• The LIGO detection of gravitational waves (GW) from merging black holes in 2015 marked the beginning of a new era in observational astronomy. The detection of an electromagnetic signal from a GW source is the critical next step to explore in detail the physics involved. The Antarctic Survey Telescopes (AST3), located at Dome A, Antarctica, is uniquely situated for rapid response time-domain astronomy with its continuous night-time coverage during the austral winter. We report optical observations of the GW source (GW~170817) in the nearby galaxy NGC 4993 using AST3. The data show a rapidly fading transient at around 1 day after the GW trigger, with the $i$-band magnitude declining from $17.23\pm0.13$ magnitude to $17.72\pm0.09$ magnitude in $\sim 1.8$ hour. The brightness and time evolution of the optical transient associated with GW~170817 are broadly consistent with the predictions of models involving merging binary neutron stars. We infer from our data that the merging process ejected about $\sim 10^{-2}$ solar mass of radioactive material at a speed of up to $30\%$ the speed of light.
  • AST3-1 is the second-generation wide-field optical photometric telescope dedicated to time domain astronomy at Dome A, Antarctica. Here we present the results of $i$ band images survey from AST3-1 towards one Galactic disk field. Based on time-series photometry of 92,583 stars, 560 variable stars were detected with $i$ magnitude $\leq$ 16.5 mag during eight days of observations; 339 of these are previously unknown variables. We tentatively classify the 560 variables as 285 eclipsing binaries (EW, EB, EA), 27 pulsating variable stars ($\delta$~Scuti, $\gamma$~Doradus, $\delta$~Cephei variable and RR Lyrae stars) and 248 other types of variables (unclassified periodic, multi-periodic and aperiodic variable stars). Among the eclipsing binaries, 34 show O'Connell effects. One of the aperiodic variables shows a plateau light curve and another one shows a secondary maximum after peak brightness. We also detected a complex binary system with RS CVn-like light curve morphology; this object is being followed-up spectroscopically using the Gemini South telescope.
  • The terahertz and far-infrared (FIR) band, from approximately 0.3 THz to 15 THz (1 mm to 20 micron), is important for astrophysics as the thermal radiation of much of the universe peaks at these wavelengths and many spectral lines that trace the cycle of interstellar matter also lie within this band. However, water vapor renders the terrestrial atmosphere opaque to this frequency band over nearly all of the Earth's surface. Early radiometric measurements below 1 THz at Dome A, the highest point of the cold and dry Antarctic ice sheet, suggest that it may offer the best possible access for ground-based astronomical observations in the terahertz and FIR band. To address uncertainty in radiative transfer modelling, we carried out measurements of atmospheric radiation from Dome A spanning the entire water vapor pure rotation band from 20 micron to 350 micron wavelength by a Fourier transform spectrometer. Our measurements expose atmospheric windows having significant transmission throughout this band. Furthermore, by combining our broadband spectra with auxiliary data on the atmospheric state over Dome A, we set new constraints on the spectral absorption of water vapor at upper tropospheric temperatures important for accurately modeling the terrestrial climate. In particular, we find that current spectral models significantly underestimate the H2O continuum absorption.
  • The masses of the black holes powering quasars represent a fundamental parameter of active galaxies. Estimates of quasar black hole masses using single-epoch spectra are quite uncertain, and require quantitative improvement. We recently identified a correction for C IV $\lambda$1549-based scaling relationships used to estimate quasar black hole masses that relies on the continuum-subtracted peak flux ratio of the ultraviolet emission-line blend Si IV + OIV] (the $\lambda$1400 feature) to that of C IV. This parameter correlates with the suite of associated quasar spectral properties collectively known as "Eigenvector 1" (EV1). Here we use a sample of 85 quasars with quasi-simultaneous optical-ultraviolet spectrophotometry to demonstrate how biases in the average EV1 properties can create systematic biases in C IV-based black hole mass scaling relationships. This effect results in nearly an order of magnitude moving from objects with small $<$ peak $\lambda$1400/C IV $>$, which have overestimated black hole masses, to objects with large $<$ peak $\lambda1400$/C IV $>$, which have underestimated values. We show that existing reverberation-mapped samples of quasars with ultraviolet spectra -- used to calibrate C IV-based scaling relationships -- have significant EV1 biases that result in predictions of black hole masses nearly 50\% too high for the average quasar. We offer corrections and suggestions to account for this bias.
