We present six new time-delay measurements obtained from $R_c$-band monitoring data acquired at the MPIA 2.2m telescope at La Silla observatory between October 2016 and February 2020. The lensed quasars HE 0047-1756, WG 0214-2105, DES 0407-5006, 2M 1134-2103, PSJ 1606-2333 and DES 2325-5229 were observed almost daily at high signal-to-noise ratio to obtain high-quality light curves where we can record fast and small-amplitude variations of the quasars. We measure time delays between all pairs of multiple images with only one or two seasons of monitoring at the exception of the time delays relative to image D of PSJ 1606-2333. The most precise estimate is obtained for the delay between image A and image B of DES 0407-5006, with $\tau_{AB} = -128.4^{+3.5}_{-3.8}$ days (2.8% precision) including systematics due to extrinsic variability in the light curves. For HE 0047-1756, we combine our high-cadence data with measurements from decade-long light-curves from previous COSMOGRAIL campaigns, and reach a precision of 0.9 day on the final measurement. The present work demonstrates the feasibility of measuring time delays in lensed quasars in only one or two seasons, provided high signal-to-noise ratio data are obtained at a cadence close to daily.

We present a systematic search for wide-separation (Einstein radius >1.5"), galaxy-scale strong lenses in the 30 000 sq.deg of the Pan-STARRS 3pi survey on the Northern sky. With long time delays of a few days to weeks, such systems are particularly well suited for catching strongly lensed supernovae with spatially-resolved multiple images and open new perspectives on early-phase supernova spectroscopy and cosmography. We produce a set of realistic simulations by painting lensed COSMOS sources on Pan-STARRS image cutouts of lens luminous red galaxies with known redshift and velocity dispersion from SDSS. First of all, we compute the photometry of mock lenses in gri bands and apply a simple catalog-level neural network to identify a sample of 1050207 galaxies with similar colors and magnitudes as the mocks. Secondly, we train a convolutional neural network (CNN) on Pan-STARRS gri image cutouts to classify this sample and obtain sets of 105760 and 12382 lens candidates with scores pCNN>0.5 and >0.9, respectively. Extensive tests show that CNN performances rely heavily on the design of lens simulations and choice of negative examples for training, but little on the network architecture. Finally, we visually inspect all galaxies with pCNN>0.9 to assemble a final set of 330 high-quality newly-discovered lens candidates while recovering 23 published systems. For a subset, SDSS spectroscopy on the lens central regions proves our method correctly identifies lens LRGs at z~0.1-0.7. Five spectra also show robust signatures of high-redshift background sources and Pan-STARRS imaging confirms one of them as a quadruply-imaged red source at z_s = 1.185 strongly lensed by a foreground LRG at z_d = 0.3155. In the future, we expect that the efficient and automated two-step classification method presented in this paper will be applicable to the deeper gri stacks from the LSST with minor adjustments.

We use spectroscopic observations of the gravitationally lensed systems SDSS0924+0219(BC), Q1355-2257(AB), and SDSS1029+2623(BC) to analyze microlensing and dust extinction in the observed components. We detect chromatic microlensing effects in the continuum and microlensing in the broad emission line profiles of the systems SDSS0924+0219(BC), and Q1355-2257(AB). Using magnification maps to simulate microlensing and modeling the emitting region as a Gaussian intensity profile with size $r_s \propto \lambda ^p$, we obtain the probability density functions for a logarithmic size prior at $\lambda_{rest-frame}=3533$ {\AA}. In the case of SDSS0924+0219, we obtain: $r_s = 4^{+3}_{-2}$ $\sqrt{M/M_{\odot}}$ light-days (at $1 \sigma$), which is larger than the range of other estimates, and $p = 0.8 \pm 0.2$ (at $1 \sigma$), which is smaller than predicted by the thin disk theory, but still in agreement with previous results. In the case of Q1355-2257 we obtain (at $1 \sigma$): $r_s = 3.6^{+3.0}_{-1.6}$ $\sqrt{M/M_{\odot}}$ light-days, which is also larger than the theoretical prediction, and $p = 2.0 \pm 0.7$ that is in agreement with the theory within errors. SDSS1029+2326 spectra show evidence of extinction, probably produced by a galaxy in the vicinity of image C. Fitting an extinction curve to the data we estimate $\Delta E \sim 0.2$ in agreement with previous results. We found no evidence of microlensing for this system.

