+593 PKS 1302þ102 1306+3021 1307+0835 Ton 153 1317+5203 Ton 157 3C 286.0 PG 1333+176 PG 1338+416 3C 288.1 B2 1340+29 4C 53.28 PG 1352+011 PKS 1354+19 4C 58.29 1355þ4138 1401+0951 PG 1402+261 PG 1407+265 NGC 5548 PG 1415+451 MC 1415+172 1425+2003 PKS 1424þ11 S4þ1435+63 Q1435þ0134 Mrk 478 PG 1444+407 PKS 1451þ375 B2 1512+37 PG 1522+101 PG 1538+477 3CR 323.1 B2 1555+33 DA 406 1612+2611 3C 334 1631+3930 Type
a

Redshift 0.177 0.858 1.187 0.504 0.646 0.560 0.009 0.969 0.594 0.165 1.158 3.822 0.003 0.081 0.331 0.071 0.001 0.158 0.004 1.045 0.415 1.273 0.949 0.321 0.871 0.536 1.375 0.462 0.278 0.806 0.155 1.022 1.060 0.960 0.849 0.553 1.219 0.961 0.905 0.980 1.117 0.720 1.371 0.313 0.441 0.164 0.940 0.017 0.114 0.821 0.113 0.806 2.068 1.310 0.079 0.267 0.314 0.371 1.321 0.772 0.264 0.942 1.401 0.131 0.555 1.023

NHb 1.40 0.87 3.83 2.88 0.91 3.55 9.59 0.96 1.67 1.76 1.11 1.59 2.17 2.76 1.56 2.74 1.43 1.68 2.51 2.34 1.57 1.93 1.27 1.03 2.55 2.22 1.13 1.54 3.26 1.04 2.13 1.27 1.26 1.17 1.14 1.71 0.76 2.09 1.19 1.19 2.07 2.90 1.29 5.57 1.95 1.40 1.36 1.61 0.96 1.63 2.46 7.88 1.57 3.66 0.96 1.08 6.24 1.40 2.67 1.62 4.04 2.57 1.44 3.77 4.07 0.92

Spectra 1119+2119ra 1125+5910ra 1130þ1449ra 1132+1023ra 1139+6547ra 1139þ1350ra 1139þ3744ra 1151+5437ra 1158+6254ra 1204+2754ra 1208+4540ra 1210+0954rA, 1210+0954rB 1210+3924ra, 1210+3924rb, 1210+3924rc 1214+1403ra 1219+0638ra 1221+7518ra, 1221+7518rb 1225+3332ra 1229+0203ra 1230+1223rA, 1230+1223rB 1231þ0224ra 1233+0931ra 1244+1721ra 1247+3209ra 1252+5634ra 1254+1141ra 1256þ0547ra 1259+3423ra 1301+5902ra 1305þ1033ra, 1305þ1033rb 1308+3005ra 1309+0819ra 1319+2728ra 1319+5148ra 1323+2910ra 1331+3030ra 1336+1725ra 1341+4123ra 1342+6021ra 1343+2844ra 1349+5341ra 1354+0052ra 1357+1919ra 1358+5752ra 1359þ4152ra 1404+0937ra 1405+2555ra 1409+2618ra 1417+2508ra, 1417+2508rb 1417+4456ra 1418+1703ra 1427+1949ra 1427þ1203ra 1436+6336ra 1437þ0147ra 1442+3526ra 1446+4035ra 1454þ3747ra 1514+3650ra 1524+0958ra 1539+4735ra, 1539+4735rb 1547+2052ra 1557+3304ra 1613+3412ra 1614+2604ra 1620+1736ra, 1620+1736rb 1633+3924ra

Q Q Q Q Q Q S1 Q Q Q Q Q S1 Q Q S1 S1.8 Q Q Q Q Q Q Q Q Q Q Q Q Q S1 Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q S1 Q Q S1 Q Q Q S1 Q Q S1 Q Q Q Q Q S1 Q Q


EMISSION LINE PROPERTIES OF AGNs
TABLE 1--Continued Designation 1634+7031 1638+5720 1642+3948 1658+0515 1704+6044 1719+4804 1821+6420 1927+7358 2044þ1043 2114+0607 2131þ1207 2137þ1432 2143+1743 2148+0657 2203+3145 2218þ0335 2225þ0457 2232+1143 2246þ1206 2253+1608 2254+1136 2303þ6807 2311+1008 2342þ0322 2346+0930 2351þ0109 2355þ3357 ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ Name PG 1634+706 OHIO S 562 3C 345 PKS 1656+053 3C 351 PG 1718+481 E1821+643 4C 73.18 MRK 509 PG 2112+059 PKS 2128þ12 PKS 2135þ147 OHIO X 169 PKS 2145+06 B2 2201+315 A PKS 2216þ03 3C 446 CTA 102 PKS 2243þ123 3C 454.3 4C 11.72 PKS 2300þ683 PG 2308+098 PKS 2340þ036 PKS 2344+09 2349þ0125 PKS 2352þ34 Type Q Q Q Q Q Q S1 Q S1 Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q Q S1 Q
a

261

Redshift 1.334 0.751 0.593 0.879 0.372 1.084 0.297 0.302 0.034 0.466 0.501 0.200 0.211 0.990 0.295 0.901 1.404 1.037 0.630 0.859 0.326 0.512 0.433 0.896 0.677 0.174 0.706

NHb 5.47 1.77 0.89 6.11 2.02 2.11 3.81 7.24 4.44 6.26 4.77 4.45 8.21 4.70 5.00 6.18 5.26 5.05 4.58 6.94 5.42 2.65 3.81 3.55 4.90 3.22 1.06

Spectra 1634+7031ra 1638+5720ra 1642+3948ra 1658+0515ra 1704+6044ra 1719+4804ra 1821+6420ra 1927+7358ra 2044þ1043ra 2114+0607ra 2131þ1207ra 2137þ1432ra 2143+1743ra 2148+0657ra 2203+3145ra 2218þ0335ra, 2218þ0335rc 2225þ0457ra 2232+1143ra 2246þ1206ra 2253+1608ra, 2253+1608rb 2254+1136ra 2303þ6807ra 2311+1008ra 2342þ0322ra 2346+0930ra 2351þ0109ra 2355þ3357ra

Note.--Table 1 is also available in machine-readable form in the electronic edition of the Astrophysical Journal Supplement. a Q, QSO; S1, Seyfert 1; S2, Seyfert 2. b N is in units of 1020 cmþ2. H

4. ANALYSIS OF THE SPECTRA

TABLE 2 Continuum and Iron Fitting Windows ° Rest Frame Wavelength Range (A) Continuum 1140­1150a 1275­1280b 1320­1330 1455­1470 1690­1700 2160­2180 2225­2250 3010­3040c 3240­3270 3790­3810 4210­4230 5080­5100 5600­5630 5970­6000 Iron Emission Lines Nearby Blueward O vi 1035 N v 1243 O i 1305 Redward

The analysis of the spectra follows the procedure described in Paper I, which modeled spectra of the LBQS sample. Small adjustments, however, have been made to accommodate the FOS spectra (see below). The modeling software `` Sherpa '' (Freeman, Doe, & Siemiginowska 2001), developed for the Chandra mission, was used, where the model parameters were determined from a minimization of the 2 statistic with modified calculation of uncertainties in each bin (Gehrels 1986) and using the Powell optimization method. The procedure begins with fitting a power-law continuum to regions of the spectrum redward of Ly that are uncontaminated by emission lines and away from blended iron emission. The continuum windows used are presented in Table 2 and are the same as those in Paper I with one minor exception. In four cases (0351þ1429ra, 0958+3224ra, 1010+4132ra, 1107+1628ra) we added an additional continuum window at the red side of C iii] 1909 to better constrain the continuum. In most cases one power law was sufficient. However, in a small number of cases (see Appendix), where the spectra covered a large wavelength range (both UV and optical), a second power law, extending redward ° of rest ¼ 4200 A was introduced. A Galactic reddening correction was applied to the power-law continua following the prescription of Cardelli, Clayton, & Mathis (1989; see Paper I for details), where the neutral hydrogen column density (NH in units of 1020 cmþ2) used for each object is given in Table 1. The values are mostly taken

Ly1215 O i 1305 Si iv +O iv] 1400 Si iv +O iv] 1400 C iv 1549 He ii 1640 Al iii 1859 2020­2120 2250­2650 2900­3000 C iii] 1909 Mg ii 2800 [O ii] 3728 H4102 [O iii] 4363 [O iii] 5007 He i 5876 Mg ii 2800 [Ne v] 3426 [Ne iii] 3869 H 4340 H4861 He i 5876 [N ii] 6549

4400­4750d 5150­5500

Notes.--In three cases: 0351þ1429ra, 0958+3224ra, and 1010+4132ra, 1107+1628ra an additional continuum window was added redward of ° C iii] at 2000­2020 A rest frame, to obtain a better power-law continuum fit. In the Ly + O vi region a flat `` pseudo '' continuum was fitted to the ° ° following continuum windows: 980­1010 A and 1060­1090 A rest frame. a This window lies on the blue side of the Ly emission line and is only used where no other continuum window is available. b This window is used only when windows at larger wavelength are unavailable. c May have some iron emission contamination. d He ii 4686 lies in this window.


262

KURASZKIEWICZ ET AL.

Vol. 143

from the Bell Laboratory H i survey (Stark et al. 1992). However, for a few AGNs for which NH had been specifically measured, we quote the values from the literature (Lockman & Savage 1995; Elvis, Wilkes, & Lockman 1989), and for objects with declination >40 , NH is from Heiles & Cleary (1979). The next step was to model the blended iron emission, which is particularly strong around Mg ii, H and redward of C iii]. We use the UV template of Vestergaard & Wilkes (2001) covering 1250 G < rest < 3100 G and the optical template from Boroson & Green (1992), covering 4400 G < rest < 7000 G. The iron template spectrum was broadened by convolving with Gaussian functions of widths between 900 km sþ1 and 10,000 km sþ1 and separated by steps of 250 km sþ1, while conserving the total flux in each template. The strengths of the iron emission in UV and optical were measured independently. The first step was to estimate a crude flux normalization amplitude for a 2000 km sþ1 template at wavelengths where iron emission is strongest (see Table 2). Next the FWHM of the template was estimated, and then a fit of both the amplitude and FWHM was made. This was followed by another iteration of the continuum and the iron emission modeling.