  • We present an analysis of the meteorological data collected at Dome A, Antarctica by the Kunlun Automated Weather Station, including temperatures and wind speeds at eight elevations above the snow surface between 0m and 14.5m. The average temperatures at 2m and 14.5m are $-54^{\circ}$C and $-46^{\circ}$C, respectively. We find that a strong temperature inversion existed at all heights for more than 70% of the time, and the temperature inversion typically lasts longer than 25 hours, indicating an extremely stable atmosphere. The temperature gradient is larger at lower elevations than higher elevations. The average wind speed was 1.5m/s at 4m elevation. We find that the temperature inversion is stronger when the wind speed is lower and the temperature gradient decreases sharply at a specific wind speed for each elevation. The strong temperature inversion and low wind speed results in a shallow and stable boundary layer with weak atmospheric turbulence above it, suggesting that Dome A should be an excellent site for astronomical observations. All the data from the weather station are available for download.
  • We have developed a new method to correct dark current at relatively high temperatures for Charge-Coupled Device (CCD) images when dark frames cannot be obtained on the telescope. For images taken with the Antarctic Survey Telescopes (AST3) in 2012, due to the low cooling efficiency, the median CCD temperature was -46$^\circ$C, resulting in a high dark current level of about 3$e^-$/pix/sec, even comparable to the sky brightness (10$e^-$/pix/sec). If not corrected, the nonuniformity of the dark current could even overweight the photon noise of the sky background. However, dark frames could not be obtained during the observing season because the camera was operated in frame-transfer mode without a shutter, and the telescope was unattended in winter. Here we present an alternative, but simple and effective method to derive the dark current frame from the scientific images. Then we can scale this dark frame to the temperature at which the scientific images were taken, and apply the dark frame corrections to the scientific images. We have applied this method to the AST3 data, and demonstrated that it can reduce the noise to a level roughly as low as the photon noise of the sky brightness, solving the high noise problem and improving the photometric precision. This method will also be helpful for other projects that suffer from similar issues.
  • The photon transfer curve (PTC, variance vs. signal level) is a commonly used and effective tool in characterizing CCD performance. It is theoretically linear in the range where photon shot noise dominates, and its slope is utilized to derive the gain of the CCD. However, recent researches on different CCDs have revealed that the variance progressively drops at high signal levels, while the linearity shown by signal versus exposure time is still excellent and unaffected. On the other hand, bright stars are found to exhibit fatter point spread function (PSF). Both nonlinear PTC and the brighter-fatter effect are regarded as the result of spreading of charges between pixels, an interaction progress increasing with signal level. In this work we investigate the nonlinear PTC based on the images with a STA1600FT CCD camera, whose PTC starts to become nonlinear at about 1/3 full well. To explain the phenomenon, we present a model to characterize the charge-sharing PSF. This signal-dependent PSF can be derived from flat-field frames, and allow us to quantify the effects on photometry and measured shape of stars. This effect is essentially critical for projects requiring accurate photometry and shape parameters.
  • Twilight/night sky images are often used for flat-fielding CCD images, but the brightness gradient in twilight/night sky causes problems of accurate flat-field correction in astronomical images for wide-field telescopes. Using data from the Antarctic Survey Telescope (AST3), we found that when the sky brightness gradient is minimum and stable, there is still a gradient of 1% across AST3's field-of-view of 4.3 square degrees. We tested various approaches to remove the varying gradients in individual flat-field images. Our final optimal method can reduce the spatially dependent errors caused by the gradient to the negligible level. We also suggest a guideline of flat-fielding using twilight/night sky images for wide-field robotic autonomous telescopes.
  • We present results from a season of observations with the Chinese Small Telescope ARray (CSTAR), obtained over 183 days of the 2010 Antarctic winter. We carried out high-cadence time-series aperture photometry of 20,000 stars with i<15.3 mag located in a 23 square-degree region centered on the south celestial pole. We identified 188 variable stars, including 67 new objects relative to our 2008 observations, thanks to broader synoptic coverage, a deeper magnitude limit and a larger field of view. We used the photometric data set to derive site statistics from Dome A. Based on two years of observations, we find that extinction due to clouds at this site is less than 0.1 and 0.4 mag during 45% and 75% of the dark time, respectively.