We present new measurements of the time delays of WFI2033-4723. The data sets used in this work include 14 years of data taken at the 1.2m Leonhard Euler Swiss telescope, 13 years of data from the SMARTS 1.3m telescope at Las Campanas Observatory and a single year of high-cadence and high-precision monitoring at the MPIA 2.2m telescope. The time delays measured from these different data sets, all taken in the R-band, are in good agreement with each other and with previous measurements from the literature. Combining all the time-delay estimates from our data sets results in Dt_AB = 36.2-0.8+0.7 days (2.1% precision), Dt_AC = -23.3-1.4+1.2 days (5.6%) and Dt_BC = -59.4-1.3+1.3 days (2.2%). In addition, the close image pair A1-A2 of the lensed quasars can be resolved in the MPIA 2.2m data. We measure a time delay consistent with zero in this pair of images. We also explore the prior distributions of microlensing time-delay potentially affecting the cosmological time-delay measurements of WFI2033-4723. There is however no strong indication in our measurements that microlensing time delay is neither present nor absent. This work is part of a H0LiCOW series focusing on measuring the Hubble constant from WFI2033-4723.

Microlensing by stars within distant galaxies acting as strong gravitational lenses of multiply-imaged quasars, provides a unique and direct measurement of the internal structure of the lensed quasar on nano-arcsecond scales. The measurement relies on the temporal variation of high-magnification caustic crossings which vary on timescales of days to years. Multiwavelength observations provide information from distinct emission regions in the quasar. Through monitoring of these strong gravitational lenses, a full tomographic view can emerge with Astronomical-Unit scale resolution. Work to date has demonstrated the potential of this technique in about a dozen systems. In the 2020s there will be orders of magnitude more systems to work with. Monitoring of lens systems for caustic-crossing events to enable triggering of multi-platform, multi-wavelength observations in the 2020s will fulfill the potential of quasar microlensing as a unique and comprehensive probe of active black hole structure and dynamics.

Optical and X-ray observations of strongly gravitationally lensed quasars (especially when four separate images of the quasar are produced) determine not only the amount of matter in the lensing galaxy but also how much is in a smooth component and how much is composed of compact masses (e.g., stars, stellar remnants, primordial black holes, CDM sub-halos, and planets). Future optical surveys will discover hundreds to thousands of quadruply lensed quasars, and sensitive X-ray observations will unambiguously determine the ratio of smooth to clumpy matter at specific locations in the lensing galaxies and calibrate the stellar mass fundamental plane, providing a determination of the stellar $M/L$. A modest observing program with a sensitive, sub-arcsecond X-ray imager, combined with the planned optical observations, can make those determinations for a large number (hundreds) of the lensing galaxies, which will span a redshift range of $\sim$$0.25<z<1.5$

The primary goals of the STRong lensing Insights into the Dark Energy Survey (STRIDES) collaboration are to measure the dark energy equation of state parameter and the free streaming length of dark matter. To this aim, STRIDES is discovering strongly lensed quasars in the imaging data of the Dark Energy Survey and following them up to measure time delays, high resolution imaging, and spectroscopy sufficient to construct accurate lens models. In this paper, we first present forecasts for STRIDES. Then, we describe the STRIDES classification scheme, and give an overview of the Fall 2016 follow-up campaign. We continue by detailing the results of two selection methods, the Outlier Selection Technique and a morphological algorithm, and presenting lens models of a system, which could possibly be a lensed quasar in an unusual configuration. We conclude with the summary statistics of the Fall 2016 campaign. Including searches presented in companion papers (Anguita et al.; Ostrovski et al.), STRIDES followed up 117 targets identifying 7 new strongly lensed systems, and 7 nearly identical quasars (NIQs), which could be confirmed as lenses by the detection of the lens galaxy. 76 candidates were rejected and 27 remain otherwise inconclusive, for a success rate in the range 6-35\%. This rate is comparable to that of previous searches like SQLS even though the parent dataset of STRIDES is purely photometric and our selection of candidates cannot rely on spectroscopic information.