The next step was to model the emission lines with multiple Gaussians (free parameters in the fit: height of the emission line peak, FWHM, and peak position). The inventory of emission lines and their components is presented in Table 3. In most cases strong emission lines such as Ly C iv,C iii], Mg ii, and H were modeled using two Gaussian components (a very broad line region component (VBLR) and the intermediate line region (ILR) component (see, e.g., Brotherton et al. 1994) which here we call the broad and narrow components, respectively. However, in a small number of cases (where the S/N was small or two components were clearly an excess) these lines were modeled using a single Gaussian. Due to the better quality of FOS spectra compared to the LBQS, additional line components representing blended lines were added: C iii] 1909 (modeled as C iii] narrow + C iii] broad + Si iii] 1892 + Al iii 1859 in higher S/N spectra, and as C iii] single + Al iii in lower S/N spectra), H (H + [S ii] 4072), and H (H + [O iii] 4363). The continuum, iron, and emission line model obtained this way was then used as an input `` continuum '' to the FINDSL (Aldcroft 1993) routine, which identifies narrow absorption features and fits them with Gaussian profiles.

TABLE 3 Inventory of Emission Lines c ° (A) 1030.0 1215.7 1241.5 1305.0 1400.0 1549.0 1640.0 1859.0 1892.0 1909.0 2800.0 3426.0 3728.0 3869.0 4101.7 4072.5 4340.5 4263.0 4686.5 4861.3 4959.0 5007.0 5875.6 6548.1 6562.8 6583.4 6716.4 6730.8 FWHMa (km sþ1) 5000 7000s 2500n, 8000b 6000 2500 5800 6500s 3000n, 11000b 10000 3500 4000 5500s 1500n, 6000b 4000s 3000n, 8000b 1000 600 900 450 450 3000 1000 1200 4000s 1000n, 5500b 700 600 2000 700 1000n, 6000b 700 700 700 Window ° (A)
b

Emission Line Ly1025.7 + O vi 1035....... Ly1215.7 ............................ N O Si C v 1241.5............................. i 1305................................. iv +O iv] 1400 .................. iv 1549 ............................... 1640 .............................. 1859.............................. 1892 ............................. 1909 ..............................

Notes See Appendix.

He ii Al iii Si iii] C iii]

Mg ii 2800 ............................. [Ne v] 3426 ............................ [O ii] 3728.............................. [Ne iii] 3869 ........................... H4101.7 .............................. [S ii] 4072.5............................ H 4340.5 ............................. [O iii] 4363............................. He ii 4686.5 ........................... H4861.3 ............................. [O iii] 4959............................. [O iii] 5007............................. He i 5875.6 ............................ [N ii] 6548 ............................. H6563 ................................ [N ii] 6583 ............................. [S ii] 6716.4............................ [S ii] 6731 ..............................

1010­1060 1170­1350 1170­1350 1170­1350 1170­1350 1350­1450 1350­1720 1350­1720 1350­1720 1820­1970 1820­1970 1820­1970 1820­1970 2700­2900 2700­2900 3390­3460 3700­3760 3810­3930 4000­4200 4000­4200 4240­4440 4240­4440 4580­4790 4750­5100 4750­5100 4750­5100 4750­5100 5825­5900 6300­6800 6300­6800 6300­6800 6300­6800 6300­6800

See Appendix.

Iron emission may be strong in this window. Iron emission may be strong in this window. See Appendix. See Appendix. See Appendix.

Iron Iron Iron Iron Iron

emission emission emission emission emission

may may may may may

be be be be be

strong strong strong strong strong

in in in in in

these these these these these

windows. windows. windows. windows. windows.

a This is the FWHM used for estimating the upper limits of W in weak emission lines. More than one width is given for emission lines that can have two Gaussian components; s, single component; n, narrow component; b, broad component. b The wavelength range over which the emission lines are modeled.


No. 2, 2002

EMISSION LINE PROPERTIES OF AGNs

263

We excluded from the analysis the Ly forest region ° blueward of rest ¼ 1065 A and also the Balmer contin° uum region (3360­3960 A), where the global power-law continuum may not fit the spectra well. The minimum significance level for identification of absorption lines was set to 4.5 . The absorption line parameters were then used in a next iterative modeling step where the emission

lines and the absorption lines were fitted simultaneously by the Sherpa program. After each of the above steps, the results of the modeling were inspected and adjustments were made to those spectra that were not fitted successfully. This is not surprising as no automated procedure can be fully successful while dealing with the plethora of AGN spectral shapes.

Fig. 1.--Example of spectral modeling for quasar PKS 0003+158. Panel (a) shows the reddened continuum model (fitted redward of Ly) plotted over the observed spectrum. Panel (b) shows iron modeling and is divided into three frames: top, showing the continuum+iron model plotted over the overall spectrum; middle, showing the iron subtracted spectrum; and bottom, showing the fitted iron template. Panels (c)to( f ) show modeling of the O vi,Ly,C iv and C iii] emission line regions. Each panel for each emission line region is divided into three frames, where the top frame shows the total best-fit model plotted over the relevant region of each spectrum, middle frame the residuals, and bottom frame the individual Gaussian components. (Note that Ly, C iv and C iii] are modeled with two components: narrow and broad, while other emission lines are modeled using one Gaussian.) The absorption lines that overlap each emission line are plotted at the top of the bottom panel. The dashed vertical lines in the emission line panels are drawn at the expected emission line position ° ° calculated using the redshift quoted at the top of the figure. Flux units are 10þ14 ergs cmþ2 sþ1 Aþ1, wavelength units are in A and are observed frame values. [See the electronic edition of the Supplement for a color version of this figure.]


264

KURASZKIEWICZ ET AL.
TABLE 4 Continuum Parameters Designation (1) 0005+1609ra ........ 0010+1058ra ........ 0027+2241ra ........ 0047+0319ra ........ 0053+1241ra ........ 0054+2525ra ........ 0103+0221ra ........ 0113+2958ra ........ 0115þ0127ra......... 0118+0258ra ........ 0123þ5848ra......... 0125þ0005ra......... 0136+2057ra ........ 0242þ0000ra......... 0242þ0000rb ........ 0256þ0126ra......... 0319+4130ra ........ 0323þ4930ra......... 0336þ3607ra......... 0351þ1429ra......... 0352þ0711ra......... 0357þ4812ra......... 0405þ1308ra......... 0407þ1211ra......... 0417þ0553ra......... 0420þ5456ra......... 0441þ4313ra......... 0456þ2159ra......... 0506þ6109ra......... 0615+7102ra ........ 0630+6905ra ........ 0630+6905rb ........ 0635þ7516ra......... 0713+1146ra ........ 0743þ6726ra......... 0744+3753ra ........ 0745+3142ra ........ 0755+2542ra ........ 0837+4450ra ........ 0840+1312ra ........ 0847+3445ra ........ 0853+4349ra ........ 0859+4637ra ........ 0902þ1415ra......... 0906+1646ra ........ 0909+4253ra ........ 0919+5106ra ........ 0927+3902ra ........ 0927+3902rb ........ 0955+6903ra ........ 0955+6903rb ........ 0955+6903rc ......... 0956+4115ra ........ 0957+5522ra ........ 0958+3224ra ........ 1003+6813ra ........ 1004+2855ra ........ 1010+4132ra ........ 1011+1304ra ........ 1013+3551ra ........ 1024+1912ra ........ 1041+0610ra ........ 1042+1203ra ........ CUVa (2)
× 1.71þ × 0.94þ × 0.77þ × 1.17þ × 0.75þ × 0.21þ × þ0.07þ × 0.86þ × 1.38þ × 2.51þ × 1.46þ × 1.00þ × 1.86þ × 1.56þ × 1.35þ × 1.56þ × 3.12þ × 0.01þ × 2.23þ × 2.43þ × 1.57þ × 1.80þ × 1.46þ × 2.02þ × 1.62þ × 0.36þ × 1.54þ × 1.91þ × 1.05þ × 0.00þ 1 × 1.49þ × 2.57þ × 1.15þ 1 × 1.49þ × 1.05þ × þ1.30þ × þ0.90þ × 0.96þ × 1.12þ × 1.83þ 1 1 × 1.31þ × 0.25þ × 1.42þ × 1.86þ × 1.80þ × þ1.86þ × 0.28þ × þ0.60þ × 1.72þ × 1.00þ × 1.52þ × 0.38þ × 1.74þ × 1.96þ 1 × 1.50þ 1 1 1 0:02 0:02 0:09 0:05 0:33 0:37 0:61 0:04 0:03 0:02 0:48 0:45 3:42 5:47 0:82 0:64 0:22 0:22 0:43 0:43 0:03 0:01 0:07 0:31 0:05 0:05 0:03 0:03 0:06 0:06 0:20 0:20 2:92 3:96 0:18 0:16 0:50 0:48 0:03 0:30 0:81 0:77 0:26 0:26 0:05 0:03 0:06 0:06 0:16 0:15 0:02 0:03 0:12 0:12 0:55 0:02 0:87 0:87 0:05 0:74 0:11 0:09 0:24 0:24 0:13 0:01 0:07 0:05 0:04 0:03 2:47 3:27 3:94 2:62 0:05 0:04 0:04 0:03 0:17 0:10

The error analysis follows the procedure from Paper I (for details see x 3.5 of that paper), in which the 2 errors for each emission line parameter were determined from 2 confidence interval bounds (D2 ¼ 4:0). We determine upper limits for the amplitude of the line from the 2 positive error, estimated while fixing the line position at the expected wavelength and the FWHM at the median value found for that line in the LBQS sample (see Table 3). Because of the large number of spectra in the preCOSTAR FOS Spectral Atlas plots of spectral fits are available only in electronic form from our Web site.1 In Figure 1 we present an example of one spectrum 0005+1609ra (PKS 0003+158) and its continuum, iron, emission, and absorption line modeling.
5. CONTINUUM AND EMISSION LINE MEASUREMENTS