  • We investigate the orientation dependence of the spectral energy distributions in a sample of radio-loud quasars. Selected specifically to study orientation issues, the sample contains 52 sources with redshifts in the range 0.16<z<1.4 and measurements of radio core dominance, a radio orientation indicator. Measured properties include monochromatic luminosities at a range of wavelengths between the infrared and X-rays, integrated infrared luminosity, spectral slopes, and the covering fraction of the obscuring circumnuclear dust. We estimate dust covering fraction assuming that the accretion disk emits anisotropically and discuss the shortcomings and technical difficulties of this calculation. Luminosities are found to depend on orientation, with face-on sources factors of a 2-3 brighter than more edge-on sources, depending on wavelength. The degree of anisotropy varies very little with wavelength such that the overall shape of the spectral energy distribution does not vary significantly with orientation. In the infrared, we do not observe a decrease in anisotropy with increasing wavelength. The spectral slopes and estimates of covering fraction are not significantly orientation dependent. We construct composite spectral energy distributions as a function of radio core dominance and find that these illustrate the results determined from the measured properties.
  • Currently, the ability to produce black hole mass estimates using the CIV 1549 line that are consistent with Hb mass estimates is uncertain, due in large part to disagreement between velocity line width measurements for the two lines. This discrepancy between Hb and CIV arises from the fact that both line profiles are treated the same way in single-epoch scaling relationships based on the assumption that the broad-line region is virialized, even though a non-virialized emission component is often significant in the CIV line and absent or weak in the Hb line. Using quasi-simultaneous optical and ultra-violet spectra for a sample of 85 optically bright quasars with redshifts in the range z=0.03-1.4, we present a significant step along the path to rehabilitating the CIV line for black hole mass estimates. We show that the residuals of velocity line width between CIV and Hb are significantly correlated with the peak flux ratio of SiIV+OIV] 1400 to CIV. Using this relationship, we develop a prescription for estimating black hole masses from the ultra-violet spectrum that agree better with Hb-based masses than the traditional CIV masses. The scatter between Hb and CIV masses is initially 0.43 dex in our sample and is reduced to 0.33 dex when using our prescription. The peak flux ratio of SiIV+OIV] 1400 to CIV is an ultraviolet indicator of the suite of spectral properties commonly known as "Eigenvector 1", thus the reduction in scatter between CIV and Hb black hole masses is essentially due to removing an Eigenvector 1 bias in CIV-based masses.
  • We present Spitzer InfraRed Spectrograph (IRS) low-resolution spectra of 16 spectroscopically selected post-starburst quasars (PSQs) at z ~ 0.3. The optical spectra of these broad-lined active galactic nuclei (AGNs) simultaneously show spectral signatures of massive intermediate-aged stellar populations making them good candidates for studying the connections between AGNs and their hosts. The resulting spectra show relatively strong polycyclic aromatic hydrocarbon (PAH) emission features at 6.2 and 11.3\micron and a very weak silicate feature, indicative of ongoing star formation and low dust obscuration levels for the AGNs. We find that the mid-infrared composite spectrum of PSQs has spectral properties between ULIRGs and QSOs, suggesting that PSQs are hybrid AGN and starburst systems as also seen in their optical spectra. We also find that PSQs in early-type host galaxies tend to have relatively strong AGN activities, while those in spiral hosts have stronger PAH emission, indicating more star formation.
  • We present optical spectroscopy of a sample of 38 post-starburst quasars (PSQs) at z ~ 0.3, 29 of which have morphological classifications based on Hubble Space Telescope imaging. These broad-lined active galactic nuclei (AGNs) possess the spectral signatures of massive intermediate-aged stellar populations making them potentially useful for studying connections between nuclear activity and host galaxy evolution. We model the spectra in order to determine the ages and masses of the host stellar populations, and the black hole masses and Eddington fractions of the AGNs. Our model components include an instantaneous starburst, a power-law, and emission lines. We find the PSQs have MBH ~ 10^8 Msun accreting at a few percent of Eddington luminosity and host ~ 10^10.5 Msun stellar populations which are several hundred Myr to a few Gyr old. We investigate relationships among these derived properties, spectral properties, and morphologies. We find that PSQs hosted in spiral galaxies have significantly weaker AGN luminosities, older starburst ages, and narrow emission-line ratios diagnostic of ongoing star-formation when compared to their early-type counterparts. We conclude that the early-type PSQs are likely the result of major mergers and were likely luminous infrared galaxies in the past, while spiral PSQs with more complex star-formation histories are triggered by less dramatic events (e.g., harassment, bars). We provide diagnostics to distinguish the early-type and spiral hosts when high spatial resolution imaging is not available.
