9512.net
甜梦文库
当前位置:首页 >> >>

Spectral Variability of Quasars in the Sloan Digital Sky Survey. II The C IV Line


Spectral Variability of Quasars in the Sloan Digital Sky Survey. II: The C iv Line1
Brian C. Wilhite2,3,4 , Daniel E. Vanden Berk5 , Robert J. Brunner3,4 , Jonathan V. Brinkmann6

arXiv:astro-ph/0512313v1 12 Dec 2005

ABSTRACT We examine the variability of the high-ionizaton C ivλ1549 line in a sample of 105 quasars observed at multiple epochs by the Sloan Digital Sky Survey. We ?nd a strong correlation between the change in the C iv line ?ux and the change in the line width, but no correlations between the change in ?ux and changes in line center and skewness. The relation between line ?ux change and line width change is consistent with a model in which a broad line base varies with greater amplitude than the line core. The objects studied here are more luminous and at higher redshift than those normally studied for variability, ranging in redshift from 1.65 to 4.00 and in absolute r-band magnitude from roughly ?24 to ?28. Using moment analysis line-?tting techniques, we measure line ?uxes, centers, widths and skewnesses for the C iv line at two epochs for each object. The wellknown Baldwin E?ect is seen for these objects, with a slope β = ?0.22. The sample has a median intrinsic Baldwin E?ect slope of βint = ?0.85; the C iv lines in these high-luminosity quasars appear to be less responsive to continuum variations than those in lower luminosity AGN. Additionally, we ?nd no evidence for variability of the well known blueshift of the C iv line with respect to the low-ionization Mg iiλ2798 line in the highest ?ux objects, indicating that this blueshift might be useful as a measure of orientation. Subject headings: galaxies: active – quasars: general – techniques: spectroscopic
Presented as part of a dissertation to the Department of Astronomy and Astrophysics, The University of Chicago, in partial ful?llment of the requirements for the Ph.D. degree. The University of Chicago, Department of Astronomy and Astrophysics, 5640 S. Ellis Ave., Chicago, IL 60637
3 4 5 2 1

The University of Illinois, Department of Astronomy, 1002 W. Green St., Urbana, IL 61801 National Center for Supercomputing Applications, 605 E. Spring?eld Ave., Champaign, IL 61820

The Pennsylvania State University, Department of Astronomy and Astrophysics, 525 Davey Lab, University Park, PA 16802
6

Apache Point Observatory, P.O. Box 59, Sunspot, NM 88349

–2– 1. Introduction

Quasar emission lines represent light reprocessed by high-velocity ionized gas surrounding a central continuum source. As the central source continuum varies, the emission lines vary in response. Quasar continuum variability is well studied. Long-known anticorrelations between variability amplitude and luminosity (e.g., Uomoto, Wills, & Wills 1976; Cristiani, Trentini, La Franca, & Andreani 1997) and between variability amplitude and wavelength (e.g., Giveon et al. 1999; Tr`vese, Kron, & Bunone 2001; Wilhite et al. 2005), as well as the e correlation between variability and time lag (e.g., Hook, McMahon, Boyle, & Irwin 1994; Hawkins 2002; de Vries, Becker, & White 2003) were recently parameterized by Vanden Berk et al. (2004). Variability of quasar emission lines is also well studied, but for a much smaller number of objects. Much of this work has come as a part of reverberation mapping e?orts to determine black hole masses and study the structure of the broad-line region (e.g., Peterson 1993; Wandel, Peterson, & Malkan 1999), and has generally focused on line response to continuum variability. These emission line variability studies have been critical in characterize the structure of the broad-line region. By measuring the response delays of various emission lines, with respect to the continuum, reverberation mapping has shown that the broad-line emitting region species are strati?ed by ionization potential (e.g. Peterson 1993), and that the size of the BLR is dependent upon continuum luminosity (e.g. Wandel, Peterson, & Malkan 1999; Kaspi et al. 2005). Until recently, C ivλ1549 had been monitored in only a few low-redshift, low-luminosity objects, such as NGC5548, monitored for years with both IUE and HST (Clavel et al. 1991; Korista et al. 1995), and NGC 5141, observed for short-term variability with IUE (Crenshaw et al. 1996). Kaspi et al. (2003) have begun a campaign to use C iv variability in reverberation mapping of high-redshift, high-luminosity quasars, but the results are not yet conclusive. Pro?le variability of AGN emission lines has been studied (e.g., Wanders & Peterson 1996; Sergeev, Pronik, & Sergeeva 2001), but most of this has been done for nearby, lowluminosity Seyferts and has been limited to the rest-frame optical. The reasons for this are well motivated—low-luminosity objects are known to be more variable and optical spectroscopy is more common—but this means there has been relatively little study of line variability in higher luminosity objects, or of rest-frame ultraviolet lines like C iv. Although not well studied in the time domain, the C iv line has demonstrated several intriguing properties at a single epoch which suggest study of C iv variability could prove useful in understanding the structure of the broad emission line region and quasars as a whole.

–3– 1.1. The C iv Line Profile

Wills, Fang, & Brotherton (1992) and Francis, Hewett, Foltz, & Cha?ee (1992) found that C iv line width is anticorrelated with the equivalent width of the line, a result even more clearly demonstrated by Wills et al. (1993). The highest ?ux C iv lines tend to be the most narrow. Wills et al. (1993) suggested that this might be due to di?erent relative importances of an intermediate width line region (ILR) and the very broad line region (VLBR) from quasar to quasar. In this scenario, the narrow (? 2000 km/s) line core is produced in the ILR, which lies near the outer edge of the broad-line region (BLR). The broad (? 7000 km/s) line base is a product of the VBLR, which comprises the inner portion of the BLR. According to the ILR model, possible line core ?uxes extend over a larger range in values than the line base ?uxes. Thus, the width of an individual line is strongly dependent upon the strength of the ILR line core. Strong C iv lines have dominant cores and are therefore narrow. Similarly, weak C iv lines have less dominant cores and, thus, relatively more important line bases, and are preferentially broader as a result. Murray & Chiang (1997) found that the C iv pro?le could be reproduced with a continuous disk-wind model and did not distinguish between intermediate-width and very broad line regions. Wills et al. (1993) also found that the C iv line is typically asymmetric, in the sense that C iv lines are generally skewed to shorter wavelengths. Richards et al. (2002a) found that this asymmetry tends to increase with increasing C iv blueshift; the most blueshifted lines also tend to be the most skewed toward the blue end of the spectrum.

1.2.

The Baldwin E?ect in the C iv Line

Baldwin (1977), and others later (Kinney, Rivolo, & Koratkar 1990; Baskin & Laor 2004) demonstrated that the equivalent width of the C iv line is anticorrelated with the luminosity of the nearby continuum for quasars observed at a single epoch: WC iv ∝ Lβ λ (1)

The initial ?t to the slope was β = ?0.64 (Baldwin, Burke, Gaskell, & Wampler 1978). Kinney, Rivolo, & Koratkar (1990) found a lower value of β = ?0.17 ± 0.04 and later studies (c.f., Dietrich et al. 2002) have found similar results. The Baldwin E?ect may also be recast in terms of line luminosity, giving the similar

–4–

LC iv ∝ Lb , λ

(2)

where b = β + 1. In this form, the relation is easier to understand. From quasar to quasar, as the continuum luminosity increases, the C iv line luminosity increases, but at a slower rate. By using equivalent width as a proxy for luminosity, it had originally been hoped that the Baldwin E?ect could be used as a cosmological probe. Unfortunately, the roughly halfmagnitude scatter about the original relation is too large to allow for precision cosmology. Richards et al. (2002a) found that the Baldwin E?ect in C iv appears to be related to the blueshift of the line with respect to lower ionization lines, such as Mg ii (see § 1.3). They separated almost 800 quasars into four equally populated bins, splitting by the size of the blueshift; from the quasars in each bin, they created a composite spectrum. They found a clear anticorrelation between the strength of the C iv line and the size of the blueshift. The bin with the largest C iv-Mg ii blueshift also had the lowest equivalent width composite C iv line, and vice versa. It has been suggested that the C iv Baldwin E?ect could be largely reproduced through a softening of the continuum slope with increasing luminosity and luminosity-dependent quasar metallicity (Korista et al. 1998). Wang et al. (1998) found that the ultraviolet to X-ray spectral index is correlated with quasar luminosity: more luminous quasars have softer ionizing continua slopes. They also found that the UV to X-ray index is strongly correlated with C iv equivalent width. The combination of these e?ects leads directly to the Baldwin E?ect: quasars with high luminosity display low C iv equivalent width. Though these relations are consistent with the Baldwin E?ect, the physical driver itself is not yet understood. Recently, Baskin & Laor (2004) found that the correlation between C iv equivalent width was much stronger with L1/2 (Hβ F W HM)?2 , a proxy for L/LEDD (since the black hole mass scales as L1/2 (Hβ F W HM)2 ), than it was with the simple continuum luminosity. They have suggested that the Baldwin E?ect may in fact be a secondary e?ect spawned by a more fundamental relation between C iv equivalent width and the relative accretion rate, L/LEDD . However, the potential physical mechanism driving the relation is unknown. The roughly half-magnitude scatter in the Baldwin E?ect was shown by Kinney, Rivolo, & Koratkar (1990) to be at least partially due to continuum and C iv line variability. As a quasar’s continuum luminosity increases or decreases, the C iv line luminosity (which consists largely of reprocessed continuum photons) increases or decreases in turn, with a small delay owing to the light travel time. An intrinsic relationship between continuum and line luminosities may be written in forms identical to those for the global Baldwin relation (WC iv ∝ Lβint and LC iv ∝ Lbint ). Kinney, Rivolo, & Koratkar (1990) found that the λ λ

–5– so-called ”intrinsic Baldwin E?ect” (IBE) slope ranged from β = -0.4 to -0.9 for 6 Seyfert galaxies and 3C 273 with an average of βint ≈ ?0.65 (bint ≈ 0.35). The intrinsic Baldwin E?ect is, for historical reasons, usually cast in terms of equivalent width, but is more straightforward when expressed in terms of luminosity. As was the case with the global Baldwin E?ect, the slope of the IBE is between 0 and 1. This indicates that, for an individual quasar, the BLR reprocessing of the incident continuum light is not perfectly e?cient. As the continuum luminosity of an individual quasar ?uctuates, so too does the CIV line luminosity, but to a lesser degree. If an object had an intrinsic Baldwin E?ect slope of 0.35, a doubling in a quasar’s continuum luminosity would only lead to a roughly 25% increase in CIV line luminosity (after allowing for the light-travel time delay). The intrinsic Baldwin E?ect (IBE) slope itself has been found to vary. Over 13 years of monitoring, the IBE slope of the Hβ line in the Seyfert I galaxy NGC 5548 ranged from b = 0.4 to 1.0 on time scales of roughly one year (Goad, Korista, & Knigge 2004). The slope was strongly anti-correlated with continuum ?ux, indicating a lower line responsivity at higher continuum ?ux levels, which is consistent with photoionization models (Korista & Goad 2004). Pogge & Peterson (1992) determined that the Baldwin E?ect scatter may be further reduced by accounting for the light travel time, τ , between the continuum source and the broad emission line region: LC iv (t) ∝ L(t ? τ )β .