Norm.b (3) 1.284×0:: þ0 1.208×0:: þ0 0.219×0:: þ0 0.338×0:: þ0 1.920×0:: þ0 2.016×0:: þ0 0.066×0:: þ0 0.004×0:: þ0 0.079×0:: þ0 0.051×0:: þ0 4.459×0:: þ0 0.221×0:: þ0 0.169×0:: þ0 0.477×0:: þ0 0.308×0:: þ0 0.301×0:: þ0 0.389×0:: þ0 0.027×0:: þ0 0.048×0:: þ0 1.223×0:: þ0 0.235×0:: þ0 0.255×0:: þ0 0.197×0:: þ0 2.062×0:: þ0 0.491×0:: þ0 0.145×0:: þ0 0.271×0:: þ0 0.625×0:: þ0 0.094×0:: þ0 0.023×0:: þ0 2.627×0:: þ0 0.163×0:: þ0 0.960×0:: þ0 0.385×0:: þ0 0.489×0:: þ0 0.079×0:: þ0 0.565×0:: þ0 0.018×0:: þ0 0.003×0:: þ0 0.090×0:: þ0 2.757×0:: þ0 0.301×0:: þ0 0.319×0:: þ0 0.207×0:: þ0 0.093×0:: þ0 0.027×0:: þ0 0.510×0:: þ0 0.414×0:: þ0 0.213×0:: þ0 0.117×0:: þ0 0.174×0:: þ0 0.133×0:: þ0 1.460×0:: þ0 0.062×0:: þ0 0.355×0:: þ0 0.354×0:: þ0 1.324×0:: þ0 0.542×0:: þ0 0.356×0:: þ0 1.084×0:: þ0 0.128×0:: þ0 0.409×0:: þ0 0.119×0:: þ0
009 009 028 041 004 004 032 001 011 013 106 107 014 014 001 001 002 002 002 002 033 046 008 002 003 003 006 006 006 006 005 005 071 071 002 002 002 002 019 008 011 011 004 004 002 003 010 011 008 008 002 002 003 003 012 001 005 005 001 001 072 072 005 006 019 019 018 002 009 009 001 002 005 005 002 002 001 001 001 001 050 059 003 010 029 029 012 012 001 001 001 001 006 006 002 003 002 002 053 055 012 012 010 010 012 014 001 001 004 004 006 006 009 009 005 005 006 006 108 108 007 007 036 036 008 008

norm (4) 2120.6 1593.1 3099.0 2373.9 2302.6 1689.2 2037.3 1993.4 3458.8 2445.3 1531.2 3027.4 2084.1 4235.6 4235.6 2748.0 1488.1 1566.3 3071.2 2363.7 2869.4 2932.3 2296.9 2299.8 2595.9 4240.7 2329.8 2242.6 3061.0 4277.0 1568.7 2003.6 2417.5 2588.6 3206.5 3017.1 2136.7 2114.8 1766.0 2458.2 1556.1 2214.1 2457.9 3083.3 2065.2 2442.4 2276.1 2478.6 2478.6 1462.4 4219.6 2238.1 1804.9 2791.9 2238.4 2593.0 1944.7 2358.0 3030.3 1564.9 2422.1 2503.4 2323.2

Copta (5) .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . ×0 06 0.71þ0::06 ×0 12 0.23þ0::11 .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . ×0 07 1.16þ0::07 .. . .. . .. . þ1:21×0::11 þ0 10 .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. .

The parameters of the power-law continuum are presented in Table 4, where column (1) gives the name of the spectrum, and column (2) the slope of the UV power-law continuum CUV (where f / þþUV ). For spectra with only one continuum window present, a constant slope of þ ¼ 1 (with no errors) was adopted, since the mean of the continuum slopes for the spectra in the FOS sample with more than one window, was 1:16 ô 0:89. In column (3) the normalization of the power law in units of 10þ14 ergs cmþ2 sþ1 ° Aþ1 is listed for the observed wavelength norm given in column (4). For spectra extending toward the optical, a second power law was introduced with a slope Copt presented in column (5). The optical power law was normalized at ° rest ¼ 4220 A to have the same continuum flux as the UV power law. The slopes and normalizations of the optical/ UV continuum parameters are quoted with 2 errors. In Table 5 we present in detail the emission line measurements for each object/spectrum from the pre-COSTAR FOS Spectral Atlas. However, due to the large size of this table it is available only in the electronic edition of the Supplement. For readers' convenience we present here a printed version of Table 5 in a digestible format for one example spectrum, 0005+1609ra.
6. STATISTICS

The number of emission lines we measured in the preCOSTAR FOS/HST sample is shown in Table 6. The total number of emission lines modeled adds up to 1710 among which 86 are upper limits. Most measured lines lie in the UV range, and only a small number (9) of spectra extend toward optical wavelengths. The means and medians of the rest frame equivalent widths and FWHM for each emission line and their standard deviations are presented in Table 6. These are calculated for the detections (cols. [4]­[6] and [10]­[12]), and when upper limits are present a nonparametric survival analysis technique was used and a KaplanMeier estimator applied to estimate the means and medians of emission line parameters (cols. [7]­[8]; for reference see Isobe, Feigelson, & Nelson 1986 and Lavalley, Isobe, & Feigelson 1992). The means and medians of equivalent width for Ly,C iv,C iii], Mg ii, and H lines were calculated separately for single, narrow and broad components, as well as
1

0:03 0:03 0:30 0:29 0:14 0:14 0:03 0:02 0:04 0:03 6:34 84:78 0:39 0:43 0:25 0:23 0:03 0:02 0:18 0:18 0:20 0:19 0:14 0:15 0:03 0:02 0:04 0:04 1:62 1:81

See http://hea-www.harvard.edu/~ pgreen/HRCULES.html.


TABLE 4--Continued Designation (1) 1048þ2509ra......... 1051þ0051ra......... 1058+1951ra ........ 1104+7658ra ........ 1106þ0052ra......... 1107+1628ra ........ 1114+4037ra ........ 1119+2119ra ........ 1125+5910ra ........ 1130þ1449ra......... 1132+1023ra ........ 1139+6547ra ........ 1139þ1350ra......... 1139þ3744ra......... 1151+5437ra ........ 1158+6254ra ........ 1204+2754ra ........ 1208+4540ra ........ 1210+0954rA ....... 1210+0954rB ........ 1210+3924ra*....... 1210+3924rb ........ 1210+3924rc ......... 1214+1403ra ........ 1219+0638ra ........ 1221+7518ra ........ 1221+7518rb ........ 1225+3332ra ........ 1229+0203ra ........ 1230+1223rA ....... 1230+1223rB ........ 1231þ0224ra......... 1233+0931ra ........ 1244+1721ra ........ 1247+3209ra ........ 1252+5634ra ........ 1254+1141ra ........ 1256þ0547ra......... 1259+3423ra ........ 1301+5902ra ........ 1305þ1033ra......... 1305þ1033rb ........ 1308+3005ra ........ 1309+0819ra ........ 1319+2728ra ........ 1319+5148ra ........ 1323+2910ra ........ 1331+3030ra ........ 1336+1725ra ........ 1341+4123ra ........ 1342+6021ra ........ 1343+2844ra ........ 1349+5341ra ........ 1354+0052ra ........ 1357+1919ra ........ 1358+5752ra ........ 1359þ4152ra......... 1404+0937ra ........ 1405+2555ra ........ 1409+2618ra ........ CUVa (2)
× 0.00þ × 1.62þ 1 × 1.28þ × 1.43þ × 1.83þ × 1.66þ × 1.61þ 1 × 0.53þ × 1.12þ × 1.81þ × 1.59þ × 1.11þ × 1.27þ × 1.33þ × 0.49þ × 1.00þ × 1.00þ × 1.00þ × 1.70þ × 1.05þ × 0.89þ × 1.12þ × 1.56þ × 2.04þ × 1.52þ × 0.04þ × 1.72þ × 0.01þ × 0.00þ 1 × þ0.25þ 1 1 × 1.44þ × 1.32þ × 0.31þ 1 × 1.55þ × 1.43þ 1 × 1.70þ × 0.30þ × 1.76þ × 1.74þ 1 1 × 1.23þ × 1.73þ × 0.86þ 1 1 × 1.14þ × 1.46þ 1 × 0.30þ × þ4.63þ × 0.69þ × 2.16þ 0:10 0:14 0:04 0:03 0:03 0:02 0:07 0:05 0:08 0:08 0:06 0:04 0:02 0:02 0:75 0:82 0:30 0:01 0:05 0:04 0:17 0:16 0:02 0:01 0:26 0:25 0:38 0:36 0:05 0:06 1:49 0:04 1:60 1:45 2:75 0:80 0:01 0:01 0:44 0:45 0:48 0:51 0:01 0:01 0:03 0:02 1:06 0:97 0:02 0:01 0:01 0:01 0:02 0:02 0:07 0:07 0:18 0:90 0:96 1:06