  • Black hole masses are estimated for radio-loud quasars using several self-consistent scaling relationships based on emission-line widths and continuum luminosities. The emission lines used, H-beta, Mg II, and C IV, have different dependencies on orientation as estimated by radio core dominance. We compare differences in the log of black hole masses estimated from different emission lines and show that they depend on radio core dominance in the sense that core-dominated, jet-on objects have systematically smaller H-beta and Mg II determined masses compared to those from C IV, while lobe-dominated edge-on objects have systematically larger H-beta and Mg II determined masses compared to those from C IV. The effect is consistent with the H-beta line width, and to a lesser extent that of Mg II, being dependent upon orientation in the sense of a axisymmetric velocity field plus a projection effect. The size of the effect is nearly an order of magnitude in black hole mass going from one extreme orientation to the other. We find that radio spectral index is a good proxy for radio core dominance and repeating this analysis with radio spectral index yields similar results. Accounting for orientation could in principle significantly reduce the scatter in black hole mass scaling relationships, and we quantify and offer a correction for this effect cast in terms of radio core dominance and radio spectral index.
  • We present measurements and statistical properties of the optical and ultraviolet emission lines present in the spectra of 85 bright quasars which have detailed spectral energy distributions. This heterogeneous sample has redshifts up to z=1.5 and is comprised of three subsamples that may be of particular utility: ultraviolet excess Palomar-Green quasars, quasars with far-ultraviolet coverage from FUSE, and radio-loud quasars selected to have similar extended radio luminosity originally selected for orientation studies. Most of the objects have quasi-simultaneous optical-ultraviolet spectra, with significant coverage in the radio-to-X-ray wavebands. The parameters of all strong emission lines are measured by detailed spectral fitting. Many significant correlations previously found among quasar emission-line properties are also present in this sample, e.g., the Baldwin effect, the optical correlations collectively known as eigenvector 1, and others. Finally, we use our measurements plus scaling relationships to estimate black hole masses and Eddington fractions. We show the mass estimates from different emission lines are usually in agreement within a factor of 2, but nearly a third show larger differences. We suggest using multiple mass scaling relationships to estimate black hole masses when possible, and adopting a median of the estimates as the black hole mass for individual objects. Line measurements and derived AGN properties will be used for future studies examining the relationships among quasar emission lines and their spectral energy distributions.
  • We present infrared bolometric luminosity corrections derived from the detailed spectral energy distributions of 62 bright quasars of low- to moderate-redshift (z=0.03-1.4). At 1.5, 2, 3, 7, 12, 15, and 24 microns we provide bolometric corrections of the mathematical forms L_iso=\zeta \lambda L_\lambda and log(L_iso)=A+B log(\lambda L_\lambda). Bolometric corrections for radio-loud and radio-quiet objects are consistent within 95% confidence intervals, so we do not separate them. Bolometric luminosities estimated using these corrections are typically smaller than those derived from some commonly used in the literature. We investigate the possibility of a luminosity dependent bolometric correction and find that, while the data are consistent with such a correction, the dispersion is too large and the luminosity range too small to warrant such a detailed interpretation. Bolometric corrections at 1.5 $\mu$m are appropriate for objects with properties that fall in the range log(L_bol)=45.4-47.3 and bolometric corrections at all other wavelengths are appropriate for objects with properties that fall in the range log(L_bol)=45.1-47.0.
  • Bolometric corrections are used in quasar studies to quantify total energy output based on a measurement of a monochromatic luminosity. First, we enumerate and discuss the practical difficulties of determining such corrections, then we present bolometric luminosities between 1 \mu m and 8 keV rest frame and corrections derived from the detailed spectral energy distributions of 63 bright quasars of low to moderate redshift (z = 0.03-1.4). Exploring several mathematical fittings, we provide practical bolometric corrections of the forms L_iso=\zeta \lambda L_{\lambda} and log(L_iso)=A+B log(\lambda L_{\lambda}) for \lambda= 1450, 3000, and 5100 \AA, where L_iso is the bolometric luminosity calculated under the assumption of isotropy. The significant scatter in the 5100 \AA\ bolometric correction can be reduced by adding a first order correction using the optical slope, \alpha_\lambda,opt. We recommend an adjustment to the bolometric correction to account for viewing angle and the anisotropic emission expected from accretion discs. For optical/UV monochromatic luminosities, radio-loud and radio-quiet bolometric corrections are consistent within 95% confidence intervals so we do not make separate radio-loud and radio-quiet corrections. In addition, we provide several bolometric corrections to the 2-10 keV X-ray luminosity, which are shown to have very large scatter. Separate radio-loud and radio-quiet corrections are warranted by the X-ray data.