1.3.

C iv Line Shifts

Gaskell (1982) ?rst demonstrated that high-ionization quasar broad emission lines (such as C iv) are typically blueshifted by hundreds of kilometers per second with respect to the low-ionization lines, thought to represent the true systemic redshift of the quasar. This was veri?ed in a number of later studies (Wilkes 1984; Espey et al. 1989; Corbin 1990; Tytler & Fan 1992; McIntosh, Rix, Rieke, & Foltz 1999; Sulentic, Marziani, & Dultzin-Hacyan 2000). Recently, Richards et al. (2002a) measured the blueshift of the C iv line with respect to Mg ii for ? 800 quasars in the SDSS Early Data Release Quasar Catalog (Schneider et al. 2002). A possible correlation between C iv blueshift and radio-determined orientation measures, as well as a similarity between the spectra of broad absorption line quasars and quasars with large C iv blueshifts, prompted Richards et al. (2002a) to suggest the possibility that C iv blueshift could be used as a measure of quasar orientation, either internal (related to the disk wind opening angle) or external (related to the line of sight to the observer). They proposed that the blueshift might be a result of the obscuration or suppression of the C iv ?ux on the

–6– red side of the line. If the blueshift of the C iv line, relative to low-ionization lines like Mg ii, is related to the observer’s viewing angle, it could represent the ?rst technique to measure orientation for radio-quiet quasars.

1.4.

The Present Work

This is the second paper reporting results of a quasar spectral variability program using data from the Sloan Digital Sky Survey (SDSS; York et al. 2000). The ?rst paper (Wilhite et al. 2005, hereafter Paper I) examined the detailed wavelength dependence of quasar variability. This paper focuses on the high-ionization C ivλ1549 line. We brie?y summarize the SDSS data acquisition, our previous spectrophotometric recalibration work, and the creation of the variable quasar sample in § 2. In § 3, we describe the line-?tting algorithm used here. The variability of the C iv line ?ux and pro?le is studied in § 4. Interesting individual objects are identi?ed in § 5. The results are discussed in § 6 and we conclude in § 7. Throughout the paper we assume a ?at, cosmological-constant-dominated cosmology with parameter values ?Λ = 0.7, ?M = 0.3, and H0 = 70km s?1 Mpc?1 .

2.

The Sloan Digital Sky Survey and the Variable Quasar Sample 2.1. The Sloan Digital Sky Survey

Through Summer 2004, the Sloan Digital Sky Survey (York et al. 2000) had imaged almost ? 8200deg2 and obtained follow-up spectra for roughly 5 × 105 galaxies and 5 × 104 quasars. All imaging and spectroscopic observations are made with a dedicated 2.5-meter telescope at the Apache Point Observatory in the Sacramento Mountains of New Mexico. Imaging data are acquired by a 54-chip drift-scan camera (Gunn et al. 1998) equipped with the SDSS u, g, r, i and z ?lters (Fukugita et al. 1996); they are then reduced and calibrated by the PHOTO software pipeline (Lupton et al. 2001). The photometric system is normalized such that SDSS magnitudes are on the AB system (Smith et al. 2002). A 0.5-meter telescope monitors site photometricity and extinction (Hogg et al. 2001). Point source astrometry for the survey is accurate to less than 100 milliarcseconds (Pier et al. 2003). Ivezi? et al. (2004) c discusses imaging quality control. Objects are targeted for follow-up spectroscopy as candidate galaxies (Strauss et al. 2002; Eisenstein et al. 2001), quasars (Richards et al. 2002b) or stars (Stoughton et al.

–7– 2002). Targeted objects are grouped in 3-degree diameter ”tiles” (Blanton et al. 2003) and aluminum plates are drilled with 640 holes whose locations on the plate correspond to the objects’ sky locations. Each plate is placed in the imaging plane of the telescope and plugged with optical ?bers assigned to roughly 500 galaxies, 50 quasars and 50 stars. Fibers run from the telescope to twin spectrographs. ? ? SDSS spectra cover the observer-frame optical and near infrared, from 3900A—9100A. Spectra are obtained in three or four consecutive 15-minute observations until an average minimum signal-to-noise ratio is met. The spectra and calibrated by observations of 32 sky ?bers, 8 reddening standard stars, and 8 spectrophotometric standard stars. Spectra are ?at-?elded and ?ux calibrated by the Spectro2d pipeline. Next, Spectro1d identi?es spectral features and classi?es objects by spectral type (Stoughton et al. 2002). Ninety-four percent of all SDSS quasars are identi?ed spectroscopically by this automated calibration; the remaining quasars are identi?ed through manual inspection. Quasars are de?ned to be those extragalactic objects with broad emissions lines (full width at half maximum velocity width of 1000km s?1 , regardless of luminosity. Through June 2004, objects corresponding to 181 plates had been observed multiple times, with time lags between observations ranging from days to years. As discussed in Paper I, spectra from plates observed greater than 50 days apart have not been co-added and are more suitable for use in variability studies. There are 53 such large time-lag plate pairs containing almost 2200 quasars; 47 of these plate pairs are contained in the Third Data Release (DR3; Abazajian et al. 2005).

2.2.

Re?nement of Spectroscopic Calibration

Vanden Berk et al. (2004, hereafter VB04) and Paper I demonstrated that additional spectrophotometric calibration of SDSS spectra is necessary for variability studies. We summarize here the calibration methods used in Paper I; see that work for a complete discussion. The Spectro1d pipeline calculates three values of signal-to-noise ratio for each spectrum by calculating the median S/N ratio per pixel in the portions of the spectrum corresponding to the SDSS g, r and i ?lter transmission curves. Hereafter, when referring to the two halves of a plate pair, we use the phrase ”high-S/N epoch” to refer to the plate with the higher median r-band signal-to-noise ratio. The plate with the lower median r-band signal-to-noise ratio will be called the ”low-S/N epoch.” It is worth emphasizing that this is a plate-wide designation; although most objects follow the plate-wide trend, this does not speak to the relative S/N values for any given individual object, nor does it correspond to an object’s relative line or continuum ?ux at a given epoch. The stars on a plate are used to resolve

–8– calibration di?erences between the high- and low-S/N epochs, under the assumption that the majority of stars are non-variable (precautions are taken to remove the obviously variable stars from re-calibration). For each plate pair, we create a re-calibration spectrum, equal to the ratio of the median stellar high-S/N epoch ?ux to the median stellar low-S/N ?ux, as a function of wavelength. This re-calibration spectrum is ?tted with a 5th-order polynomial to preserve real wavelength dependences, but remove pixel-to-pixel noise (see Figure 5 of Paper I), leaving a smooth, relatively featureless curve as a function of wavelength. All low-S/N epoch spectra are then scaled by this ”correction” spectrum.

2.3.

Variable Quasar Sample

In this study, we make use of the sample of variable quasars created in Paper I. Many quasars at low redshift appear as extended objects; due to the ?ber nature of the spectrograph, these low-redshift objects are problematic for accurate relative spectrophotometry between epochs. To avoid such problems with extended objects, only quasars with z > 0.5 were used in Paper I. With respect to the assumed non-variable stellar population, 315 quasars were determined to have varied signi?cantly between epochs. These variable quasars have larger rest-frame time lags than the average for all SDSS quasars with multi-epoch spectroscopy, but are otherwise indistinguishable from the main sample. This sample was ?rst used to study the detailed dependence of variability on wavelength; for more information, see Paper I.

3.

Fitting the C iv Line Region of Interest

3.1.

Using moment analysis techniques, we ?t the C iv line for all objects where the entire line has been observed by the SDSS spectrograph. Both epochs are ?t individually. Thus, for each object, we obtain ?ux, position and pro?le information for the C iv line at two epochs. The line-?tting techniques used are similar to those used by Vanden Berk et al. (2001) to ?t composite spectra created from the SDSS quasar survey, with a few modi?cations. Vanden Berk et al. (2001) ?t the composite spectrum C iv line over the rest-frame wavelength range 1494?—1620?. However, the C iv line is typically ?anked on the red side by He iiλ1640 A A and emission from other species, such as Fe ii, O iii] and Al ii, as well as some unidenti?ed ?ux above the continuum redward of 1600? (e.g., Wilkes 1984; Boyle 1990; Laor et al. 1994; A

–9– Vanden Berk et al. 2001). However, over intervals centered on roughly 1480? and 1690?, A A quasar spectra are relatively free of emission, making these logical intervals to use in ?tting the underlying continuum (see §,3.2). Spectra are not de-redshifted for ?tting; the region of interest for each quasar is determined by scaling the [1472?, 1700?] interval by 1 + z for that quasar. Only those spectra A A containing this region in its entirety are used; quasars with redshifts between 1.65 and 4.35 are available for study. Table 1 lists these 105 objects, as well as their dates of observation (MJD), redshifts (z), rest-frame time lags (?τ ), absolute magnitudes (Mr ) and both epochs’ S/N ratio. (Absolute magnitudes are calculated assuming a power law spectral energy distribution fλ ∝ λαλ , with a slope of αλ = ?1.5. ) This leaves 105 objects with redshifts ranging from 1.65 to 4.00. The ?tting procedures are explained in sections § 3.2—§ 3.6. Results of the ?ts are in Tables 2 and ??.

3.2.

Continuum Fitting and Total Line Flux

Accurately ?tting the underlying continuum near a line is critical. Too low or high a continuum ?t will lead to an overestimate or underestimate of the line ?ux. An incorrectly ?t local continuum slope could introduce an apparent asymmetry not inherent to the line itself. Despite its importance, there is no widely accepted method of continuum ?tting. The ?ts to quasar continua in Wills et al. (1993) employed either a power-law or a low-order polynomial. Vanden Berk et al. (2001) ?t the local continuum with a straight line. Over such a small wavelength range (? 150?), it appears that any reasonable function will work, A provided care is taken. For simplicity, we employ a straight-line local continuum ?t. To ensure that the ?ts were well behaved, numerous visual checks were done. We varied the size of the region of interest and the size of the region used to ?t the continuum. We use here those values which appeared to give the most accurate and stable ?ts to the continuum and the line center. Before ?tting the continuum, the region of interest in each spectrum is manually inspected for poor night sky subtraction or absorption lines. In the case of poor subtraction of a night sky line (most commonly O iλ5577), we interpolate over the a?ected region, using the pixels within 25? on either side. Night sky lines were removed from 8 spectra. There are no A cases where poor night sky subtraction occurs near the peak of a line. Absorption lines are only removed in the cases where they a?ect the ?t to the continuum; they are not removed if they lie on top of the C iv emission line itself. Only 4 absorption lines are removed.