TABLE 4--Continued norm (4) 4270.6 1988.9 2416.0 1918.1 2081.6 2386.8 2536.0 1720.6 2461.9 3198.5 2199.6 2407.3 2281.5 1474.9 2879.7 2331.2 1704.3 3156.1 7045.3 7045.3 4233.9 1467.3 1467.3 4561.4 1947.0 1566.0 1815.0 1464.0 1694.0 4237.7 4237.7 2341.5 2069.4 3011.7 2489.8 1932.0 2736.3 2246.7 3146.9 2138.5 1869.7 1463.8 2641.3 1689.2 2957.2 3012.8 2503.4 2362.1 2271.7 3245.3 2868.0 2433.6 2267.1 3096.1 2515.5 3141.6 1920.3 2107.5 1702.4 2837.2 Copta (5) þ1.90×0::22 þ0 06 .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . ×0 04 1.6þ0::03 .. . .. . ×0 11 2.69þ0::11 .. . .. . .. . .. . .. . ×0 10 0.46þ0::09 þ2.25×0::26 þ0 24 .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . Designation (1) 1417+2508ra ........ 1417+2508rb ........ 1417+4456ra ........ 1418+1703ra ........ 1427+1949ra ........ 1427þ1203ra......... 1436+6336ra ........ 1437þ0147ra......... 1442+3526ra ........ 1445+0958ra ........ 1454þ3747ra......... 1514+3650ra ........ 1524+0958ra ........ 1539+4735ra ........ 1539+4735rb ........ 1547+2052ra ........ 1557+3304ra ........ 1613+3412ra ........ 1614+2604ra ........ 1620+1736ra ........ 1620+1736rb ........ 1633+3924ra ........ 1634+7031ra ........ 1638+5720ra ........ 1642+3948ra ........ 1658+0515ra ........ 1704+6044ra ........ 1719+4804ra ........ 1821+6420ra ........ 1927+7358ra ........ 2044þ1043ra......... 2114+0607ra ........ 2131þ1207ra......... 2137þ1432ra......... 2143+1743ra ........ 2148+0657ra ........ 2203+3145ra ........ 2218þ0335ra......... 2218þ0335rc ......... 2225þ0457ra......... 2232+1143ra ........ 2246þ1206ra......... 2253+1608ra ........ 2253+1608rb ........ 2254+1136ra ........ 2303þ6807ra......... 2311+1008ra ........ 2342þ0322ra......... 2346+0930ra ........ 2351þ0109ra......... 2355þ3357ra......... CUVa (2)
× 0.58þ × 1.04þ × 1.00þ 1 × 1.00þ × 2.27þ 1 1 1 1 × 0.12þ × 2.08þ 1 1 × 1.35þ × 1.37þ × 1.06þ × 1.31þ 1 × 1.62þ × 1.70þ × 2.43þ 1 × 1.87þ × 0.91þ 1 × 1.50þ × 1.58þ × 1.52þ × 1.24þ × 1.15þ × 0.82þ × 1.56þ × þ0.40þ × 1.01þ × 1.78þ × 1.43þ × 1.01þ × 1.29þ × 0.41þ × 1.37þ × 2.30þ × 1.10þ × 1.06þ × 0.82þ × 1.48þ × 1.63þ × 1.57þ × 1.20þ × 1.79þ × 1.59þ 0:04 0:01 0:11 0:09 1:01 0:34 0:22 1:02 0:19 0:19

Norm.b (3) 0.027×0:: þ0 0.866×0:: þ0 0.243×0:: þ0 1.087×0:: þ0 0.418×0:: þ0 0.657×0:: þ0 0.076×0:: þ0 4.088×0:: þ0 0.358×0:: þ0 0.130×0:: þ0 0.315×0:: þ0 0.547×0:: þ0 0.457×0:: þ0 22.191×0:: þ0 0.444×0:: þ0 0.362×0:: þ0 0.221×0:: þ0 0.425×0:: þ0 0.017×0:: þ0 0.005×0:: þ0 8.390×0:: þ0 33.921×0:: þ0 9.335×0:: þ0 0.874×0:: þ0 1.077×0:: þ0 2.023×0:: þ0 2.493×0:: þ0 0.138×0:: þ0 24.331×0:: þ0 0.109×0:: þ0 0.007×0:: þ0 0.210×0:: þ0 0.577×0:: þ0 0.406×0:: þ0 0.150×0:: þ0 0.171×0:: þ0 0.217×0:: þ0 0.267×0:: þ0 0.092×0:: þ0 1.005×0:: þ0 3.001×0:: þ0 0.109×0:: þ0 0.245×0:: þ0 1.135×0:: þ0 0.510×0:: þ0 0.217×0:: þ0 0.142×0:: þ0 0.178×0:: þ0 0.507×0:: þ0 0.250×0:: þ0 0.082×0:: þ0 0.199×0:: þ0 0.101×0:: þ0 0.337×0:: þ0 0.333×0:: þ0 0.166×0:: þ0 1.689×0:: þ0 0.162×0:: þ0 1.556×0:: þ0 0.506×0:: þ0
002 002 007 009 011 011 006 009 007 009 008 008 001 001 023 027 017 017 006 006 001 032 007 009 006 006 086 152 007 007 014 014 004 005 006 006 002 002 001 001 008 008 695 696 213 212 006 007 006 008 126 126 012 020 001 001 135 149 002 002 001 001 025 025 029 029 006 006 012 012 002 002 003 003 003 004 002 002 002 029 018 019 022 022 004 004 069 069 006 006 004 004 036 036 028 028 006 006 004 004 002 002 013 013 007 007 006 006 006 007 009 009 057 057 016 016 070 070 007 007

Norm.b (3) 0.898×0:: þ0 3.797×0:: þ0 0.459×0:: þ0 0.102×0:: þ0 1.756×0:: þ0 0.327×0:: þ0 0.194×0:: þ0 0.710×0:: þ0 2.106×0:: þ0 1.536×0:: þ0 0.377×0:: þ0 0.716×0:: þ0 0.514×0:: þ0 0.635×0:: þ0 0.588×0:: þ0 0.769×0:: þ0 0.028×0:: þ0 0.081×0:: þ0 0.686×0:: þ0 0.572×0:: þ0 0.446×0:: þ0 0.110×0:: þ0 1.945×0:: þ0 0.279×0:: þ0 0.324×0:: þ0 0.209×0:: þ0 1.087×0:: þ0 1.338×0:: þ0 5.336×0:: þ0 1.017×0:: þ0 11.316×0:: þ0 0.569×0:: þ0 1.156×0:: þ0 0.599×0:: þ0 1.097×0:: þ0 0.360×0:: þ0 2.133×0:: þ0 0.224×0:: þ0 0.206×0:: þ0 0.037×0:: þ0 0.198×0:: þ0 0.581×0:: þ0 0.476×0:: þ0 0.286×0:: þ0 0.739×0:: þ0 0.236×0:: þ0 0.683×0:: þ0 0.467×0:: þ0 0.429×0:: þ0 0.410×0:: þ0 0.355×0:: þ0
007 017 135 148 006 006 007 007 112 111 006 006 007 007 076 076 118 118 009 011 011 011 005 008 005 005 012 012 006 006 011 011 001 001 001 001 069 069 007 007 004 006 005 005 013 013 003 004 005 006 015 015 010 012 010 010 016 024 008 011 031 049 003 010 008 008 014 014 013 014 007 007 018 020 003 004 003 003 001 001 002 002 008 008 007 007 003 003 011 012 005 005 006 006 006 006 007 007 046 046 004 004

norm (4) 1487.5 1487.5 2417.4 2412.8 1627.8 2641.3 3271.7 2503.4 1828.9 1853.4 1921.7 2004.6 3075.3 2263.9 2591.7 1849.0 2840.2 3511.5 1654.1 2274.3 2274.3 2958.6 3092.6 2560.3 2329.5 2489.7 2006.7 3047.9 1896.9 1904.3 1512.8 2144.0 2195.2 1755.4 1771.2 2910.4 1893.9 2780.2 2780.2 3515.9 2979.1 2383.9 2718.8 2718.8 1938.5 2211.3 2096.2 2772.9 2452.6 1717.0 2495.0

Copta (5) .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

0:27 0:26 0:04 0:03

0:10 0:10 0:03 0:03 0:46 0:47 0:20 0:20 0:14 0:13 0:04 0:03 0:68 0:75 0:05 0:04 0:06 0:06 0:03 0:03 0:12 0:12 0:01 0:01 0:03 0:02 0:01 0:01 0:05 0:03 0:03 0:03 0:19 0:22 0:03 0:03 0:30 0:29 0:02 0:02 0:06 0:05 0:06 0:06 0:34 0:34 0:04 0:04 0:12 0:12 0:16 0:15 0:04 0:03 0:03 0:02 0:13 0:12 0:03 0:03 0:14 0:14 0:16 0:15 1:12 0:96 0:12 0:12

0:02 0:02 0:16 0:15 0:02 0:02 0:11 0:01 0:02 0:02 0:19 0:19 0:53 0:55 0:19 0:20 0:26 0:26

0:16 0:16 0:26 0:24 0:33 0:35

0:34 0:35 0:07 0:06 0:32 0:30 2:37 3:09 0:38 0:42 0:20 0:20

Notes.--Table 4 is also available in machine-readable form in the electronic edition of the Astrophysical Journal Supplement. (*) See notes on individual spectra in the Appendix. a The power-law continuum slopes C UV and Copt are defined as f / þþ . CUV is fitted at rest < 4200, Copt at rest > 4200. Slopes with no listed errors show the assumed slope value in cases where only a single continuum window was available. b Normalization of the UV power law in units of 10þ14 ergs cmþ2 sþ1 ° Aþ1, at observed wavelength norm.


TABLE 5 Representative Emission Line Measurements Observed Flux (10þ14 ergs cmþ2 sþ1) (5)

Emission Line (1)

FWHM (km sþ1) (2)

Dvpeak (km sþ1) (3)

W ° (A) (4)

Absorption Lines (6)

Designation: 0005+1609ra, Redshift: z = 0.45000 UV iron................ Optical iron .......... Ly ...................... Ly narrow .......... Ly broad ............ N v ....................... O i ........................ Si iv +O iv].......... C iv narrow .......... C iv broad ............ He ii blend............ Al iii ..................... Si iii]..................... C iii] narrow ......... C iii] broad ........... 1000×9000 þ250 ... ×220 6400þ200 1680×40 þ40 7350×80 þ80 7000×3600 þ80 3600×1200 þ600 5400×320 þ320 ×30 2260þ30 10200×90 þ80 13000×400 þ400 4500×2800 þ2100 2200×420 þ380 500×220 þ200 3100×220 þ200 .. . ... 1200×100 þ100 þ60×20 þ20 þ120×40 þ40 800×320 þ90 þ700×380 þ380 þ350×160 þ160 ×10 þ50þ10 420×40 þ40 300×240 þ240 2500×1200 þ1000 100×200 þ200 þ200×90 þ90 350×120 þ120 2:6×3:: þ2 ... ×1: 15:5þ1: 24:1×1:: þ1 70:7×1:: þ1 14:6×8:: þ0 1:9×1:: þ0 7:1×0:: þ0 ×0: 22:3þ0: 61:2×0:: þ1 25:3×1:: þ1 1:0×1:: þ0 2:1×0:: þ0 0:7×0:: þ0 6:2×0:: þ0
7 5 0 0 2 1 6 5 1 7 0 6 8 8 5 5 9 0 7 6 3 7 7 6 5 3 8 8 ×3 4 2:4þ2::3 ... 23:9×1::6 þ1 5 61:6×3::0 þ2 9 180:4×4::1 þ3 9 36:0×207:0 þ1: ×2 4 4:4þ1::3 14:3×1::7 þ1 6 37:56×0::9 þ0 9 103:0×1::6 þ1 6 38:7×2::6 þ2 4 ×1 6 1:2þ0::9 ×0 9 2:6þ0::8 ×0 6 0:8þ0::4 7:3×1::0 þ0 9

.. .. 3 .. 2 .. .. 3 .. .. 3 .. .. .. 3

. . . . . . . . . .