  • Dome A on the Antarctic plateau is likely one of the best observing sites on Earth thanks to the excellent atmospheric conditions present at the site during the long polar winter night. We present high-cadence time-series aperture photometry of 10,000 stars with i<14.5 mag located in a 23 square-degree region centered on the south celestial pole. The photometry was obtained with one of the CSTAR telescopes during 128 days of the 2008 Antarctic winter. We used this photometric data set to derive site statistics for Dome A and to search for variable stars. Thanks to the nearly-uninterrupted synoptic coverage, we find 6 times as many variables as previous surveys with similar magnitude limits. We detected 157 variable stars, of which 55% are unclassified, 27% are likely binaries and 17% are likely pulsating stars. The latter category includes delta Scuti, gamma Doradus and RR Lyrae variables. One variable may be a transiting exoplanet.
  • We present images of 29 post-starburst quasars (PSQs) from a Hubble Space Telescope (\emph{HST}) Advanced Camera for Surveys (ACS) Wide Field Channel Snapshot program. These broad-lined active galactic nuclei (AGN) possess the spectral signatures of massive ($M_{burst} \sim 10^{10} M_{\odot}$), moderate-aged stellar populations (hundreds of Myrs). Thus, their composite nature provides insight into the AGN-starburst connection. We measure quasar-to-host galaxy light contributions via semi-automated two-dimensional light profile fits of PSF-subtracted images. We examine the host morphologies, as well as, model the separate bulge and disk components. The \emph{HST}/ACS-F606W images reveal an equal number of spiral (13/29) and early-type (13/29) hosts, with the remaining three hosts having indeterminate classifications. AGNs hosted by early-type galaxies have on average greater luminosity than those hosted by spiral galaxies. Disturbances, such as tidal tails, shells, star-forming knots, and asymmetries are seen as signposts of interaction/merger activity. Disturbances such as these were found in 17 of the 29 objects and are evenly distributed among early-type and spiral galaxies. Two of these systems are clearly merging with their companions. Compared to other AGN of similar luminosity and redshift these PSQs have a higher fraction of early-type hosts and disturbances. Our most luminous objects with disturbed early-type host galaxies appear to be consistent with merger products. Thus, these luminous disturbed galaxies may represent a phase in an evolutionary scenario for merger driven activity and of hierarchical galaxy evolution. Our less luminous objects appear to be consistent with Seyfert galaxies not requiring triggering by major mergers. Many of these Seyferts are barred spiral galaxies.
  • We have produced the next generation of quasar spectral energy distributions (SEDs), essentially updating the work of Elvis et al. (1994) by using high-quality data obtained with several space and ground-based telescopes, including NASA's Great Observatories. We present an atlas of SEDs of 85 optically bright, non-blazar quasars over the electromagnetic spectrum from radio to X-rays. The heterogeneous sample includes 27 radio-quiet and 58 radio-loud quasars. Most objects have quasi-simultaneous ultraviolet-optical spectroscopic data, supplemented with some far-ultraviolet spectra, and more than half also have Spitzer mid-infrared IRS spectra. The X-ray spectral parameters are collected from the literature where available. The radio, far-infrared, and near-infrared photometric data are also obtained from either the literature or new observations. We construct composite spectral energy distributions for radio-loud and radio-quiet objects and compare these to those of Elvis et al., finding that ours have similar overall shapes, but our improved spectral resolution reveals more detailed features, especially in the mid and near-infrared.
  • The i-band observing conditions at Dome A on the Antarctic plateau have been investigated using data acquired during 2008 with the Chinese Small Telescope ARray. The sky brightness, variations in atmospheric transparency, cloud cover, and the presence of aurorae are obtained from these images. The median sky brightness of moonless clear nights is 20.5 mag arcsec^{-2} in the SDSS $i$ band at the South Celestial Pole (which includes a contribution of about 0.06 mag from diffuse Galactic light). The median over all Moon phases in the Antarctic winter is about 19.8 mag arcsec^{-2}. There were no thick clouds in 2008. We model contributions of the Sun and the Moon to the sky background to obtain the relationship between the sky brightness and transparency. Aurorae are identified by comparing the observed sky brightness to the sky brightness expected from this model. About 2% of the images are affected by relatively strong aurorae.