– 10 – To ?t the continuum, a single straight line is ?t to pixels at either end of the region of interest. This corresponds roughly to ?tting the continuum with those pixels with rest-frame wavelengths between 1472? and 1487? or between 1685? and 1700?. This linear continuum A A A A ?t is then subtracted from every pixel in the region of interest to isolate the line ?ux: Fline (λ) = Ftotal (λ) ? Fcont. (λ) (3)

To avoid inclusion of emission ?ux from other sources, measurements are made over the range corresponding to a 100? interval centered on 1546? in the quasar’s rest frame. A A ? is the laboratory wavelength of C iv, the line is blueshifted in quasars such (Though 1549A that the mean rest wavelength position is actually 1546? (Richards et al. 2002a; Vanden Berk A et al. 2001).) The total ?ux for the line is simply the integral of the continuum-subtracted line ?ux density (Fλ = Fline (λ)) over this measurement interval: Fλ dλ, (4) iv where C iv indicates the interval [1496?,1596?], as described above. In § 3.3 and § 3.4, we A A calculate the ?rst three moments of the C iv line, using this continuum-free region.
C

f=

3.3.

1st Moment: Line Center

In a perfectly symmetric line, the meaning of the line ”center” is easily understood; the mean, median and mode of the line’s ?ux distribution all fall at the same wavelength. It is thought, however, that the C iv line is typically asymmetric (Wills et al. 1993). Therefore, one must choose which statistic to use. In the spectra used here, some of which have low signal-to-noise ratios, the median is a far more reliable measurement that the mean, which is easily a?ected by noisy pixels in the line wings. Thus, for robustness and simplicity, we calculate the line center using the median. The median is simply the midpoint of the line ?ux, the wavelength which evenly divides the continuum-subtracted ?ux in the [1496?,1596?] A A interval. We also use this measurement of the line center to calculate the local continuum; the continuum ?ux density is determined by evaluating the straight-line continuum ?t (see § 3.2) at the median-determined line center.

– 11 – 3.4. 2nd and 3rd Moments: Line Width and Skewness

We use the second moment about the median wavelength as a measurement of line width: σ2 =
C 2 iv (λ ? λmedian ) Fλ dλ . F dλ C iv λ

(5)

Visual inspection reveals that the Pearson Skewness, 3 × (mean ? median) , (6) σ is a more stable statistic between epochs than the third moment of the ?ux. Thus, as did Vanden Berk et al. (2001), we use the Pearson skewness to measure line asymmetry. To keep the mean from being unduly a?ected by the noisy tails of the C iv line, we measure the λF dλ mean of the line over the interval [1516?,1576?], using λmean = Fλλdλ . This is tantamount A A to Vanden Berk et al. (2001) using only the ”top 50%” of the line. The single-object spectra used here are too noisy to allow for a reliable determination of the ”top 50%”, so this 60? A interval, chosen because it closely approximates the ”top 50%”, is used instead. P skew = Fig. 1 is a histogram of the high-S/N epoch Pearson skewness values for the objects in our sample. The median value of ?0.012 ± 0.013 is consistent with no skewness of the C iv line. Vanden Berk et al. (2001) measured a C iv Pearson skewness of -0.04 for the SDSS composite quasar spectrum. The lower value measured here is likely due to the di?erence in A ?tting the continuum. Using the region around 1690? to ?t the red side of the continuum leads to a bluer continuum and, therefore, the inclusion of more ?ux from the red side of the C iv line than either Vanden Berk et al. (2001) or Wills et al. (1993).

3.5.

Line Flux Requirement

Accurate measurement of line width, skewness and C iv-Mg ii line shift requires accurate and precise determination of the line center. When searching for changes in these quantities with time, it is essential to have reliable measurements at both epochs. Visual inspection indicates that the ?tting procedure returns reasonable parameter values for those objects with ?ux fline,HSN > 200 × 10?17 erg/s/cm2 . We thus remove those 10 objects below this limit from future study. For reference, we do include the results of the ?tting code for these objects in Tables 2 and ??.

– 12 – 3.6. Errors in Fitted Quantities

We use a Monte Carlo method for determination of errors. We add random, Gaussian noise to each pixel in the region of interest by assigning a random number drawn from a Gaussian distribution with mean equal to the measured ?ux in that pixel and standard deviation equal to the measured error in that pixel. The moment analysis code is run on this altered spectrum and values for continuum ?ux, line ?ux, center, width, and skewness are calculated. This is repeated 1000 times; the error assigned to each quantity is equal to the standard deviation of the distribution of resulting values. At the high-S/N epoch, the median errors in the line ?ux and the continuum ?ux density are 33 × 10?17 erg/s/cm2 and 0.093 × 10?17 erg/s/cm2 /?, repsectively. The median error in A A the line center (as determined by the median) is 1.2?. This is less than half the value for any of the other methods used to measure the line center. The medians in the line width A and skewness error distributions are 2.8? and 0.05, respectively.

3.7.

Mg ii Line Fitting

For use in studying the e?ect of variability on the C iv-Mg ii line shift, we also ?t the Mg ii line for those objects where that line also falls in the SDSS spectroscopic wavelength range. For Mg ii, the region of interest for each line was de?ned to be those pixels which correspond to rest-frame wavelengths between 2684? and 2913?, the same range used in A A Vanden Berk et al. (2001). Requiring SDSS coverage of the entire region of interest means that the Mg ii line is ?t to those 77 objects in the C iv sample with z < 2.12. The Mg ii line is ?t at both the high- and low-S/N epochs, using the same algorithm as the C iv lines. It should be noted that there is signi?cant Fe ii emission on either side of the Mg ii line (e.g., Vestergaard & Wilkes 2001; Sigut & Pradhan 2003; Baldwin et al. 2004). No attempt has been made in this work to remove this Fe ii ?ux, as only the robust Mg ii line median measurement is used. In any detailed study of the Mg ii line pro?le, however, this Fe ii emission should be removed. See § 4.3 for a discussion of variability of the blueshift of the C iv line relative to Mg ii.

– 13 – 4. Variability of the C iv Line 4.1. Pro?le

There appears to be a strong correlation between the change in the line ?ux and the change in the line width. Fig. 2 shows the epoch-to-epoch ?ux ratio ( fHSN ) versus the fLSN ratio of line widths ( σHSN ). A Spearman rank correlation test yields a correlation coe?cient σLSN of 0.42 with a signi?cance of 2.3 × 10?5 . The Kendall correlation coe?cient is 0.29 with a signi?cance of 2.5 × 10?5 . These tests indicate there is a very low probability that no correlation exists. This strongly suggests that, for an individual object, the C iv line width increases with the line ?ux. This is opposite the sense of the single-epoch anticorrelation between C iv equivalent width and FWHM seen by Wills, Fang, & Brotherton (1992). It appears that there are two separate relations between line strength and line width: a global relation, like that seen by Wills, Fang, & Brotherton (1992) and Francis, Hewett, Foltz, & Cha?ee (1992), which suggests that, from object to object, line strength is anti-correlated with line width; and an intrinsic relation seen in Fig. 2 which suggests that line width and line strength are correlated for an individual object. There is no obvious correlation between C iv line ?ux and either line center or skewness. Fig. 3 shows the epoch-to-epoch ?ux ratio ( fHSN ) versus the change in median-determined fLSN redshift (?z = zHSN ? zLSN ) and Fig. 4 shows the ?ux ratio versus the change in Pearson skewness (?P skew = P skewHSN ? P skewLSN ). No trend is apparent in either plot. This is reinforced by the Spearman rank correlation tests. The Spearman signi?cances are 0.86 and 0.734 for the ?ux-redshift and ?ux-skewness distributions, respectively, indicating that there is no signi?cant correlation with ?ux change for either redshift change or skewness change.

4.2.

Flux

4.2.1. Luminosity and Time Dependences The upper panels of Figs. 5 and 6 show the well-known dependence of continuum variability amplitude upon rest-frame time lag and absolute magnitude, respectively. To make comparisons with past photometric studies easier we measure the change in ?ux between epochs by computing the logarithm of the ratio of the ?tted continuum ?uxes at the two f epochs for each object: ?f = ?2.5log( fcont,HSN ). The error in ?f , σ?f , is calculated for cont,LSN each object through standard error propagation: σ?f =
2.5 ln10

(

σfcont,HSN fcont,HSN

)2 + (

σfcont,LSN fcont,LSN

)2 .

We then create four equally populated bins in rest-frame time lag and absolute magni-

– 14 – tude. In each bin, we calculate the average continuum variability by removing the average ?ux change error (σ?f ) from the average ?ux change (?f ): π 2 ( ) < ?f >2 ? < σ?f >, 2

V =

(7)

as was done in VB04. Because of the small number of objects in each bin (? 25), we use the median for the average (instead of the mean as in VB04). The rest-frame time-lag dependence of variability is commonly referred to as the structure function. The upper panel of Fig. 5 shows the continuum variability (|?f |) versus rest-frame time lag for all individual objects, with the binned ensemble structure function overlaid. Similarly, the upper panel of Fig. 6 shows the binned ensemble luminosity dependence overlaid on the individual objects’ continuum variability versus luminosity. Qualitatively, the structure function in the upper panel of Fig. 5 is similar to that seen in previous studies of continuum variability (see VB04; Hook, McMahon, Boyle, & Irwin 1994). The structure function increases with time, indicating that quasar continua are more likely to appear to have varied when the time interval between observations is long, as seen in Rengstorf, Brunner, & Wilhite (2005). The well-known (e.g. VB04; Giveon et al. 1999) anti-correlation between variability amplitude and quasar luminosity is seen in the upper panel of Fig. 6; more luminous quasars tend to exhibit less continuum variability. The bin containing the intrinsically faintest objects does not follow this trend; it is unlikely that this dip in variability amplitude is statistically signi?cant. With such a small number of objects per bin, this measure of the variability is easily a?ected by large or overestimated errors in the variability of individual objects. To avoid this problem, VB04 required that each bin contain a minimum of 75 objects; with only 94 total objects, we do not have this luxury here. Quantitatively, the amplitude of the variability in the upper panels of Figs. 5 and 6 is larger than that of previous studies. This can be at least partially understood as an artifact of the creation of the sample; only those quasars which had demonstrated signi?cant variability were chosen for study in Paper I. The well-known relationships between continuum variability amplitude and time lag and luminosity do not appear to hold for the C iv line ?ux. Replacing continuum ?ux with line f ?ux, we again calculate the relative ?ux change between epochs (?f = ?2.5log( fline,HSN )), line,LSN
line,HSN line,LSN 2.5 as well as the error in the relative ?ux change (σ?f = ln10 ( fline,HSN )2 + ( fline,LSN )2 ). We then calculate the line ?ux variability, V, in each of the four bins in L and ?τ and Mr , using equation 9. These are plotted in the lower panels of Figs. 5 and 6. The lower panel of Fig. 5 appears to show a decrease in line variability amplitude with rest-frame time lag. The error bars (representing the standard deviation of the ?f distribution in each bin) are

σf

σf

– 15 – quite large, though, indicating that this decrease is not statistically signi?cant. Regardless, the variability amplitude does not obviously demonstrate the same time lag dependence as the continuum variability amplitude. One is cautioned not too read too much into this lack of dependence, however. We have no knowledge of the quasars’ individual light travel times from the central source to the C iv-emitting portion of the BLR. A true C iv structure function would likely require some correction for the light travel time to the BLR, something impossible to obtain from only-two epoch data. Thus. without this correction, is perhaps unsurprising that no dependence is evident. As seen in the lower panel of Fig. 6, the line variability luminosity dependence also fails to duplicate the relation seen in the continuum variability.