Notes.--Table 5 is available in its entirety in the electronic edition of the Astrophysical Journal Supplement. A portion is shown here for guidance regarding its form and content. We present here line measurements for one example spectrum 0005+1609ra. All measurements are rest frame except for flux. Dvpeak is the offset of the peak of the Gaussian emission line model, in km sþ1, from the expected position based on the tabulated redshift. Spectral modeling of this object is shown in Fig. 1.

TABLE 6 Rest Frame Emission Line Parameter Distributions ° W (A) Detected Emission Line (1) UV iron................ Optical iron .......... Ly +O vi ........... Ly: Singlea .............. Narrow ............ Broad ............... Sumb ................ N v ....................... O i ........................ Si iv +O iv].......... C iv: Singlea .............. Narrow ............ Broad ............... Sumb ................ He ii 1640 ........... Al iii ..................... C iii]: Singlea .............. Narrow ............ Broad ............... Sumb ................ Si iii]..................... Mg ii: Singlea .............. Narrow ............ Broad ............... Sumb ................ Total (2) 85 9 97 10 133 133 143 141 133 129 4 121 121 125 110 91 45 47 47 92 45 8 25 25 33 Limits (3) 22 1 2 0 0 0 0 12 17 0 0 0 0 0 2 17 0 0 0 0 7 2 0 0 2 Mean (4) 47 ô 7 85 ô 25 12 ô 2 229 29 74 112 15 2.9 13 21 30 63 91 20 6 21 7 20 24 3 136 29 33 76 ô ô ô ô ô ô ô ô ô ô ô ô ô ô ô ô ô ô ô ô ô ô 156 3 8 15 2 0.4 2 16 4 7 11 3 1 5 13 4 3 1 92 9 8 21 SD (5) 55 70 15 493 38 93 175 19 4 25 32 46 79 118 26 11 32 10 24 32 5 227 44 41 115 Median (6) 43.6 58.5 11.8 64.5 23.8 62.5 87.2 12.5 2.2 8.0 20.6 21.1 49.5 68.8 16.9 3.6 13.9 4.0 18.0 17.9 1.9 62.1 18.0 27.9 50.5 Kaplan-Meier Mean (7) 35.7 ô 3.6 76.0 ô 28.4 12.2 ô 0.8 .. . .. . .. . ... 13.5 ô 1.1 4.8 ô 2.2 .. . .. . .. . .. . ... 19.7 ô 1.5 5.0 ô 1.0 .. . .. . .. . ... 2.8 ô 0.6 101.9 ô 53.0 .. . .. . 71.5 ô 14.6 Median (8) 35.5 46.7 11.4 . .. . .. . .. ... 11.4 1.9 . .. . .. . .. . .. ... 16.7 2.5 . .. . .. . .. ... 1.5 50.5 . .. . .. 49.4 Num. (9) 64 9 95 10 133 133 .. . 133 122 129 4 121 121 .. . 109 75 45 47 47 .. . 38 6 26 26 .. . Mean (10) 4960 ô 730 8080 ô 3130 6130 ô 690 6220 ô 2540 2380 ô 220 10050 ô 960 . .. 5430 ô 540 2930 ô 300 6360 ô 650 4430 ô 3070 2900 ô 290 11090 ô 1070 . .. 10890 ô 1110 5550 ô 770 4890 ô 780 1730 ô 310 8270 ô 1660 . .. 2780 ô 600 3840 ô 1850 3260 ô 760 8230 ô 1850 . .. FWHM (km sþ1) Detected SD (11) 5880 9380 6720 8040 2560 11080 .. . 6220 3270 7400 6140 3140 11790 .. . 11560 6660 5260 2140 11370 .. . 3680 4530 3880 9430 .. . Median (12) 4500 10000 5500 6100 2260 8900 ... 5800 3200 5600 3950 2650 10400 ... 11500 5400 4800 1400 5700 ... 2000 3620 2600 8400 ...


EMISSION LINE PROPERTIES OF AGNs
TABLE 6--Continued ° W (A) Detected Emission Line (1) [Ne v] ................... [O ii] ..................... [Ne iii] .................. H ........................ [S ii] 4072.5......... H ....................... [O iii] 4363.......... He ii 4686 ........... H: Singlea .............. Narrow ............ Broad ............... Sumb ................ [O iii] 4959.......... [O iii] 5007.......... He i ...................... [N ii] 6548 .......... H: Singlea .............. Narrow ............ Broad ............... Sumb ................ [N ii] 6583 .......... [S ii] 6716.4......... [S ii] 6731 ........... Total (2) 13 10 9 8 7 7 7 8 2 5 5 7 7 7 7 6 0 6 6 6 6 5 5 Limits (3) 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 Mean (4) 39 142 98 32 22 25 16 31 70 46 88 116 218 629 8 25 ô ô ô ô ô ô ô ô ô ô ô ô ô ô ô ô .. . ô ô ô ô ô ô 20 92 56 18 13 14 8 15 96 26 59 61 134 384 4 13 SD (5) 69 291 169 51 35 38 21 41 135 58 132 162 354 1018 11 31 ... 98 575 633 88 23 9 Median (6) 4.5 24.7 70.0 26.9 10.1 9.9 15.8 28.3 70.4 36.1 35.5 61.4 208.20 611.2 5.9 26.2 .. . 61.5 198.0 308.5 56.9 23.0 3.7 Kaplan-Meier Mean (7) 35.9 ô 14.5 .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . ... .. . .. . 7.7 ô 2.2 .. . . .. .. . .. . ... .. . .. . .. . Median (8) 4.4 . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. ... . .. . .. 4.8 . .. .. . . .. . .. ... . .. . .. . .. Num. (9) 12 10 9 9 8 8 8 9 2 6 6 .. . 8 8 7 7 ... 7 7 .. . 7 6 6 Mean (10) 920 1210 1500 2610 1050 1600 1300 2870 ô ô ô ô ô ô ô ô 370 460 620 1380 440 980 520 1380 FWHM (km sþ1) Detected SD (11) 1260 1470 1860 4130 1250 2780 1480 4130 12080 1040 13230 .. . 1290 1300 2500 1010 ... 990 3380 .. . 620 1180 1050 Median (12) 700 1210 1450 1500 1080 950 1300 1550 6700 770 4550 ... 1140 1280 900 700 .. . 800 2150 ... 500 1030 300

267

6700 ô 8540 870 ô 420 8310 ô 5400 . .. 1070 ô 460 1100 ô 460 1530 ô 950 840 ô 380 ... 810 ô 370 2700 ô 1280 . .. 530 ô 230 910 ô 480 680 ô 430

68 325 392 67 20 6

40 234 258 36 10 4

Notes.--Col. (1), Emission line or line blend; col. (2), total number of emission lines modeled; col. (3), number of upper limits; col. (4), mean W of detected emission lines; col. (5), standard deviation (SD) of W measurements for detected emission lines; col. (6), median of W for detections, cols. (7)­(8), KaplanMeier reconstructed mean and median of W distribution; cols. (9)­(12), the number, mean, and median of the distribution of FWHM of the Gaussian components used to model each emission feature. a The distribution for single Gaussian component models are tabulated separately from narrow and broad components. b The distribution of the sum of the broad and narrow component W measurements included with the single component measurements.

for the whole line (called `` sum '' in Table 6) defined as the sum of narrow and broad components for the two Gaussian models and single component for the single Gaussian model. The histograms of W and FWHM for emission lines blueward of [Ne v] are presented in Figure 2. The first and third rows give the distributions for W , while the second and fourth rows give the FWHM distributions. In all panels, solid lines represent distributions for detections, while the dotted lines show the estimated W distributions from the Kaplan-Meier estimator (see, e.g., Fe UV). In the panels that show the sum of Ly, C iv, C iii] and Mg ii distributions, the shaded histograms represent results from single Gaussian component fits.

7. DISCUSSION

Comparison of the FOS sample with the previously studied LBQS sample (Paper I) shows that the means and medians of the equivalent width of the strong emission lines (UV iron, Ly , Ly, C iv, Mg ii, and He ii) are larger in the FOS sample than in the LBQS sample. The one-tailed Kolmogorov-Smirnov test showed that the values of the equivalent widths of the FOS sample were larger than the values of the equivalent widths of the LBQS sample at more than 95%, 99%, 99%, 99.9%, 99.9%, and 99.9% significance level for UV iron, Ly,

Ly, C iv, He ii, and Mg ii lines, respectively. This trend reflects the Baldwin effect (BEff), i.e., the known anticorrelation between the line equivalent width and luminosity found in AGNs. Since the FOS sample extends toward lower redshifts and luminosities than the LBQS (see Fig. 3 for Lopt vs. z dependence of the two samples), extension toward higher equivalent widths is also expected. The weaker lines or blends such as N v, O i, Si iv + O vi], C iii], and Al iii have mean and median values, similar to or smaller than the LBQS values. The one-tailed Kolmogorov-Smirnov test showed that the values of the equivalent widths of the FOS sample were smaller than the values of the equivalent widths of the LBQS sample at more than 99.9%, 95%, 99.9%, 99.9%, and 99.9% significance level for N v, O i, Si iv + O vi], C iii], and Al iii lines, respectively, which is inconsistent with the BEff for these lines. However, since N v and O i do not show a BEff (see x 7.1 below) and the C iii] + Al iii blend has been modeled including the Si iii] line in half of the FOS spectra (making the equivalent widths smaller) and none of the LBQS spectra, this is expected. 7.1. The Baldwin Effect We now nosity, i.e., chromatic magnitude study the correlations between W and UV lumithe Baldwin effect. We use the rest frame mono° luminosity at 2500 A obtained from the B (assembled from the Veron-Cetty & Veron 1996


268

KURASZKIEWICZ ET AL.

Vol. 143

Fig. 2.--Distribution of rest frame W (top and third row) and FWHM (second from top and bottom row) of the emission line properties of AGNs in the pre-COSTAR FOS sample. Dotted lines represent the estimated Kaplan-Meier distributions, shown for emission lines which have more than 10% of upper limits in their measurements. Shaded areas represent single Gaussian component distributions.