Time-delay strong lensing provides a unique way to directly measure the Hubble constant ($H_{0}$). The precision of the $H_{0}$ measurement depends on the uncertainties in the time-delay measurements, the mass distribution of the main deflector(s), and the mass distribution along the line of sight. Tie and Kochanek (2018) have proposed a new microlensing effect on time delays based on differential magnification of the coherent accretion disc variability of the lensed quasar. If real, this effect could significantly broaden the uncertainty on the time delay measurements by up to $30\%$ for lens systems such as PG1115+080, which have relatively short time delays and monitoring over several different epochs. In this paper we develop a new technique that uses the time-delay ratios and simulated microlensing maps within a Bayesian framework in order to limit the allowed combinations of microlensing delays and thus to lessen the uncertainties due to the proposed effect. We show that, under the assumption of Tie and Kochanek (2018), the uncertainty on the time-delay distance ($D_{\Delta t}$, which is proportional to 1/$H_{0}$) of short time-delay ($\sim18$ days) lens, PG1115+080, increases from $\sim7\%$ to $\sim10\%$ by simultaneously fitting the three time-delay measurements from the three different datasets across twenty years, while in the case of long time-delay ($\sim90$ days) lens, the microlensing effect on time delays is negligible as the uncertainty on $D_{\Delta t}$ of RXJ1131-1231 only increases from $\sim2.5\%$ to $\sim2.6\%$.

We report upon the follow up of 34 candidate lensed quasars found in the Dark Energy Survey using NTT-EFOSC, Magellan-IMACS, KECK-ESI and SOAR-SAMI. These candidates were selected by a combination of double component fitting, morphological assessment and color analysis. Most systems followed up are indeed composed of at least one quasar image and 13 with two or more quasar images: two lenses, four projected binaries and seven Nearly Identical Quasar Pairs (NIQs). The two systems confirmed as genuine gravitationally lensed quasars are one quadruple at $z_s=1.713$ and one double at $z_s=1.515$. Lens modeling of these two systems reveals that both systems require very little contribution from the environment to reproduce the image configuration. Nevertheless, small flux anomalies can be observed in one of the images of the quad. Further observations of 9 inconclusive systems (including 7 NIQs) will allow to confirm (or not) their gravitational lens nature.

We present time-delay estimates for the quadruply imaged quasar PG 1115+080. Our resuls are based on almost daily observations for seven months at the ESO MPIA 2.2m telescope at La Silla Observatory, reaching a signal-to-noise ratio of about 1000 per quasar image. In addition, we re-analyse existing light curves from the literature that we complete with an additional three seasons of monitoring with the Mercator telescope at La Palma Observatory. When exploring the possible source of bias we consider the so-called microlensing time delay, a potential source of systematic error so far never directly accounted for in previous time-delay publications. In fifteen years of data on PG 1115+080, we find no strong evidence of microlensing time delay. Therefore not accounting for this effect, our time-delay estimates on the individual data sets are in good agreement with each other and with the literature. Combining the data sets, we obtain the most precise time-delay estimates to date on PG 1115+080, with Dt(AB) = 8.3+1.5-1.6 days (18.7% precision), Dt(AC) = 9.9+1.1-1.1 days (11.1%) and Dt(BC) = 18.8+1.6-1.6 days (8.5%). Turning these time delays into cosmological constraints is done in a companion paper that makes use of ground-based Adaptive Optics (AO) with the Keck telescope.