4.2.2. The Baldwin E?ect and the Internal Baldwin E?ect Using the ?t values for the continuum ?ux at the C iv line center and the ?ux of the line itself determined in § 3, we are able to recreate the Baldwin E?ect. The upper panel of Fig. 7 shows continuum luminosity versus line luminosity for all 105 objects at the high-S/N epoch. Using the IDL routine POLYFITW, the power-law slope of this relation is measured to be b = 0.78 ± 0.03 for the high-S/N epoch (and b = 0.82 ± 0.03 for the low-S/N epoch), corresponding to β = ?0.22 (and β = ?0.18) for the equivalent width formulation of equation 1. These values are in rough agreement with the β = ?0.17 ± 0.04 value measured by Kinney, Rivolo, & Koratkar (1990). Combining the data from both epochs, we calculate an Intrinsic Baldwin E?ect (IBE) slope for every object in the sample: bint = log(Lline,HSN /Lline,LSN ) log(Lcont,HSN /Lcont,LSN ) (8)

The calculated IBE slope for any one object should not be taken as a de?nitive measurement of the IBE slope for that object, since it is determined with data from only two epochs, but the distribution of IBE slopes should be meaningful. This distribution is quite wide, as is seen in the lower panel of Fig. 7. The median IBE slope of the entire sample is bint = 0.15 ± 0.49. The error in the mean is large because of the few objects with large values for the IBE slope. These large slopes are not likely to be trustworthy (a very small change in continuum luminosity between epochs can lead to a very large, but essentially meaningless, value for the Intrinsic Baldwin E?ect slope). Excluding those 12 objects with an absolute value for the IBE slope greater than 2, we ?nd bint = 0.15 ± 0.06. This is shallower than the bint ≈ 0.35 value found by Kinney, Rivolo, & Koratkar (1990). The small number of objects (6) in the

– 16 – Kinney, Rivolo, & Koratkar (1990) sample precludes us from drawing strong conclusions about this di?erence. However, if this di?erence is real, it is an indication that the C iv lines of high-luminosity quasars are less responsive to continuum variations than those of low-luminosity quasars, in agreement with the results of Kaspi et al. (2003).

4.3.

Line Shifts

We are able to reproduce the single-epoch C iv line shift with respect to Mg ii, seen recently by Richards et al. (2002a). The upper panel of Fig. 8 shows a histogram of these line shifts at the high-S/N epoch. In this representation, a positive velocity (vHSN > 0 km/s) indicates a blueshift of C iv with respect to Mg ii. The median line shift for our 77 objects is 722 km/s, with a standard deviation of 1750 km/s. With over 700 objects, Richards et al. (2002a) found a median blueshift of 824 km/s and a dispersion of 511 km/s. The lower panel of Fig. 8 shows the di?erences in line shift between the two epochs (?v = vHSN ? vLSN ). The median line shift di?erence is near 0; < ?v > = 36 km/s, but the width of the ?v distribution (1680 km/s) is almost as wide as the distribution of single-epoch line shifts. Fig. 9 is is similar to Fig. 8, but only includes those objects with the highest C iv line ?ux (fline,HSN > 800 × 10?17 erg/s/cm2 ). This high line ?ux sample shows a similar median line shift (719 km/s), and a smaller standard deviation (710 km/s) than the sample as a whole. The lower panel of Fig. 9 shows the distribution of di?erences in line shift between epochs for the high line ?ux sample. This distribution has a median of 120 km/s and dispersion of 280 km/s, much narrower than its counterpart for the entire sample. As the distribution of line shift di?erences for these objects is essentially centered at 0 km/s and has a much narrower width than the single-epoch distribution of line shifts (300 km/s vs. 940 km/s), it is consistent with no di?erence in C iv-Mg ii line shift between epochs. As will be discussed in § 6, this is further indication that C iv line shifts with respect to Mg ii might be useful as an orientation measure for radio-quiet quasars, if a high enough line ?ux can be obtained.

– 17 – 5. 5.1. Individual Objects High Line-Flux Objects

We have selected some of the highest C iv line ?ux objects in our sample to illustrate the relationship between line ?ux change and line width change. SDSS J115154.83?005904.6 was selected based on odd change in C iv line pro?le (see 5.2). In all Figures, the dark curve represents the spectrum from the high-S/N epoch, the light curve the low-S/N epoch. SDSS J081931.48+055523.6 (Fig. 10) has line ?ux ratio σHSN = 1.12. σLSN SDSS J100013.37+011203.2 (Fig. 11) has line ?ux ratio σHSN = 1.03. σLSN SDSS J231147.90+002941.9 (Fig. 12) has line ?ux ratio σHSN = 1.03 σLSN SDSS J160126.31+511038.1 (Fig. 13) has line ?ux ratio σHSN = 0.97. σLSN
fHSN fLSN

= 1.39 and width ratio = 1.09 and width ratio = 1.03 and width ratio = 0.92 and width ratio

fHSN fLSN

fHSN fLSN

fHSN fLSN

5.2.

SDSS J115154.83?005904.6

Fig. 14 shows the C iv line of SDSS J115154.83?005904.6, a quasar at redshift z = 1.93. This C iv line was noticed as a part of the manual inspection for night sky and absorption lines done before line ?tting. The low-S/N epoch line (light curve) appears bifurcated, while the high-S/N epoch line (dark curve) does not; it is thus unlikely to be the result of an intervening absorption system. These spectra were taken 281 days apart, corresponding to roughly 96 days’ separation in the quasar rest-frame. The depression near the peak of the line at the low-S/N epoch lies at ? 4500? in the observed frame, where there is not expected A to be a large contamination from night sky lines. This dip in ?ux near 4500? is not seen A in other objects observed simultaneously with the same plate. If this bifurcation is a real e?ect, this could potentially be an intriguing object for follow-up study. The line pro?le looks similar to the C iv line of NGC 3516, a well-studied intrinsic absorption system which varied by roughly 50% in absorption equivalent width (Voit, Shull, & Begelman 1987). However, in the case of NGC 3516, the absorption was visible in all 11 epochs. It should also be noted that NGC 3783, a nearby (z = 0.0097) type I AGN has shown strong variations in many of its absorption features (Kraemer et al. 2001; Crenshaw et al. 2003; Gabel et al. 2003). SDSS J115154.83?005904.6 shows little or no sign of absorption in the high-S/N epoch spectrum.

– 18 – 6. 6.1. Discussion

Broad Line Region Structure

Wills et al. (1993) found that the equivalent width of the C iv line decreased with increasing line width and that C iv lines were typically asymmetric, skewed blueward of the ?ux peak. They proposed the so-called ILR model to describe the structure of the broad-line region (BLR), in which the C iv line ?ux is produced in two distinct portions of the BLR. In this scheme, the narrow core of the C iv line is produced in the ”intermediate width emission line region” (ILR), located far from the quasar central engine. The considerably broader base of the line arises in the very broad line region (VBLR), thought to be the portion of the BLR close to the central continuum source. The line base is blueshifted with respect to the line core, presumably due to some relative bulk motion towards the observer by the VBLR gas; this blueshift of the line base produces a composite line which is skewed towards the blue. From quasar to quasar, the range of possible equivalent widths for the narrow line core appears to be much larger than the range of possible equivalent widths for the broad line base. This results in the overall line width depending primarily on the relative importance of the line core. In high-?ux lines, the line core is dominant, resulting in high equivalent widths and narrow lines. In low-?ux lines, the line core does not dominate, the line base is relatively more important, and the line is broad. This yields the single-epoch relation between equivalent width and line FWHM seen in Fig. 3 of Wills et al. (1993). Disk-wind models of the BLR, such as that of Murray & Chiang (1997), are able to reproduce the C iv line pro?les ?rst seen in Wills et al. (1993), although they do not produce enough ?ux in the blue tail of the line. In such models, the line core and base are produced in di?erent portions of one continuous BLR, rather than two distinct regions. The line base photons are simply emitted at a smaller distance from the central source, where the gas velocities are greater. Instead of coming from just two distinct line production regions, the overall line pro?le can be thought of as the superposition of in?nitely many lines lines whose width depends on the distance from the central black hole at which they were produced. The intrinsic relation between C iv ?ux and width—as an individual line gets stronger, it also becomes wider—may be explained through similar geometric arguments. In fact, one should expect individual lines to broaden with increasing ?ux if one believes a) the broad portion of the line is produced near to the central engine and b) the portion of the line produced near to the central engine is more variable. That the line base is produced nearer to the central source is widely believed. It is assumed that the increased width comes with the higher rotational velocities of the BLR gas nearer to the central black hole and thus, in a deeper potential well.

– 19 – That the ?ux from the inner region of the BLR should be more variable is an essentially geometric argument. The relatively small size allows for more coherent variability (Korista & Goad 2004). The larger the region, the more ”washed out” the ?uctuations in the continuum ?ux will be, due to a larger range in light-travel times from the various portions of the BLR to the observer. This e?ect has been seen in reverberation mapping studies; recent attempts at reverberation mapping in high-luminosity objects have not yet produced conclusive results, due to the relative lack of variability in the emission lines (Kaspi et al. 2003). The large ionizing ?ux in these high-luminosity systems pushes the BLR to great distances from the central engine, resulting in larger ranges in light-travel times throughout the BLR and decreased coherent variability. Most current models make both of the assumptions necessary to explain the observed intrinsic ?ux-width relation. Both assumptions are a part of the ILR and disk-wind models. While this relation does not currently have the power to di?erentiate between discrete and continuous models of the BLR, it does strongly rule out any models which cannot predict such a relation. It is also worth noting that in the Wills et al. (1993) ILR model, the C iv line base is blueshifted with respect to the line core. This suggests that the line skewness and median should also change with ?ux; as the more variable, blueshifted broad line base increases in relative importance, the median should be ”dragged” to the blue and the line should be skewed bluewards as well. That this is not seen at all in Figs. 3 and 4 is intriguing and merits further study.

6.2.