Catalog of quasars and AGNs), assuming a continuum slope ¼ 0:5 (where f $ ) and H0 ¼ 50 km sþ1 Mpcþ1, q0 ¼ 0:5, and ö ¼ 0. We use the survival analysis package ASURV (Lavalley et al. 1992) to allow for the presence of a small number of upper limits in our data and applied the following tests: the Cox proportional hazard model, the generalized Kendall rank and the Spearman rank test. We considered a correlation significant if the probability of the correlation occurring by chance was 1% in all these tests. The probabilities, slopes, and intercept coefficients of the

Baldwin effect regressions are quoted in Table 7. We found a significant Baldwin effect for the major emission lines: Ly, Si iv, and C iv. These correlations have been found previously in the LBQS sample as well as by other authors (Kinney, Rivolo, & Koratkar 1990; Zamorani et al. 1992; Espey & Andreadis 1999). We confirm the Ly BEff discovered by Zheng, Kriss, & Davidsen (1995). This was not significant in the LBQS data (unless only the detections were analyzed), probably due to an insufficient luminosity range (1 dex in LBQS cf. 2.5 dex in FOS).


No. 2, 2002

EMISSION LINE PROPERTIES OF AGNs

269

Fig. 2.--Continued

We find a weak Baldwin effect in Fe UV when all data are analyzed (P ¼ 0:03=0:05=0:05 in Cox/Kendall/Spearman test) and a marginally significant correlation when only the detections are analyzed (P ¼ 0:02=0:00=0:00). When the FOS and LBQS data are combined together the Baldwin effect is significant with the probability of a chance correlation less than 0.01%, confirming the BEff discovered in Paper I. The optical iron will not be studied here since the number of new points (10) is too small to conduct any reasonable statistical analysis. We also find a significant BEff for C iii] which was not detected in the LBQS. This could be due either to a larger

° L(2500 A) range covered by the FOS sample or to a more detailed model fit of the C iii] region here which includes the Si iii] emission feature. We also find a Si iii] BEff which has not been previously reported. No BEff is present for N v. This line is sensitive to the abundances (Hamann & Ferland 1993), and the abundances are claimed to increase with redshift or luminosity, working against the BEff (Ferland et al. 1996; Dietrich & Wilhelm-Erkens 2000). No BEff was found for O i (either here or in the LBQS) or for Al iii (present in LBQS). Mg ii does not show any BEff perhaps due to too few data points (33).


270

KURASZKIEWICZ ET AL.

Vol. 143

Fig. 2.--Continued

° We also looked at the correlations between L(2500 A) and the broad and narrow components of strong emission lines. We find that both the narrow and the broad components of Ly and C iv show significant BEff with the probability of a chance correlation in the Cox, Kendall, and Spearman tests less than 0.01%, while C iii] shows a marginal correlation for the narrow component (P ¼ 0:02=0:00=0:01 in Cox/Kendall/Spearman tests, respectively) and an even weaker trend for the broad component (P ¼ 0:11=0:06=0:07). The BEff for the narrow component is stronger than for the broad component in all these lines (Spearman's for Ly is þ0.476 and þ0.411, for

C iv is þ0.523 and þ0.510, and for C iii] is þ0.372 and þ0.272, for the narrow and broad components, respectively). This is consistent with Francis et al. (1992) and Osmer, Porter, & Green (1994) who found, performing the spectral principal component analysis on the LBQS spectra, that most variance in the spectra is caused by the change in emission line cores. However, adding the broad and narrow components, and analyzing the line as a whole does not add scatter to the BEff. The probability of the BEff correlation occurring by chance is smaller (or the amplitude of Spearman's is larger) for the sum of the narrow and broad components than for the narrow or broad components


No. 2, 2002

EMISSION LINE PROPERTIES OF AGNs

271

Fig. 2.--Continued

analyzed separately (Spearman's for Ly sum is þ0.494, for C iv sum is þ0.524, and for C iii] sum is þ0.372). 7.2. Evolution We find significant correlations between the line equivalent widths and redshift for: Fe UV, Ly, Si iv, C iv, He ii, Si iii], C iii], and a weak trend for Ly and Al iii (see Table 8 for the chance probabilities, slopes, and intercepts of the evolution regressions). We also examine, following the LBQS analysis, whether the primary (stronger) correlations are with z or L using the Partial

Spearman Rank Analysis (see Green et al. 2001 for details). For Fe UV and the blended lines of Si iv, He ii, Al iii, Si iii], and C iii] the primary correlations are with redshift (see Table 9). However, for other emission lines--Ly, Ly, N v, O i, C iv, and Mg ii--the primary correlations are with luminosity. The LBQS analysis, which showed that all the emission lines correlated more strongly with redshift than with luminosity, implied that an evolutionary effect could be the cause of the BEff. However, the LBQS is a magnitude-limited sample (16:0 mBJ 18:85) where the L-z correlation is extremely strong, making it difficult to disentangle


272

KURASZKIEWICZ ET AL.

Vol. 143

Fig. 3.--Luminosity vs. redshift in the FOS and the LBQS samples. Open squares represent FOS AGNs, dots the LBQS quasars.

between the luminosity and evolutionary effects. The analysis of the more heterogeneous FOS sample, which has a better coverage of the L-z plane (see Fig. 3), shows that the primary correlations with redshift found in the LBQS sample for the strong emission lines are possibly due to the LBQS being magnitude limited. However, for Fe UV and Si iv lines the primary correlations are with redshift both for the LBQS alone and the FOS sample

(and also when FOS+LBQS samples are analyzed together) and probably is a real effect. Whether this also applies to the He ii, Al iii, Si iii], and C iii] lines is more difficult to answer, as these lines are heavily blended: He ii is a blend of six lines (see Appendix), and although we do our best to deblend the C iii] + Si iii] + Al iii complex, any conclusions for these individual lines may be tenuous.

TABLE 7 Baldwin Effect Regressions Emission Line Line Name Fe UV ..................... Fe Optical ............... O vi +Ly .............. Ly ......................... N v .......................... O i ........................... Si iv ......................... C iv ......................... He ii ........................ Al iii ........................ Si iii]........................ C iii] ........................ Mg ii ....................... Total 84 9 96 141 139 132 128 123 109 91 45 92 33 Limits 21 1 2 0 11 12 0 0 2 17 7 0 2 All Data (Including Upper Limits) C/K/S
a

Detections Only C/K/S
a

Slope þ0.102 þ0.258 þ0.260 þ0.137 þ0.046 þ0.078 þ0.150 þ0.145 þ0.069 þ0.142 þ0.203 þ0.104 þ0.038 ô ô ô ô ô ô ô ô ô ô ô ô ô 0.061 0.099 0.064 0.019 0.063 0.030 0.026 0.029 0.037 0.070 0.084 0.021 0.052

Intercept 4.22 8.71 9.09 6.14 2.26 2.61 5.52 6.25 3.29 4.60 6.30 4.41 2.76 ô ô ô ô ô ô ô ô ô ô ô ô ô 1.87 2.83 1.97 0.60 1.92 0.91 0.79 0.87 1.14 2.12 2.55 0.63 1.62

Slope þ0.091 þ0.095 þ0.257 þ0.137 þ0.060 þ0.097 þ0.150 þ0.145 þ0.064 þ0.127 þ0.184 þ0.104 þ0.085 ô ô ô ô ô ô ô ô ô ô ô ô ô 0.023 0.420 0.068 0.019 0.055 0.035 0.026 0.029 0.035 0.065 0.084 0.021 0.037

Intercept 4.30 4.34 9.00 6.14 2.81 3.24 5.52 6.25 3.16 4.33 5.82 4.41 4.27 ô ô ô ô ô ô ô ô ô ô ô ô ô 0.65 11.83 2.14 0.60 1.69 1.04 0.79 0.87 1.08 2.00 2.56 0.63 1.11

0.03/0.05/0.05 0.10/0.29/0.28 0.00/0.00/0.00 0.00/0.00/0.00 0.46/0.08/0.10 0.02/0.02/0.03 0.00/0.00/0.00 0.00/0.00/0.00 0.04/0.00/0.00 0.54/0.55/0.47 0.00/0.01/0.01 0.00/0.00/0.00 0.31/0.27/0.28

0.02/0.00/0.00 0.90/0.80/0.82 0.00/0.00/0.00 0.00/0.00/0.00 0.00/0.03/0.04 0.00/0.01/0.01 0.00/0.00/0.00 0.00/0.00/0.00 0.00/0.00/0.00 0.02/0.14/0.11 0.01/0.02/0.02 0.00/0.00/0.00 0.03/0.10/0.10

Notes.--Schmitt two-dimensional Kaplan-Meier regression fits and errors from ASURV. The Baldwin effect examined here is log LÏ2500 G÷ / log W (line). a Probability of a correlation occurring by chance from ASURV for (C) Cox proportional Hazard model, (K) generalized Kendall's tau, and (S) Spearman's tests.