  • In 2008 January the 24th Chinese expedition team successfully deployed the Chinese Small Telescope ARray (CSTAR) to DomeA, the highest point on the Antarctic plateau. CSTAR consists of four 14.5cm optical telescopes, each with a different filter (g, r, i and open) and has a 4.5degree x 4.5degree field of view (FOV). It operates robotically as part of the Plateau Observatory, PLATO, with each telescope taking an image every 30 seconds throughout the year whenever it is dark. During 2008, CSTAR #1 performed almost flawlessly, acquiring more than 0.3 million i-band images for a total integration time of 1728 hours during 158 days of observations. For each image taken under good sky conditions, more than 10,000 sources down to 16 mag could be detected. We performed aperture photometry on all the sources in the field to create the catalog described herein. Since CSTAR has a fixed pointing centered on the South Celestial Pole (Dec =-90 degree), all the sources within the FOV of CSTAR were monitored continuously for several months. The photometric catalog can be used for studying any variability in these sources, and for the discovery of transient sources such as supernovae, gamma-ray bursts and minor planets.
  • We have obtained quasi-simultaneous ultraviolet-optical spectra for 22 out of 23 quasars in the complete PG-X-ray sample with redshift, z<0.4, and M_B<-23. The spectra cover rest-frame wavelengths from at least Lyman-alpha to H-alpha. Here we provide a detailed description of the data, including careful spectrophotometry and redshift determination. We also present direct measurements of the continua, strong emission lines and features, including Lyman-alpha, SiIV+OIV], CIV, CIII], SiIII], MgII, H-beta, [OIII], He5876+NaI5890,5896, H-alpha, and blended iron emission in the UV and optical. The widths, asymmetries and velocity shifts of profiles of strong emission lines show that CIV and Lyman-alpha are very different from H-beta and H-alpha. This suggests that the motion of the broad line region is related to the ionization structure, but the data appears not agree with the radially stratified ionization structure supported by reverberation mapping studies, and therefore suggest that outflows contribute additional velocity components to the broad emission line profiles.
  • We have detected FIRST J101614.3+520916 with the XMM-Newton X-ray Observatory. FIRST J101614.3+520916, one of the most extreme radio-loud, broad absorption line (BAL) quasars so far discovered, is also a Fanaroff-Riley type II (FR II) radio source. We find that, compared to its estimated intrinsic X-ray flux, the observed X-rays are likely suppressed, and that the observed hardness ratio indicates significant soft X-ray photons. This is inconsistent with the simplest model, a normal quasar spectrum absorbed by a large neutral HI column density, which would primarily absorb the softer photons. More complex models, involving partial covering, an ionized absorber, ionized mirror reflection, or jet contributions need to be invoked to explain this source. The suppressed but soft X-ray emission in this radio-loud BAL quasar is consistent with the behavior displayed by other BAL quasars, both radio-loud and radio-quiet.
  • We investigate the ultraviolet-to-optical spectral energy distributions (SEDs) of 17 active galactic nuclei (AGNs) using quasi-simultaneous spectrophotometry spanning 900-9000 Angstrom (rest frame). We employ data from the Far Ultraviolet Spectroscopic Explorer (FUSE), the Hubble Space Telescope (HST), and the 2.1-meter telescope at Kitt Peak National Observatory (KPNO). Taking advantage of the short-wavelength coverage, we are able to study the so-called "big blue bump," the region where the energy output peaks, in detail. Most objects exhibit a spectral break around 1100 Angstrom. Although this result is formally associated with large uncertainty for some objects, there is strong evidence in the data that the far-ultraviolet spectral region is below the extrapolation of the near-ultraviolet-optical slope, indicating a spectral break around 1100 Angstrom. We compare the behavior of our sample to those of non-LTE thin-disk models covering a range in black-hole mass, Eddington ratio, disk inclination, and other parameters. The distribution of ultraviolet-optical spectral indices redward of the break, and far-ultraviolet indices shortward of the break, are in rough agreement with the models. However, we do not see a correlation between the far-ultraviolet spectral index and the black hole mass, as seen in some accretion disk models. We argue that the observed spectral break is intrinsic to AGNs, although intrinsic reddening as well as Comptonization can strongly affect the far-ultraviolet spectral index. We make our data available online in digital format.