Orientation

Richards et al. (2002a) suggested that the size of the C iv-Mg ii line shift could be a function of the orientation of the quasar with respect to the observer or of the disk-wind opening angle. They noted that the C iv blueshift seems to be the result of the degradation of the red side of the C iv line, rather than a systemic shift of the entire emission line. A wide distribution of ?v values (as seen in the lower panel of Fig. 8) suggests that individual blueshift measurements may not be reliable. Even if the C iv-Mg ii blueshift is due to orientation, the measurements of the shift must be non-variable and reproducible in order to be a useful measure of viewing angle for an individual object, though the ensemble average would still be useful. Fig. 8 shows a large scatter in ?v when the entire sample is included, casting doubt on the reliability of any single measurement of blueshift. However, for the highest ?ux C iv lines, we have demonstrated that the line shifts for these objects

– 20 – appear to be robust (see Fig. 9). While this certainly does not constitute proof that the blueshift is an orientation e?ect, the opposite result would have caused serious problems for its use as a measure of orientation. It is certainly possible to imagine a scenario where the C iv-Mg ii line shift is a product of viewing angle but varies in size. However, in such a system, the line shift would not be useful as an orientation measure, even if it is an orientation e?ect. Although the distribution of line shift di?erences between epochs, ?v, for the whole sample is impractically large, the narrow distribution for the highest ?ux objects indicates that the Mg ii-C iv line shift may hold promise as an orientation measure. Observers are warned, however, that an adequate line ?ux is necessary to reliably measure the line shift.

6.3.

Line Flux Variability

Reverberation mapping studies indicate that the distance the BLR lies from the central source increases with continuum luminosity (e.g. Wandel, Peterson, & Malkan 1999; Kaspi et al. 2005). More luminous central sources produce a greater number of ionizing photons, ”pushing” the BLR to greater distances. As mentioned in § 6.1, one would expect the emission lines from more luminous quasars to be less variable, as their larger sizes allow less coherent variability, ”smearing out” the variations in the incident continuum over a greater range in light-travel times. The median Intrinsic Baldwin E?ect Slope for our sample (bint = 0.15 ± 0.06) is shallower than that of the value of bint ≈ 0.35 from Kinney, Rivolo, & Koratkar (1990). This is to be expected, given that objects in our sample are much higher luminosity (< Mr >= ?25.6) than those studied in Kinney, Rivolo, & Koratkar (1990). NGC 5548, for example, has an absolute magnitude MR ≈ ?22.5 (Tyson et al. 1998). It is not clear, however, why there is no sign of a decrease in C iv line ?ux variability with increasing luminosity (decreasing absolute magnitude) in Fig. 6. If more luminous sources are less variable, and line variability is a result of continuum variations, one might expect line variability to share the anti-correlation with continuum luminosity. It could simply be a function of binning; although our sample spans four magnitudes in luminosity, the di?erence between the median luminosities of the two most extreme bins is only 1.5 magnitudes. However, as this was only a two-epoch study, and we have no knowledge, for individual objects, of the time lag between continuum and CIV variability, the issue may not be that simple. A larger dynamic range, or a larger number of objects per bin, might yield a more illuminating result.

– 21 – 7. Conclusions

Using a sample of 105 quasars observed multiple times by the Sloan Digital Sky Survey, we have studied the variability of the C iv line. Spectra were ?t using moment analysis techniques and four main conclusions are drawn: 1) We ?nd a strong correlation between the change in C iv line ?ux and the change in line width. As an individual quasar’s C iv line ?ux increases, so does the C iv line width. This is consistent with any picture of the BLR in which the broad line base is produced nearer to the central engine and the portion of the BLR nearer to the central engine exhibits more coherent line ?ux variability. 2) We demonstrate that there is no apparent variability in the the blueshift of the C iv line with respect to the Mg ii line for the highest ?ux C iv lines, a possibly positive sign for the use of line shifts as an orientation measure. 3) With our measurements of continuum and line ?uxes, we are able to reproduce the Baldwin E?ect, deriving a slope of b = 0.78. We also calculate a median slope for the Intrinsic Baldwin E?ect of bint = 0.15, shallower than the bint ≈ 0.35 determined by Kinney, Rivolo, & Koratkar (1990) for lower luminosity AGN. 4) Using the continuum ?ux at the position of the line center, we reproduce well-known dependences of continuum variability amplitude on quasar luminosity and rest-frame time lag. However, these same dependences are not evident for the amplitude of the C iv line variability. This may be due to the ”smearing out” of continuum variability by the extended BLR. Funding for the creation and distribution of the SDSS Archive has been provided by the Alfred P. Sloan Foundation, the Participating Institutions, the National Aeronautics and Space Administration, the National Science Foundation, the U.S. Department of Energy, the Japanese Monbukagakusho, and the Max Planck Society. The SDSS Web site is http://www.sdss.org/. The SDSS is managed by the Astrophysical Research Consortium (ARC) for the Participating Institutions. The Participating Institutions are The University of Chicago, Fermilab, the Institute for Advanced Study, the Japan Participation Group, The Johns Hopkins University, the Korean Scientist Group, Los Alamos National Laboratory, the Max-PlanckInstitute for Astronomy (MPIA), the Max-Planck-Institute for Astrophysics (MPA), New Mexico State University, University of Pittsburgh, University of Portsmouth, Princeton University, the United States Naval Observatory, and the University of Washington.

– 22 – REFERENCES Abazajian, K., et al. 2005, AJ, 129, 1755 Baldwin, J. A. 1977, ApJ, 214, 679 Baldwin, J. A., Burke, W. L., Gaskell, C. M., & Wampler, E. J. 1978, Nature, 273, 431 Baldwin, J. A., Ferland, G. J., Korista, K. T., Hamann, F., & LaCluyz?, A. 2004, ApJ, 615, e 610 Baskin, A. & Laor, A. 2004, MNRAS, 350, L31 Blanton, M. R., Lin, H., Lupton, R. H., Maley, F. M., Young, N., Zehavi, I., & Loveday, J. 2003, AJ, 125, 2276 Boyle, B. J. 1990, MNRAS, 243, 231 Clavel, J., et al. 1991, ApJ, 366, 64 Corbin, M. R. 1990, ApJ, 357, 346 Crenshaw, D. M., et al. 1996, ApJ, 470, 322 Crenshaw, D. M., Kraemer, S. B., & George, I. M. 2003, ARA&A, 41, 117 Cristiani, S., Trentini, S., La Franca, F., & Andreani, P. 1997, A&A, 321, 123 de Vries, W. H., Becker, R. H., & White, R. L. 2003, AJ, 126, 1217 Dietrich, M., Hamann, F., Shields, J. C., Constantin, A., Vestergaard, M., Cha?ee, F., Foltz, C. B., & Junkkarinen, V. T. 2002, ApJ, 581, 912 Eisenstein, D. J. et al. 2001, AJ, 122, 2267 Espey, B. R., Carswell, R. F., Bailey, J. A., Smith, M. G., & Ward, M. J. 1989, ApJ, 342, 666 Francis, P. J., Hewett, P. C., Foltz, C. B., & Cha?ee, F. H. 1992, ApJ, 398, 476 Fukugita, M., Ichikawa, T., Gunn, J. E., Doi, M., Shimasaku, K., & Schneider, D. P. 1996, AJ, 111, 1748 Gabel, J. R., et al. 2003, ApJ, 595, 120 Gaskell, C. M. 1982, ApJ, 263, 79

– 23 – Giveon, U., Maoz, D., Kaspi, S., Netzer, H., & Smith, P. S. 1999, MNRAS, 306, 637 Goad, M. R., Korista, K. T., & Knigge, C. 2004, MNRAS, 352, 277 Gunn, J. E. et al. 1998, AJ, 116, 3040 Hawkins, M. R. S. 2002, MNRAS, 329, 76 Hogg, D. W., Finkbeiner, D. P., Schlegel, D. J., & Gunn, J. E. 2001, AJ, 122, 2129 Hook, I. M., McMahon, R. G., Boyle, B. J., & Irwin, M. J. 1994, MNRAS, 268, 305 Ivezi?, Z., et al. 2004, AGN Physics with the Sloan Digital Sky Survey, Proceedings of a conc ference held in Princeton, NJ, USA, 27-31 July 2003, Edited by Gordon T. Richards and Patrick B. Hall, ASP Conference Series, Volume 311. San Francisco: Astronomical Society of the Paci?c, 2004., p.437, 437 Kaspi, S., Netzer, H., Maoz, D., Shemmer, O., Brandt, W. N., & Schneider, D. P. 2003, ASP Conf. Ser. 290: Active Galactic Nuclei: From Central Engine to Host Galaxy, 615 Kaspi, S., Maoz, D., Netzer, H., Peterson, B. M., Vestergaard, M., & Jannuzi, B. T. 2005, ApJ, accepted for publication (astro-ph/0504484) Kinney, A. L., Rivolo, A. R., & Koratkar, A. P. 1990, ApJ, 357, 338 Korista, K. T., et al. 1995, ApJS, 97, 285 Korista, K., Baldwin, J., & Ferland, G. 1998, ApJ, 507, 24 Korista, K. T. & Goad, M. R. 2004, ApJ, 606, 749 Kraemer, S. B., Crenshaw, D. M., & Gabel, J. R. 2001, ApJ, 557, 30 Laor, A., Bahcall, J. N., Jannuzi, B. T., Schneider, D. P., Green, R. F., & Hartig, G. F. 1994, ApJ, 420, 110 Lupton, R., Gunn, J. E., Ivezi?, Z., Knapp, G. R., Kent, S., & Yasuda, N. 2001, in ASP Conf. c Ser. 238, Astronomical Data Analysis Software and Systems X, ed. F. R. Harnden, Jr., F. A. Primini, and H. E. Payne (San Francisco: Astr. Soc. Pac.), p. 269, astroph[0101420] McIntosh, D. H., Rix, H.-W., Rieke, M. J., & Foltz, C. B. 1999, ApJ, 517, L73 Murray, N. & Chiang, J. 1997, ApJ, 474, 91

– 24 – Peterson, B. M. 1993, PASP, 105, 247 Pier, J. R., et al. 2003, AJ, 125, 1559 Pogge, R. W. & Peterson, B. M. 1992, AJ, 103, 1084 Rengstorf, A. .W., Brunner, R. B., & Wilhite, B. C. 2005, AJ, submitted Richards, G. T., Vanden Berk, D. E., Reichard, T. A., Hall, P. B., Schneider, D. P., SubbaRao, M., Thakar, A. R., & York, D. G. 2002, AJ, 124, 1 Richards, G. T. et al. 2002, AJ, 123, 2945 Schneider, D. P., et al. 2002, AJ, 123, 567 Sergeev, V. I., Pronik, S. G., & Sergeeva, E. A. 2001, ApJ, 554, 245 Sigut, T. A. A., & Pradhan, A. K. 2003, ApJS, 145, 15 Smith, J. A. et al. 2002, AJ, 123, 2121 Stoughton, C. et al. 2002, AJ, 123, 485 Strauss, M. A. et al. 2002, AJ, 124, 1810 Sulentic, J. W., Marziani, P., & Dultzin-Hacyan, D. 2000, ARA&A, 38, 521 Tr`vese, D., Kron, R. G., & Bunone, A. 2001, ApJ, 551, 103 e Tyson, J. A., et al. 1998, AJ, 116, 102 Tytler, D. & Fan, X. 1992, ApJS, 79, 1 Uomoto, A. K., Wills, B. J., & Wills, D. 1976, AJ, 81, 905 Vanden Berk, D. E. et al. 2001, AJ, 122, 549 Vanden Berk, D. E. et al. 2004, ApJ, 601, 692 Vestergaard, M., & Wilkes, B. J. 2001, ApJS, 134, 1 Voit, G. M., Shull, J. m., & Begelman, M.C. 1987, ApJ, 316, 573 Wandel, A., Peterson, B. M., & Malkan, M. A. 1999, ApJ, 526, 579 Wanders, I. & Peterson, B. M. 1996, ApJ, 466, 174

– 25 – Wang, T., Lu, Y., & Zhou, Y. 1998, ApJ, 493, 1 Wilhite, B. C. et al. 2005, ApJ, 633, 638 Wilkes, B. J. 1984, MNRAS, 207, 73 Wills, B. J., Fang, D., & Brotherton, M. S. 1992, Physics of Active Galactic Nuclei, 264 Wills, B. J., Brotherton, M. S., Fang, D., Steidel, C. C., & Sargent, W. L. W. 1993, ApJ, 415, 563 York, D. G. et al. 2000, AJ, 120, 1579

A This preprint was prepared with the AAS L TEX macros v5.2.