No. 2, 2002

EMISSION LINE PROPERTIES OF AGNs
TABLE 8 Evolution Regressions

273

Emission Line Line Fe UV .............. Fe Optical ........ O vi +Ly ....... Ly .................. N v ................... O i .................... Si iv .................. C iv .................. He ii ................. Al iii ................. Si iii]................. C iii] ................. Mg ii ................ Total 85 9 97 143 141 134 129 125 110 91 45 92 33 Limits 22 1 2 0 11 13 0 0 2 17 7 0 2

All Data (Including Upper Limits) C/K/S
a

Detections Only C/K/S
a

Slope þ0.195 þ0.116 þ0.202 þ0.167 0.007 þ0.056 þ0.233 þ0.198 þ0.108 þ0.459 þ0.375 þ0.236 0.016 ô ô ô ô ô ô ô ô ô ô ô ô ô 0.070 0.306 0.124 0.028 0.118 0.051 0.053 0.036 0.049 0.114 0.180 0.029 0.162

Intercept 0.999 1.211 0.946 1.864 0.849 0.188 0.818 1.724 1.139 0.015 þ0.117 1.122 1.640 ô ô ô ô ô ô ô ô ô ô ô ô ô 0.116 0.788 0.045 0.026 0.068 0.059 0.035 0.048 0.034 0.104 0.127 0.034 0.129

Slope þ þ þ þ þ þ þ þ þ þ þ þ þ 0.187 0.076 0.241 0.167 0.058 0.090 0.233 0.198 0.119 0.310 0.298 0.236 0.170 ô ô ô ô ô ô ô ô ô ô ô ô ô 0.035 0.321 0.119 0.028 0.099 0.068 0.053 0.036 0.069 0.088 0.215 0.029 0.053

Intercept 1.399 1.461 0.957 1.864 0.917 0.217 0.818 1.724 1.140 0.276 0.016 1.122 1.534 ô ô ô ô ô ô ô ô ô ô ô ô ô 0.067 0.800 0.025 0.026 0.055 0.048 0.035 0.048 0.043 0.088 0.157 0.034 0.084

0.08/0.00/0.01 0.12/0.32/0.29 0.27/0.02/0.03 0.00/0.00/0.00 0.96/0.36/0.38 0.14/0.08/0.11 0.00/0.00/0.00 0.00/0.00/0.00 0.01/0.00/0.00 0.06/0.00/0.00 0.00/0.00/0.00 0.00/0.00/0.00 0.45/0.98/0.80

0.01/0.00/0.00 0.90/0.89/0.89 0.37/0.01/0.02 0.00/0.00/0.00 0.01/0.08/0.08 0.01/0.04/0.05 0.00/0.00/0.00 0.00/0.00/0.00 0.00/0.00/0.00 0.01/0.00/0.00 0.01/0.01/0.01 0.00/0.00/0.00 0.04/0.41/0.52

Notes.--Schmitt two-dimensional Kaplan-Meier regression fits and errors from ASURV. The relations examined here are log W Ïline÷ / log z. a Probability of a correlation occurring by chance from ASURV for (C) Cox proportional Hazard model, (K) generalized Kendall's tau, and (S) Spearman's .

7.3. Baldwin Effect and Evolution Slopes All the FOS BEff slopes are significantly flatter than in the LBQS. This finding is consistent with the Netzer, Laor, & Gondhalekar (1992) model in which the Baldwin effect is caused by geometrically thin, optically thick accretion disc with a viewing angle dependent optical/UV emission and an isotropic line emission. A sample of objects with differing disk luminosities and inclination angles produces, in this model, a Baldwin effect slope dependent on the luminosity range of the sample, with flatter slopes expected for samples of larger luminosity range. The slopes of the Ly and C iv Baldwin effect are consistent with those found for a heterogeneous sample of IUE spectra studied by Kinney et al. (1990; 0.12 ô 0.05 for Ly, and 0.17 ô 0.04 for C iv) and Espey & Andreadis (1999; 0.08 ô 0.03 for Ly, and 0.17 ô 0.03 for C iv) as well as for

a number of optically selected complete samples from Zamorani et al (1992; 0.13 ô 0.03 for C iv). Interestingly, the slopes of the evolution effect for all emission lines in the FOS (and FOS + LBQS) sample are also flatter than in the LBQS, resembling the behavior of the BEff slopes which flatten with increasing luminosity range. We find a trend (chance probability of $4% in both Kendall andSpearmantests)ofananticorrelation between the ionization potential of the emission line and the slope of the BEff, with the slope of the BEff slopes of þ0:0013 ô 0:0003 (see Fig. 4). This trend, first found by Espey & Andreadis (1999), is consistent with BEff being caused by the change of AGN spectral energy distribution (SED) with luminosity. The emission lines with higher ionization potential are expected, in this picture, to show steeper BEff slopes due to a greater difference between the higher energy line-producing continuum and that measured underneath the line.

TABLE 9 Partial Correlations of log W All Data (Including Upper Limits) ° log L(2500 A) Line UV iron............. Ly ................... Ly ................... N v .................... O i ..................... Si iv +O iv]....... C iv ................... He ii .................. Al iii .................. Si iii].................. C iii] .................. Mg ii ................. P 0.217 <0.005 <0.005 0.037 0.043 >0.400 <0.005 >0.400 <0.005 >0.400 <0.005 <0.005 0.090 þ0.334* þ0.275* þ0.155* þ0.152* þ0.009 þ0.242* þ0.011 0.465 þ0.011 0.317 þ0.484* P 0.023 0.034 >0.400 0.125 0.228 0.016 0.287 0.046 <0.005 0.086 <0.005 <0.005 log z þ0.221* 0.192 0.016 0.101 0.067 þ0.193* þ0.051 þ0.163* þ0.543* þ0.214* þ0.538* 0.455 Detections Only ° log L(2500 A) P 0.366 <0.005 <0.005 0.108 0.019 >0.400 <0.005 0.316 0.015 >0.400 <0.005 0.013 0.044 þ0.322* þ0.275* þ0.111* þ0.193* þ0.009 þ0.242* þ0.047 0.260 0.003 0.317 þ0.411* P 0.011 0.054 >0.400 0.392 0.194 0.016 0.287 0.131 <0.005 0.096 <0.005 0.045 log z þ0.293* 0.170 0.016 0.024 0.084 þ0.193* þ0.051 þ0.113* þ0.387* þ0.221* þ0.538* 0.317

Notes.--P, partial Spearman rank probability; , partial correlation coefficient of a W correlation occurring by ° chance, where W depends on two variables log L(2500 A) and log z, and the partial correlation is calculated while holding one of the variables constant. For each line the primary (stronger) correlation is denoted by an asterisk (*) on .


274

KURASZKIEWICZ ET AL.

Vol. 143

MgII NV CIII] SiIV CIV

OVI

Fig. 4.--Dependence of the slope of the Baldwin Effect on the ionization parameter of the emission line in the FOS sample. The filled squares are the slopes predicted by the LOC model of Korista et al.

Korista, Baldwin, & Ferland (1998) have calculated emission line fluxes using a `` locally optimally emitting cloud (LOC) '' model in which clouds with a large range of densities, sizes, and distances from the ionizing continuum source compose the broad emission line region, where those clouds with the optimal physical parameters for emission in a given line contribute the most flux. When the clouds are illuminated by a model power law with a changing continuum slope, the LOC model predicts the BEff slopes shown as filled squares in Figure 4. These slopes are generally consistent with the slopes we find in the FOS sample, confirming that the SED change is the likely cause of the BEff. All the above analysis of the slope of BEff slopes has been made by ignoring the N v BEff slope, which is much flatter than all the other emission line slopes. It is, however, consistent with the Korista et al. prediction, if the BEff flat slope is due to the metallicity increase in high-redshift/luminosity quasars (Hamann & Ferland 1993).
8. CONCLUSIONS

In this paper we present the emission line properties of 158 AGNs (174 spectra) observed by FOS/HST before the installation of COSTAR. Using an automated technique, which accounts for galactic extinction, blended iron emission, and galactic and intrinsic absorption lines, we uniformly measure the equivalent widths, FWHM, and peak shifts relative to the systemic redshift of the UV/ optical emission lines spanning from Ly to H. The measurements are quoted with errors, and upper limits to

equivalent widths are estimated where emission lines are undetected. The FOS spectra, together with the 993 LBQS spectra from Paper I, comprise the largest emission line database wherein emission and absorption lines have been uniformly measured. In this paper we study the relation between emission line equivalent widths and UV luminosity in the FOS sample. We find a significant Baldwin effect for Ly, Ly, Si iv, He ii, C iv, and C iii] and a previously unreported Si iii] Baldwin effect. We find that narrow components of Ly, C iv, and C iii] show a stronger Baldwin effect than the broad components, consistent with the spectral principal component analysis which showed that most variance comes from emission line cores (Francis et al. 1992; Osmer et al. 1994). We also find a Baldwin effect in UV iron, significant when both FOS and LBQS samples are analyzed together. The values of the Baldwin effect slopes and the dependence of the slopes on the sample's luminosity range point to a change of the SED as the cause of the Baldwin effect in the FOS sample. We also study the relations between the line equivalent widths and redshift and find significant anticorrelations for Fe UV, Ly, Si iv, C iv, He ii, Si iii], and C iii] and trends for Ly and Al iii. For UV iron and weak or blended lines (Si iv, He ii, Al iii, Si iii], C iii]) stronger (primary) correlations are with redshift, while for other emission lines (Ly, Ly,N v,O i,C iv,Mg ii) the primary correlations are with luminosity. This is inconsistent with the LBQS sample, where all emission lines correlated more strongly with redshift than with luminosity. We conclude that the primary


No. 2, 2002

EMISSION LINE PROPERTIES OF AGNs

275

correlations with redshift in the LBQS sample may be due to the LBQS sample being magnitude limited. This database will become useful in a number of projects, vital for the understanding of the physics of the broad emission line region. The large range of redshifts and luminosities covered by the FOS + LBQS sample [0 < z < 3:0, and 1028 < LÏ2500 G÷ < 1032 ergs sþ1] will, for example, enable tests for the dependence of abundances (estimated from the N v/C iv, N v/He ii line ratios--Hamann & Ferland 1993) on luminosity or redshift. Correlations between abundances and continuum properties such as luminosity, radio-loudness, or UV/X-ray continuum shape will give clues to the factors driving the enrichment of gas in the BLR. The FOS absorption line measurements can be used to measure the frequency of UV absorption in AGNs, whether it depends on luminosity of the AGN, and whether it is consistent with the X-ray absorption (confirmed only in Seyfert 1 galaxies, see Crenshaw et al. 1999). Finding a relation between intrinsic UV absorption lines and continuum properties (e.g., Brandt, Laor, & Wills 2000 find stronger

UV absorption for weaker soft X-ray QSOs) will help assess the physics of the absorbing gas, its location and origin. Finally, the relation of any emission line property or line ratio with the width of the BLR lines will give clues to the dependence of the BLR properties on the black hole mass (since black hole mass /FWHM2 assuming BLR cloud motions are virialized). The authors gratefully acknowledge support provided by NASA through grant NAG5-6410 (LTSA). This work has made use of the NASA/IPAC Extragalactic Database (NED) which is operated by the Jet Propulsion Laboratory, California Institute of Technology, under contract with the National Aeronautics and Space Administration. Based on observations made with the NASA/ESA Hubble Space Telescope, obtained from the data archive at the Space Telescope Science Institute. STScI is operated by the Association of Universities for Research in Astronomy, Inc., under the NASA contract NAS 5-26555.