We report on discovery results from a quasar lens search in the ATLAS public footprint, extending quasar lens searches to a regime without $u-$band or fiber-spectroscopic information, using a combination of data mining techniques on multi-band catalog magnitudes and image-cutout modelling. Spectroscopic follow-up campaigns, conducted at the 2.6m Nordic Optical Telescope (La Palma) and 3.6m New Technology Telescope (La Silla) in 2016, yielded seven pairs of quasars exhibiting the same lines at the same redshift and monotonic flux-ratios with wavelength (hereafter NIQs, Nearly Identical Quasar pairs). The quasar redshifts range between $\approx1.2$ and $\approx 2.7;$ contaminants are typically pairs of bright blue stars, quasar-star alignments along the line of sight, and narrow-line galaxies at $0.3<z<0.7.$ Magellan data of A0140-1152 (01$^h$40$^m$03.0$^s$-11$^d$52$^m$19.0$^s$, $z_{s}=1.807$) confirm it as a lens with deflector at $z_{l}=0.277$ and Einstein radius $\theta_{\rm E}=(0.73\pm0.02)^\ase$. We show the use of spatial resolution from the Gaia mission to select lenses and list additional systems from a WISE-Gaia-ATLAS search, yielding three additional lenses (02$^h$35$^m$27.4$^s$-24$^d$33$^m$13.2$^s$, 02$^h$59$^m$33.$^s$-23$^d$38$^m$01.8$^s$, 01$^h$46$^m$32.9$^s$-11$^d$33$^m$39.0$^s$). The overall sample consists of 11 lenses/NIQs, plus three lenses known before 2016, over the ATLAS-DR3 footprint ($\approx3500$~deg$^2$). Finally, we discuss future prospects for objective classification of pair/NIQ/contaminant spectra.

We report the discovery, spectroscopic confirmation and first lens models of the first two, strongly lensed quasars from a combined search in WISE and Gaia over the DES footprint. The four-image lensWGD2038-4008 (r.a.=20:38:02.65, dec.=-40:08:14.64) has source- and lens-redshifts $z_{s}=0.777 \pm 0.001$ and $z_l = 0.230 \pm 0.002$ respectively. Its deflector has effective radius $R_{\rm eff} \approx 3.4^{\prime\prime}$, stellar mass $\log(M_{\star}/M_{\odot}) = 11.64^{+0.20}_{-0.43}$, and shows extended isophotal shape variation. Simple lens models yield Einstein radii $R_{\rm E}=(1.30\pm0.04)^{\prime\prime},$ axis ratio $q=0.75\pm0.1$ (compatible with that of the starlight) and considerable shear-ellipticity degeneracies. The two-image lensWGD2021-4115 (r.a.=20:21:39.45, dec.=--41:15:57.11) has $z_{s}=1.390\pm0.001$ and $z_l = 0.335 \pm 0.002$, and Einstein radius $R_{\rm E} = (1.1\pm0.1)^{\prime\prime},$ but higher-resolution imaging is needed to accurately separate the deflector and faint quasar image. We also show high-rank candidate doubles selected this way, some of which have been independently identified with different techniques, and discuss a DES+WISE quasar multiplet selection.

We present time-delay measurements for the new quadruply imaged quasar DES J0408-5354, the first quadruply imaged quasar found in the Dark Energy Survey (DES). Our result is made possible by implementing a new observational strategy using almost daily observations with the MPIA 2.2m telescope at La Silla observatory and deep exposures reaching a signal-to-noise ratio of about 1000 per quasar image. This data quality allows us to catch small photometric variations (a few mmag rms) of the quasar, acting on temporal scales much shorter than microlensing, hence making the time delay measurement very robust against microlensing. In only 7 months we measure very accurately one of the time delays in DES J0408-5354: Dt(AB) = -112.1 +- 2.1 days (1.8%) using only the MPIA 2.2m data. In combination with data taken with the 1.2m Euler Swiss telescope, we also measure two delays involving the D component of the system Dt(AD) = -155.5 +- 12.8 days (8.2%) and Dt(BD) = -42.4 +- 17.6 days (41%), where all the error bars include systematics. Turning these time delays into cosmological constraints will require deep HST imaging or ground-based Adaptive Optics (AO), and information on the velocity field of the lensing galaxy.