– 26 –

Fig. 1.— Distribution of Pearson skewness for quasars at high-S/N epoch. The median skewness is 0.012 ± 0.013.

– 27 –

Fig. 2.— Change in C iv line width versus line ?ux change. The Spearman rank correlation coe?cient for this distribution is 0.42 with a signi?cance of 2.3 × 10?5.

– 28 –

Fig. 3.— Change in C iv line center versus line ?ux change. The Spearman rank correlation signi?cance for this distribution is 0.86.

– 29 –

Fig. 4.— Change in C iv line Pearson skewness versus line ?ux change. The Spearman rank correlation signi?cance for this distribution is 0.34.

– 30 –

Fig. 5.— Continuum ?ux (upper panel) and C iv line ?ux (lower panel) variability (V) as a function of rest-frame time lag (?τ ). The overlaid binned points correspond to the well-known Structure Function.

– 31 –

Fig. 6.— Continuum ?ux (upper panel) and C iv line ?ux (lower panel) variability (V) as a function of r-band absolute magnitude. Overlaid are the binned values for V and Mr . Large errors in the individual values of V in the lowest-luminosity bin make the binned measurements of V unreliable.

– 32 –

Fig. 7.— C iv Line ?ux versus continuum ?ux (upper panel) for objects at high-S/N epoch display the familiar Baldwin E?ect. (lower panel) Histogram of C iv Intrinsic Baldwin E?ect slopes. The median slope (after untrustworthy outliers are removed) bint = 0.15 ± 0.06.

– 33 –

Fig. 8.— Upper panel: High-S/N epoch velocity shift between C iv and Mg ii line centers. Positive velocities indicate a C iv blueshift. Lower panel: Di?erence in C iv-Mg ii line shift between the high- and low-S/N epochs.

– 34 –

Fig. 9.— Upper panel: Same as Fig. 8, for objects with fline,HSN > 800 × 10?17 erg/s/cm2 .

– 35 –

Fig. 10.— C iv line for quasar SDSS J081931.48+055523.6 at the high-S/N epoch (dark spectrum) and the low-S/N epoch (light). Vertical lines indicate median ?ts to line centers.

– 36 –

Fig. 11.— C iv line for quasar SDSS J100013.37+011203.2.

– 37 –

Fig. 12.— C iv line for quasar SDSS J231147.90+002941.9.

– 38 –

Fig. 13.— C iv line for quasar SDSS J160126.31+511038.1.

– 39 –

Fig. 14.— C iv line for quasar SDSS J115154.83?005904.6. The bifurcation in the line is likely due to highly variable absorption.

Table 1. CIV Variability Sample. HSN and LSN indicate the high- and low-S/N ratio epochs, respectively. Redshifts and magnitudes are from the high-S/N ratio epoch. Number SDSS J MJD HSN LSN 51910 51910 51959 51959 51943 51943 51943 51928 51928 51928 51689 51689 51985 51984 51959 51955 51942 51942 51942 51942 51942 51609 51609 51994 51994 51581 51581 51584 51584 51662 51662 51662 51660 51660 51660 51994 51994 51585 51665 51663 51662 51641 51641 51641 51641 51641 51957 51957 51666 51666 zHSN ?τ (days) 117.4 81.3 128.5 126.3 96.0 94.4 99.2 86.9 94.5 84.8 110.1 109.5 117.5 115.0 94.1 93.7 85.5 101.6 104.3 112.7 61.1 121.7 74.3 111.3 116.8 rHSN Mr,HSN S/Nr HSN LSN 12.3 16.4 22.7 12.9 12.9 40.0 8.0 29.7 9.8 14.2 7.9 8.5 18.3 8.2 19.6 10.4 14.2 16.7 12.6 6.3 9.5 13.4 8.3 9.3 27.6 11.4 11.2 16.4 10.9 7.6 27.2 7.3 20.8 7.2 9.4 5.5 10.3 11.7 6.8 15.0 4.6 12.8 18.0 8.9 4.5 7.1 11.2 9.1 5.5 17.9

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

100013.37+011203.2 100428.43+001825.6 114211.59?005344.2 114948.81+000855.8 115154.83?005904.6 115043.87?002354.0 115213.55+001946.7 124524.59?000937.9 124356.22?000021.8 124242.11+001157.9 125617.52?001918.2 125532.24?010608.7 131630.46+005125.5 131840.95+003103.9 132214.82+005419.9 133939.01+001021.6 140114.28?004537.1 135844.57?011055.1 135605.41?010024.4 135247.96?002351.6 135828.74+005811.5 142205.10?000120.7 142209.11+005436.3 145555.00?003713.4 145302.09?010524.4

1.80 3.04 1.92 1.97 1.93 1.98 1.83 2.08 1.84 2.16 1.77 1.78 2.40 1.77 2.15 2.13 2.52 1.96 1.89 1.67 3.92 1.86 3.68 1.95 1.81

19.1 18.7 20.1 19.4 18.3 19.7 19.6 19.5 19.4 18.4 18.1 18.1 17.5 18.9 18.9 20.3 18.8 19.1 19.2 18.8 19.4 19.4 17.5 19.1 19.3

?26.0 ?27.6 ?26.8 ?26.2 ?26.0 ?27.8 ?25.3 ?27.3 ?25.5 ?26.3 ?25.1 ?25.2 ?27.2 ?25.3 ?26.8 ?26.1 ?26.7 ?26.3 ?25.8 ?24.7 ?27.1 ?26.0 ?26.9 ?25.4 ?26.9

– 40 –

Table 1—Continued Number SDSS J MJD HSN LSN 51994 51994 51691 51694 51817 51817 51817 51816 51816 51817 51931 51929 51929 51811 51817 51817 51817 51868 52314 52314 52314 52314 52314 52295 52295 51666 51666 51990 51789 52238 52238 52238 51877 51877 51873 52254 51821 51821 51885 51884 51884 51884 51812 52027 52027 52027 52027 52027 52029 52029 zHSN ?τ (days) 92.6 89.6 68.3 34.6 153.9 157.4 146.5 21.5 19.1 17.5 117.1 37.5 37.5 26.9 19.9 23.4 23.4 18.8 103.4 102.9 101.2 101.1 95.6 99.8 95.8 rHSN Mr,HSN S/Nr HSN LSN 2.8 14.1 8.5 10.6 4.0 11.9 6.3 15.4 2.4 14.6 11.4 2.9 10.4 2.9 3.0 4.8 4.0 3.5 3.3 8.3 4.1 8.8 3.1 3.5 6.3 4.0 9.7 9.1 8.2 2.8 15.3 3.7 17.3 3.3 10.8 8.9 3.8 7.8 1.5 2.3 3.2 3.0 3.7 2.7 5.1 2.8 8.3 3.5 3.1 5.7

26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50

145246.52+003450.5 145429.65+004121.2 131728.74?024759.4 172909.93+624519.7 022534.09+000347.9 022346.42?003908.2 022526.15+010124.0 025038.68?004739.1 025701.94+010644.6 030600.41+010145.4 031127.55+005357.3 032253.09?001121.6 031544.54+004220.9 034318.37?004447.9 003732.61+144258.0 003520.91+143730.2 003240.57+143951.9 004337.73+160530.0 093622.06?004555.4 093736.74?000732.1 093233.65?003441.9 093150.57?001935.2 094149.60+003254.3 131522.44+013917.0 131439.23+021214.9

2.54 2.66 3.38 1.75 1.74 1.67 1.87 1.84 2.19 2.19 1.76 1.88 1.88 1.75 2.37 1.86 1.86 1.97 1.78 1.79 1.84 1.84 2.00 1.67 1.78

19.8 19.5 17.8 19.5 18.8 18.4 18.4 18.2 19.6 19.2 19.9 19.2 18.8 18.3 19.4 18.6 19.9 19.8 18.7 19.7 19.8 17.6 18.2 18.2 20.3

?24.6 ?26.7 ?26.7 ?25.8 ?24.2 ?25.6 ?24.9 ?26.0 ?24.5 ?26.7 ?25.5 ?24.2 ?25.6 ?23.7 ?24.5 ?24.5 ?24.3 ?24.6 ?24.3 ?25.3 ?24.7 ?25.6 ?24.4 ?24.1 ?24.8

– 41 –

Table 1—Continued Number SDSS J MJD HSN LSN 52295 52295 52295 52295 52207 52207 52207 52207 52207 52374 52375 52884 52884 52884 52884 52884 52884 52884 52884 52884 52884 52884 52884 52884 52884 52029 52029 52029 52029 51959 51959 51959 51959 51959 52442 52081 52534 52534 52534 52534 52534 52534 52534 52534 52534 52534 52534 52534 52534 52534 zHSN ?τ (days) 92.7 95.5 70.8 96.6 77.0 77.1 83.8 91.4 89.2 25.0 103.4 114.2 70.4 117.4 85.7 100.2 92.8 121.1 122.2 123.0 74.6 116.6 114.3 120.6 121.0 rHSN Mr,HSN S/Nr HSN LSN 7.9 9.8 5.8 10.9 3.7 13.6 5.2 9.3 13.9 13.8 9.4 8.7 6.7 6.4 2.6 5.8 3.3 3.0 3.1 6.5 4.3 3.4 10.4 9.5 15.9 4.4 7.7 3.3 4.2 2.3 11.4 4.8 6.8 12.6 8.7 3.7 7.2 4.5 5.6 1.1 4.5 2.4 2.6 3.1 2.7 3.6 2.8 5.6 7.1 7.7

51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75

130754.44+021820.2 130855.25+030614.2 130940.60+031826.7 130825.64+025736.0 081859.78+423327.6 081349.01+441517.7 081614.97+435640.2 081926.51+445759.9 082310.94+442048.1 153001.69+540452.3 160126.31+511038.1 231040.97?010823.0 230952.29?003138.9 230728.90?011608.9 230832.98?002332.4 230437.65?005703.3 230402.78?003855.4 230424.87?010140.8 230239.68+002702.5 230524.47+005209.7 230323.77+001615.1 230435.93+003001.5 231121.98+004959.7 231147.90+002941.9 231241.77+002450.3