APPENDIX
A1. NOTES ON EMISSION LINES

Ly + O vi 1035.--The spectrum near this emission line blend was modeled with a flat `` pseudo '' continuum and a single Gaussian, since Ly and O vi are so closely blended. The resulting equivalent width should be treated as approximate. He ii 1640.--The region on the red side of C iv 1549 includes emission from a number of lines including He ii 1640, [Ne v] 1575, 1593, [Ne iv] 1602, 1609, Si ii 1650, [O iii] 1661, 1663, 1668, and Al ii 1670. However, we model this region with a single He ii Gaussian component, as the use of multiple Gaussian components did not improve the model fit. The true value of He ii line parameters (particularly the FWHM) is likely to be significantly smaller than quoted in Table 5. [Ne v] 3426, [O ii] 3728, [Ne iii] 3869.--These lines lie on top of the Balmer continuum which was not accounted for during the global power-law continuum fit. Hence, we measured the lines above a local continuum, defined as the power-law ° fit to 30 A wide continuum windows on the red and blue side of each emission line.
A2. NOTES ON INDIVIDUAL SPECTRA

0010+1058ra.--This spectrum consists of two spectra that do not overlap and are simply joined to form a single spectrum. ° 0027+2241ra.--Blue end of the G190H spectrum is noisy and was ignored below 1630 A before merging. ° 0420þ5456ra.--Not a good continuum fit to 4000­4800 A region; for a better H fit, an intermediate H component was introduced in addition to the narrow and broad components. 0743þ6726ra.--Heavily absorbed O vi. 0837+4450ra.--A gap in spectrum at C iv line. 0847+3445ra.--Spectrum consists of two spectra, which do not overlap but have similar flux level and are simply joined. 0927+3902ra, rb.--Spectrum ends at Mg ii. ° 0956+4115ra.--Blue end of the G190H spectrum is noisy and was ignored below 1605 A before merging. 1003+6813ra.--Heavily absorbed O vi and Ly. ° 1004+2855ra.--Blue end of the G190H spectrum is noisy and was ignored below 1605 A before merging (note that this ° was ignored due to second-order grating reflection effect. region lies on Ly); the region at 1516­1523 A 1011+1304ra.--Heavily absorbed O vi and Ly. 1048þ2509ra.--Strong Galactic absorption in Mg ii. ° 1139+6547ra.--G150L spectrum was ignored above 1650 A due to a low S/N. 1208+4540ra.--Heavily absorbed O vi and Ly. 1210+3924rb, 1210+3924rc.--Heavily absorbed Ly and C iv. ° 1219+0638ra.--Blue end of the G190H spectrum is noisy and was ignored below 1605 A before merging; note that this region lies on Ly. 1225+3332ra.--Galactic absorption contaminating C iv and Mg ii. ° 1229+0203ra.--The strong absorption feature at 1209­1223 A does not appear in all spectra that were merged; this may be due to oversubtraction of geocoronal Ly in some of these spectra; we exclude this region from further analysis. Also, there is ° no overlap in the constituent spectra around 1600­1650 A, leaving a break at the Si iv + O iv. 1231þ0224ra.--Spectrum ends at Ly. 1244+1721ra.--Heavily absorbed O vi. 1259+3423ra.--Strong absorption in O vi.


276

KURASZKIEWICZ ET AL.

° 1301+5902ra.--Blue end of the G190H spectrum is noisy and was ignored below 1605 A before merging. 1341+4123ra.--Heavily absorbed O vi and Ly. 1342+6021ra.--Heavily absorbed Ly and C iv. 1354+0052ra.--Heavily absorbed Ly. ° 1357+1919ra.--Blue end of the G190H spectrum is noisy and was ignored below 1605 A before merging. 1417+2508rb.--A merger of 40 spectra of NGC 5548, which is known to have variable luminosity and emission line profiles. 1524+0958ra.--Heavily absorbed O vi and Ly. 1620+1736ra.--Gratings change at O vi. 1634+7031ra.--Heavily absorbed O vi. ° 1704+6044ra.--Blue end of the G190H spectrum is noisy and was ignored below 1605 A; heavily absorbed O vi,Ly. 1719+4804ra.--Heavily absorbed O vi and Ly. ° 1821+6420ra.--Blue end of the G190H spectrum is noisy and was ignored below 1605 A before merging. 2044þ1043ra.--Strong NAL absorption in Ly. 2137þ1432ra.--Strong NALs in Ly and C iv. 2218þ0335ra, rb, rc.--Gratings change at Ly. 2303þ6807ra.--Gratings change at O vi. ° The following spectra were fitted with continuum consisting of two power laws, which were joined at 4200 A rest frame: 0242þ0000ra, 0242þ0000rb, 0420þ5456ra, 0615+7102ra, 1048þ2509ra, 1214+1403ra, 1230+1223rA, and 1230+1223rB. However, in 1210+3924ra discontinuous optical and UV power laws provide a better fit to the UV part of the spectrum. The ° normalization of the optical power law at 4220 A rest frame is 7.439×0::001 . þ0 001
REFERENCES Kinney, A. L., Rivolo, A. R., & Koratkar, A. P. 1990, ApJ, 357, 338 Aldcroft, T. A. 1993, Ph.D. thesis, Stanford Univ. Koratkar, A., & Gaskell, C. M. 1991, ApJ, 345, 637 Alexander, T., & Netzer, H. 1994, MNRAS, 270, 781 Korista, K., Baldwin, J., & Ferland, G. 1998, ApJ, 507, 24 Baldwin, J. A., Wampler, E. J., & Gaskell, C. M. 1989, ApJ, 338, 630 Krolik, J. H., & Kallman, T. R. 1988, ApJ, 324, 714 Boroson, T. A., & Green, R. F. 1992, ApJS, 80, 109 (BG92) Laor, A., Fiore, F., Elvis, M., Wilkes, B. J., & McDowell, J. C. 1994, ApJ, Brandt, W. N., Laor, A., & Wills, B. J. 2000, ApJ, 528, 637 435, 611 Brotherton, M. S., Wills, B. J., Francis, P. J., & Steidel, C. C. 1994, ApJ, ------. 1997, ApJ, 477, 93 430, 495 Lavalley, M., Isobe, T., & Feigelson, E. D. 1992, in ASP Conf. Ser. 25, Cardelli, J. A., Clayton, G. C., & Mathis, J. S. 1989, ApJ, 345, 245 Astronomical Data Analysis Sortware and Systems I, ed. D. M. Worral, Crenshaw, D. M., Kraemer, S. B., Boggess, A., Maran, S. P., Mushotzky, C. Biemsderfer, & J. Barnes (San Francsico: ASP), 245 R. F., & Wu, C.-C. 1999, ApJ, 516, 750 Lockman, F. J., & Savage, B. D. 1995, ApJS, 97, 1 Dietrich, M., & Wilhelm-Erkens, U. 2000, A&A, 354, 17 Marziani, P., Sulentic, J. W., Dultzin-Hacyan, D., Calvani, M., & Moles, Elvis, M., Wilkes, B. J., & Lockman, F. J. 1989, AJ, 97, 77 M. 1996, ApJS, 104, 37 Espey, B., & Andreadis, S. 1999, in ASP Conf. Ser. 162, Proc. Quasars as Murray, N., & Chiang, J. 1998, ApJ, 494, 125 Standard Candles for Cosmology, ed. J. Baldwin & G. J. Ferland (San Netzer, H., & Davidson, K. 1979, MNRAS, 187, 871 Francisco: ASP), 351 Netzer, H., Laor, A., & Gondhalekar, P. M. 1992, MNRAS, 254, 15 Ferland, G. J., Baldwin, J. A., Korista, K. T., Hamann, F., Carswell, R. F., Osmer, P. S., Porter, A. C., & Green, R. F. 1994, ApJ, 436, 678 Phillips, M., Wilkes, B. J., & Williams, R. E. 1996, ApJ, 461, 683 Perry, J. J., & Dyson, J. E. 1985, MNRAS, 213, 665 Forster, K., Green, P. J., Aldcroft, T., Vestergaard, M., & Foltz, C. B. Stark, A. A., Gammie, C. F., Wilson, R. W., Bally, J., Linke, R. A., Heiles, 2001, ApJS, 134, 35 (Paper I) C., & Hurwitz, M. 1992, ApJS, 79, 77 Francis, P. J., Hewett, P. C., Foltz, C. B., & Chaffee, F. H. 1992, ApJ, 398, Veron-Cetty, M.-P., & Veron, P. 1996, Catalog of Quasars and Active 476 Galactic Nuclei (7th ed.; Garching: ESO) Freeman, P., Doe, S., & Siemiginowska, A. 2001, Proc. SPIE, 4477, 76 Vestergaard, M., & Wilkes, B. J. 2001, ApJS, 134, 1 Gehrels, N. 1986, ApJ, 303, 336 Wilkes, B. J., Kuraszkiewicz, J., Green, P. J., Mathur, S., & McDowell, Green, P. J., Forster, K., & Kuraszkiewicz, J. K. 2001, ApJ, 556, 727 (Paper J. C. 1999, ApJ, 513, 76 II) Wills, B. J., Laor, A., Brotherton, M. S., Wills, D., Wilkes, B. J., Ferland, Hamann, F., & Ferland, G. J. 1993, ApJ, 418, 11 G. J., & Shang, Z. 1999, ApJ, 515, L53 Heiles, C., & Cleary, M. N. 1979, Australian J. Phys. Astrophys Supp., 47, Zamorani, G., Marano, B., Mignoli, M., Zitelli, V., & Boyle, B. J. 1992, 1 MNRAS, 256, 238 Isobe, T., Feigelson, E. D., & Nelson, P. I. 1986, ApJ, 306, 490 Zheng, W., Kriss, G. A., & Davidsen, A. F. 1995, ApJ, 440, 606 Keyes, C. D., Koratkar, A. P., Dahlem, M., Hayes, J., Christiansen, J., & Zheng, W., & Malkan, M. A. 1993, ApJ, 415, 517 Martin, S. 1995, Faint Object Spectrograph Instrument Handbook, version 6.0 (Baltimore: STSci)