1.87 1.79 2.76 1.75 2.22 2.21 1.96 1.71 1.78 1.72 1.84 2.07 3.97 1.98 3.08 2.49 2.77 1.89 1.86 1.85 3.69 2.00 2.06 1.90 1.89

19.0 19.3 19.1 19.0 19.4 18.6 18.2 19.8 21.6 19.0 18.8 19.3 19.7 19.6 18.6 19.5 17.8 19.8 19.4 18.9 18.2 19.2 19.5 19.4 18.2

?25.2 ?25.3 ?25.7 ?25.5 ?24.4 ?26.0 ?24.5 ?25.0 ?25.5 ?25.9 ?25.4 ?25.8 ?26.8 ?25.1 ?25.1 ?25.5 ?25.3 ?24.2 ?24.2 ?25.1 ?26.1 ?24.6 ?25.8 ?25.5 ?26.2

– 42 –

Table 1—Continued Number SDSS J MJD HSN LSN 52520 52237 52237 52237 52237 52237 52346 52346 52347 52347 52443 52347 52347 52346 52346 52346 52346 52346 52346 52346 52346 52346 52346 52962 52962 52264 52577 52577 52577 52577 52577 52433 52433 52435 52435 52355 52674 52674 52669 52669 52669 52669 52669 52669 52663 52663 52663 52663 52737 52737 zHSN ?τ (days) 57.0 116.1 126.7 85.2 110.9 117.4 28.6 32.8 30.9 31.8 32.1 82.8 111.2 120.5 64.6 112.8 115.7 116.0 112.4 96.8 110.8 115.9 116.2 72.5 59.5 rHSN Mr,HSN S/Nr HSN LSN 4.8 13.9 5.8 6.3 4.3 3.8 5.9 3.9 2.7 8.9 4.0 9.7 17.5 2.7 4.8 6.2 7.4 3.8 19.1 11.8 10.9 27.8 3.7 13.2 24.9 1.1 7.6 5.1 9.1 2.8 5.0 5.7 2.3 2.5 7.0 0.7 7.8 16.9 5.9 8.5 11.1 9.6 3.6 12.4 10.0 14.0 21.7 4.8 9.0 14.6

76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100

224005.09+143147.8 075153.67+331319.8 075217.23+335524.5 074823.86+332051.2 075132.75+350535.0 075321.93+350733.5 144059.16+573724.3 143618.60+581044.2 145316.61+560750.8 143632.31+563319.5 161240.98+435749.4 101902.02+473714.5 102048.82+483908.8 105922.46+494918.2 105430.08+491947.1 105027.74+490453.0 104951.09+493156.2 104806.47+501021.5 105454.16+503123.9 074641.95+293247.9 074407.41+294707.4 074625.28+302020.7 074937.74+304021.4 082443.39+055503.7 082328.61+061146.0

3.49 1.93 1.68 2.99 2.07 1.90 2.04 1.65 1.85 1.77 1.74 2.95 1.94 1.68 4.00 1.86 1.79 1.78 1.87 2.28 1.86 1.74 1.73 2.10 2.78

19.9 20.9 19.0 19.0 19.4 17.3 20.0 19.1 18.2 18.8 19.0 19.4 20.2 18.8 17.7 20.6 19.5 19.7 19.1 18.8 19.0 19.4 19.2 19.4 19.1

?26.1 ?26.1 ?24.8 ?26.2 ?24.7 ?24.5 ?25.5 ?24.1 ?24.2 ?25.6 ?24.3 ?26.9 ?26.5 ?24.5 ?27.1 ?25.6 ?25.3 ?24.2 ?26.3 ?26.2 ?25.8 ?26.6 ?24.1 ?26.5 ?27.9

– 43 –

Table 1—Continued Number SDSS J MJD HSN LSN 52962 52962 52962 52962 52962 52737 52737 52737 52737 52737 zHSN ?τ (days) 75.4 83.3 83.7 64.1 56.9 rHSN Mr,HSN S/Nr HSN LSN 9.4 2.8 21.1 17.7 17.5 4.2 2.3 12.6 11.2 10.2

101 102 103 104 105

082256.01+060528.7 081941.12+054942.6 081931.48+055523.6 081811.50+053713.9 082257.04+070104.3

1.98 1.70 1.69 2.51 2.95

19.7 19.8 19.9 19.0 18.2

?25.7 ?24.0 ?26.6 ?27.2 ?27.5

– 44 –

– 45 – Table 2. CIV Flux Variability. Number fcont (10 erg/s/cm2 /?) A HSN LSN
?17

fline (10 erg/s/cm2 ) HSN LSN
?17

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33

11.2 8.0 19.6 9.7 10.1 50.3 5.8 20.7 7.6 8.9 5.9 4.3 12.0 4.5 12.9 6.3 6.7 10.3 8.1 4.5 2.6 10.2 2.6 5.3 24.1 0.7 5.4 2.7 8.1 2.3 8.5 4.1 9.3

20.3 6.4 27.5 13.5 7.1 44.9 8.7 17.3 4.8 6.3 3.3 6.5 9.9 6.4 15.0 3.6 8.2 14.3 4.5 3.0 2.1 8.3 1.5 3.6 19.4 1.9 4.5 3.3 7.0 1.3 14.5 2.4 14.9

1205.1 1248.3 1458.3 746.4 614.6 4691.5 712.1 1991.0 760.7 784.8 514.7 374.1 1943.9 771.7 918.3 722.0 496.0 600.1 386.1 423.3 406.8 958.0 109.5 408.2 2014.2 200.4 898.1 655.0 1335.1 466.1 359.0 311.5 748.2

1099.3 1165.5 1644.1 654.4 479.6 3729.1 853.0 2013.4 702.2 925.1 513.1 484.6 1937.0 780.5 971.4 864.0 657.7 624.7 386.2 517.9 306.3 917.8 822.0 493.3 1861.1 269.8 736.5 591.7 1290.9 489.1 374.4 297.3 1071.2

– 46 – Table 2—Continued Number fcont (10 erg/s/cm2 /?) A HSN LSN
?17

fline (10 erg/s/cm2 ) HSN LSN
?17

34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66

1.1 10.1 6.6 2.0 6.5 1.0 1.2 2.3 1.8 2.2 2.4 4.2 3.1 5.8 1.2 2.3 3.8 4.2 5.1 2.0 8.3 1.1 5.0 2.4 4.4 7.0 12.5 6.5 4.7 2.0 2.7 0.9 2.4

1.3 8.0 4.9 2.1 4.9 0.1 0.9 1.8 1.4 2.3 2.0 3.2 2.2 8.1 1.6 3.2 5.7 2.5 6.7 1.4 4.1 0.9 6.5 2.6 3.4 8.7 10.3 4.1 6.6 2.3 3.9 0.8 3.0

81.2 926.0 341.4 365.1 483.7 390.5 220.6 395.6 326.1 374.8 330.4 302.9 384.0 1217.8 294.4 173.6 451.6 437.8 584.1 77.7 411.7 81.8 507.5 213.4 499.1 769.1 572.4 958.7 459.6 218.3 120.5 20.0 275.3

160.4 717.4 432.5 388.5 492.2 244.4 209.5 331.1 304.9 451.4 242.5 241.1 319.5 1203.8 409.3 169.1 457.0 437.6 735.5 44.9 352.2 76.6 458.2 219.6 457.6 857.1 493.5 1038.7 560.6 297.8 355.3 27.6 260.5

– 47 – Table 2—Continued Number fcont (10 erg/s/cm2 /?) A HSN LSN
?17

fline (10 erg/s/cm2 ) HSN LSN
?17

67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99

1.0 1.8 2.3 4.2 0.9 1.7 6.3 6.9 11.0 1.5 8.0 5.4 2.4 1.7 2.2 4.2 2.6 2.0 7.4 3.0 5.3 15.6 4.1 2.0 7.1 6.3 2.0 14.8 6.6 7.2 21.8 1.4 10.7

1.9 1.7 2.6 2.2 1.1 2.8 5.3 8.8 6.9 0.8 4.1 3.5 3.5 2.4 3.1 5.5 2.7 0.9 3.6 2.7 3.7 18.9 5.2 2.5 10.8 8.1 3.4 11.1 5.2 9.7 16.6 2.2 8.4

359.2 278.1 274.3 452.6 368.5 377.8 445.3 1452.4 689.3 275.1 417.1 413.2 427.7 173.6 249.6 541.7 356.9 507.8 764.6 395.5 405.5 1497.7 483.2 441.3 927.3 796.7 555.5 1135.4 38.5 501.5 3051.4 245.2 898.4

380.5 357.3 329.3 631.8 332.8 312.5 326.8 1403.6 572.7 329.0 377.8 456.2 254.7 167.5 262.4 547.4 157.1 507.9 604.3 351.4 433.5 1537.6 698.9 790.4 973.0 775.6 421.5 1212.0 78.8 421.5 2626.4 218.1 863.8

– 48 –

Table 2—Continued Number fcont (10 erg/s/cm2 /?) A HSN LSN
?17

fline (10 erg/s/cm2 ) HSN LSN
?17

100 101 102 103 104 105

16.3 6.3 2.1 28.0 11.3 8.9

13.8 4.3 3.9 15.2 13.6 6.8

1383.6 456.6 786.5 1447.8 884.4 786.6

1104.5 412.7 602.5 1044.2 930.1 917.3

– 49 – Table 3. CIV Pro?le Variability. Number λmedian (?) A HSN LSN 4335.96 6254.13 4504.86 4587.67 4528.83 4609.86 4386.77 4755.69 4386.39 4885.17 4283.01 4320.61 5267.12 4292.50 4861.69 4836.85 5436.41 4577.88 4473.25 4132.72 7611.62 4426.41 7187.91 4560.30 4339.19 5473.89 5651.77 6796.05 4252.94 4229.30 4128.10 4446.01 4390.24 σCIV (?) A HSN LSN P skew HSN LSN

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33

4336.67 51.8 50.3 -0.02 -0.04 6255.22 60.2 55.8 -0.16 -0.14 4505.81 54.3 54.0 -0.00 -0.03 4591.58 55.5 45.8 -0.03 -0.07 4525.40 60.7 54.8 -0.08 -0.19 4610.34 51.4 49.1 -0.11 -0.10 4384.37 50.6 53.3 -0.15 -0.07 4755.93 60.9 61.7 0.05 0.04 4389.90 47.8 47.8 -0.13 -0.13 4884.34 46.9 55.8 -0.04 -0.04 4280.57 41.5 46.1 0.03 -0.08 4321.66 53.9 53.3 0.02 -0.01 5266.23 55.5 54.9 -0.07 -0.10 4294.13 56.6 50.0 -0.03 0.05 4864.76 66.6 60.0 -0.04 -0.03 4836.54 47.4 59.8 -0.01 -0.09 5440.46 68.1 71.6 0.10 0.10 4567.58 60.2 55.9 -0.12 0.22 4472.83 52.3 52.4 -0.11 0.04 4133.57 46.2 40.6 -0.04 -0.00 7608.21 109.3 97.5 0.09 0.14 4428.08 46.5 46.6 -0.09 -0.10 7215.40 54.9 127.7 0.28 0.19 4558.58 49.8 55.2 -0.01 0.00 4337.84 51.0 51.2 0.02 0.05 5483.46 72.3 73.9 -0.11 -0.10 5652.11 62.4 49.1 0.07 0.03 6797.81 88.2 81.0 -0.34 -0.35 4251.77 43.1 42.7 0.14 0.11 4228.77 49.7 50.4 -0.07 -0.11 4125.95 44.5 54.5 0.06 0.15 4445.38 46.6 52.5 -0.03 -0.18 4392.37 59.3 53.7 -0.03 -0.10

– 50 – Table 3—Continued Number λmedian (?) A HSN LSN 4917.46 4932.10 4270.86 4458.80 4451.02 4251.56 5200.92 4424.24 4428.53 4605.88 4295.65 4313.70 4383.73 4390.96 4661.90 4108.36 4294.97 4434.07 4305.35 5818.30 4259.49 4992.64 4960.77 4581.82 4189.89 4301.59 4203.21 4402.60 4727.93 7668.57 4633.27 6341.47 5407.90 4926.01 4934.35 4269.77 4457.78 4454.29 4253.38 5204.93 4425.90 4430.25 4604.88 4295.44 4314.42 4390.22 4391.03 4663.15 4134.69 4290.71 4434.83 4306.14 5748.08 4261.48 4986.64 4958.50 4574.54 4187.46 4302.12 4201.92 4402.21 4733.90 7671.34 4631.65 6312.69 5412.07 σCIV (?) A HSN LSN ··· 60.9 50.9 53.4 46.2 56.1 65.9 58.2 53.4 50.5 42.3 56.0 53.6 48.9 54.9 53.2 47.3 54.3 53.5 80.1 50.5 63.4 62.3 59.2 48.4 48.2 62.3 43.5 62.4 66.2 ··· ··· 69.6 80.8 59.4 54.6 54.9 49.3 48.3 62.9 53.1 54.4 58.3 35.6 61.3 46.4 42.9 48.5 45.9 49.7 56.1 51.3 177.3 39.0 57.8 59.2 60.9 52.4 48.8 63.4 44.6 68.4 64.7 63.9 ··· 54.1 P skew HSN ··· -0.01 -0.21 -0.02 0.06 -0.08 0.07 -0.16 -0.00 0.07 -0.14 0.01 -0.05 0.03 -0.21 0.31 0.03 0.10 0.09 0.07 0.17 -0.14 0.15 -0.01 0.04 0.01 0.09 0.13 0.09 -0.11 ··· ··· -0.14 LSN 0.00 -0.07 -0.20 -0.12 -0.03 -0.05 0.16 -0.25 -0.13 0.00 0.51 -0.03 0.06 0.09 -0.01 -0.31 0.05 0.17 0.11 -2.35 0.35 -0.22 0.20 -0.07 -0.00 -0.05 0.19 0.05 0.22 -0.41 -0.09 ··· -0.30

34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66

– 51 – Table 3—Continued Number λmedian (?) A HSN LSN 5836.82 4482.85 4428.81 4404.43 7269.31 4651.40 4719.64 4492.13 4480.50 6964.24 4542.01 4147.37 6148.37 4738.29 4488.77 4688.73 4105.50 4402.43 4269.64 4243.85 6087.18 4546.91 4141.26 7723.83 4431.05 4316.52 4301.31 4449.11 5050.78 4418.00 4229.98 4204.25 4788.89 5838.44 4482.89 4423.38 4403.24 7266.72 4650.14 4720.61 4492.69 4482.09 6969.20 4541.62 4149.49 6148.65 4745.81 4490.48 4694.29 4104.59 4400.79 4273.86 4244.72 6091.37 4547.56 4150.70 7725.32 4431.47 4314.17 4306.88 4449.00 5056.25 4419.19 4229.32 4216.29 4787.13 σCIV (?) A HSN LSN 73.2 47.2 50.5 53.9 79.8 64.1 64.7 54.4 52.3 82.1 48.4 41.7 85.2 63.6 54.2 64.8 53.8 40.4 57.8 61.3 70.0 47.3 41.2 ··· 47.7 53.0 49.4 52.3 ··· 61.1 48.8 56.7 47.6 64.3 57.0 57.6 61.9 89.4 52.9 56.9 52.9 52.9 95.3 45.9 46.3 73.4 64.6 47.5 57.7 ··· 44.1 63.2 58.9 81.6 51.2 47.3 73.8 48.1 49.1 41.4 56.6 59.1 54.0 48.8 64.4 53.2 P skew HSN -0.21 -0.35 0.04 0.05 -0.13 -0.13 0.11 -0.05 0.15 -0.14 -0.07 -0.12 0.28 0.09 -0.17 0.14 -0.10 0.00 0.06 0.12 0.19 0.05 -0.02 ··· 0.04 -0.05 0.09 -0.02 ··· -0.06 -0.06 0.34 -0.08 LSN -0.13 -0.08 0.10 0.06 -0.14 0.03 0.11 0.01 -0.04 -0.36 -0.05 -0.06 0.45 0.02 -0.10 0.04 ··· -0.11 -0.00 -0.16 0.02 0.06 -0.22 0.05 0.08 -0.19 0.33 -0.05 -0.63 -0.02 -0.05 0.03 -0.10

67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99

– 52 –

Table 3—Continued Number λmedian (?) A HSN LSN 5855.95 4614.60 4172.34 4154.89 5437.31 6102.28 5859.41 4614.45 4174.75 4153.73 5437.47 6105.53 σCIV (?) A HSN LSN 70.1 45.4 66.4 46.4 65.5 73.5 60.7 52.1 67.5 41.5 67.4 77.5 P skew HSN LSN

100 101 102 103 104 105

-0.20 -0.23 -0.02 0.01 0.06 0.03 -0.21 -0.11 0.22 0.18 0.03 -0.04



更多相关文章:
类星体SDSS J024304.68+000005.4光谱Mini-BAL光变分析.pdf
Keywords Quasar Spectrum, Absorption Line, Variation...(Sloan Digital Sky Survey, SDSS)为我们带来了极...(2012) Variability of Mini-BAL and BAL Outflows...
The Power of Exploratory Chandra Observations.pdf
Sloan Digital Sky Survey (SDSS) quasars with ...Based on rapid variability of soft X-rays in ...(2002) found that the C ivMg ii velocity ...
...PG1159 stars from the Sloan Digital Sky Survey_图文.pdf
Spectral analyses of DO white dwarfs and PG1159 ... discovered in the Sloan Digital Sky Survey DR?...variability driven by cyclic ionisation of carbon ...
chapter-08_图文.ppt
3C 279 Rapid and violent variability within days ...Sloan Digital Sky Survey Z = 5.00 Z = 4.90...Lyman-Alpha Forest in the spectrum of quasar QSO...
High-Redshift Quasars Found in Sloan Digital Sky Su....pdf
Doi12 , Masataka Fukugita13,3 , Zeljko Ivezi?...quasars found in the Sloan Digital Sky Survey (...at a spectral resolution of 1800, and have been...
chapter08-simple_图文.ppt
3C 279 Rapid and violent variability within days ...Sloan Digital Sky Survey Z = 5.00 Z = 4.90...Lyman-Alpha Forest in the spectrum of quasar QSO...
The science archive for the Sloan Digital Sky Survey.pdf
archive for the Sloan Digital Sky Survey_专业资料...variability and deeper imaging in the southern ... one hundred thousand quasars, and about ten ...
Large Scale Structure in the Sloan Digital Sky Surv....pdf
Scale Structure in the Sloan Digital Sky Survey...variability studies in at least one of these ...1.2. The angular power spectrum C? in di?...
The UV Continuum of Quasars Models and SDSS Spectra....pdf
wavelength spectral slopes α (Fν ∝να ) A for quasar spectra from the Sloan Digital Sky Survey. The long and short wavelength slopes are computed ...
BL Lac Objects in the Sloan Digital Sky Survey (SDSS).pdf
BL Lac Objects in the Sloan Digital Sky Survey ...variability at complete EM spectrum with the ...spectrum (BL Lac objects), (ii) spectrum with ...
...of a Field Methane Dwarf from Sloan Digital Sky ....pdf
Dwarf from Sloan Digital Sky Survey Commissioning ...nd 15 new quasars at z > 3.65 (including ...0.0014?m, and the spectral coverage is 0.4-...
附录4-变星的名字.pdf
该 项目主页 http://www.2dfquasar.org/index....暗弱变源巡天(Faint Sky Variability Survey)发现的...斯隆数字巡天(the Sloan Digital Sky Survey) 发现...
...New L and T Dwarfs from the Sloan Digital Sky Su....pdf
Candidate Type II Quasar... 暂无评价 46页 ...and T Dwarfs from the Sloan Digital Sky Survey...spectral properties that may eventually help to ...
An accretion disc model for quasar optical variability.pdf
000 quasars from the Sloan Digital Sky Survey, which is different from ...ed disc model, the spectral calculation show that the optical variability ...
...Resolution in the NOAO Deep Wide-Field Survey.pdf
ed, mostly with radio galaxies and quasars (Waddington et al. 2000). By...ed to 22 mag in the Sloan Digital Sky Survey (Ivezi? et al. 2002)....
The Shape of the Big Blue Bump as Revealed by Spectro....pdf
AGN Physics with the Sloan Digital Sky Survey ...variability, spectral energy distributions and other...Next we observed eleven quasars at the VLT Unit...
Broad absorption line quasars have the same cool dust ....pdf
AGN Physics with the Sloan Digital Sky Survey ...broad C iv absorption line quasars is discussed....dust spectral energy distributions, we will use ...
...XMM-Newton observations of the first quasars X-r....pdf
taken from the Sloan Digital Sky Survey (SDSS).... C IV) in their high-quality optical spectrum....2.1.4. X-ray Variability of the Quasars ...
PHL 1811 The Local Prototype of the Lineless High-z SDSS QSOs....pdf
Highionization lines are very weak; C IV has ...1. Introduction In the Sloan Digital Sky Survey,...spectrum of one of the high-z lineless quasars...
...from the ROSAT All-Sky and Sloan Digital Sky_免....pdf
Sky and Sloan Digital Sky Surveys: the Data ...quasars with strong, broad emission lines; (ii)...strong soft X-ray excesses and marked variability...
更多相关标签:

All rights reserved Powered by 甜梦文库 9512.net

copyright ©right 2010-2021。
甜梦文库内容来自网络,如有侵犯请联系客服。zhit325@126.com|网站地图