We present the discovery of 3 quasar lenses in the Sloan Digital Sky Survey (SDSS), selected using two novel photometry-based selection techniques. The J0941+0518 system, with two point sources separated by 5.46" on either side of a galaxy, has source and lens redshifts $z_s = 1.54$ and $z_l = 0.343$. The AO-assisted images of J2211+1929 show two point sources separated by 1.04", corresponding to the same quasar at $z_s = 1.07,$ besides the lens galaxy and Einstein ring. Images of J2257+2349 show two point sources separated by 1.67" on either side of an E/S0 galaxy. The extracted spectra show two images of the same quasar at redshift $z_s = 2.10$. In total, the two selection techniques identified 309 lens candidates, including 47 known lenses, and 6 previously ruled out candidates. 55 of the remaining candidates were observed using NIRC2 and ESI at Keck Observatory, EFOSC2 at the ESO-NTT (La Silla), and SAM and the Goodman spectrograph at SOAR. Of the candidates observed, 3 were confirmed as lenses, 36 were ruled out, and 16 remain inconclusive. Taking into account that we recovered known lenses, this gives us a success rate of at least 50/309 (16%). This initial campaign demonstrates the power of purely photometric selection techniques in finding lensed quasars. Developing and refining these techniques is essential for efficient identification of these rare lenses in ongoing and future photometric surveys.

We use single-epoch spectroscopy of three gravitationally lensed quasars, HE0435-1223, WFI2033-4723, and HE2149-2745, to study their inner structure (BLR and continuum source). We detect microlensing-induced magnification in the wings of the broad emission lines of two of the systems (HE0435-1223 and WFI2033-4723). In the case of WFI2033-4723, microlensing affects two "bumps" in the spectra which are almost symmetrically arranged on the blue (coincident with an AlIII emission line) and red wings of CIII]. These match the typical double-peaked profile that follows from disk kinematics. The presence of microlensing in the wings of the emission lines indicates the existence of two different regions in the BLR: a relatively small one with kinematics possibly related to an accretion disk, and another one that is substantially more extended and insensitive to microlensing. There is good agreement between the estimated size of the region affected by microlensing in the emission lines, $r_s=10^{+15}_{-7} \sqrt{M/M_{\odot}}$ light-days (red wing of CIV in HE0435-1223) and $r_s=11^{+28}_{-7} \sqrt{M/M_{\odot}}$ light-days (CIII] bumps in WFI2033-4723) with the sizes inferred from the continuum emission, $r_s=13^{+5}_{-4} \sqrt{M/M_{\odot}}$ light-days (HE0435-1223) and $r_s=10^{+3}_{-2} \sqrt{M/M_{\odot}}$ light-days (WFI2033-4723). For HE2149-2745 we measure an accretion disk size $r_s=8^{+11}_{-5} \sqrt{M/M_{\odot}}$ light-days. The estimates of $p$, the exponent of the size vs. wavelength ($r_s\propto\lambda^p$), are $1.2\pm0.6$, $0.8\pm0.2$, and $0.4\pm0.3$ for HE0435-1223, WFI2033-4723, and HE2149-2745, respectively. In conclusion, the continuum microlensing amplitude in the three quasars and chromaticity in WFI2033-4723 and HE2149-2745 are below expectations for the thin disk model. The disks are larger and their temperature gradients are flatter than predicted by this model.

We have developed an interactive PYTHON code and derived crucial ephemeris data of 99.4% of all stars classified as 'Mira' in the ASAS data base, referring to pulsation periods, mean maximum magnitudes and, whenever possible, the amplitudes among others. We present a statistical comparison between our results and those given by the AAVSO International Variable Star Index (VSX), as well as those determined with the machine learning automatic procedure of Richards et al. 2012. Our periods are in good agreement with those of the VSX in more than 95% of the stars. However, when comparing our periods with those of Richards et al, the coincidence rate is only 76% and most of the remaining cases refer to aliases. We conclude that automatic codes require still more refinements in order to provide reliable period values. Period distributions of the target stars show three local maxima around 215, 275 and 330 d, apparently of universal validity, their relative strength seems to depend on galactic longitude. Our visual amplitude distribution turns out to be bimodal, however 1/3 of the targets have rather small amplitudes (A $<$ 2.5$^{m}$) and could refer to semi-regular variables (SR). We estimate that about 20% of our targets belong to the SR class. We also provide a list of 63 candidates for period variations and a sample of 35 multiperiodic stars which seem to confirm the universal validity of typical sequences in the double period and in the Petersen diagrams

We present a new framework were we simultaneously fit strong lensing (SL) and dynamical data. The SL analysis is based on LENSTOOL, and the dynamical analysis uses MAMPOSSt code, which we have integrated into LENSTOOL. After describing the implementation of this new tool, we apply it on the galaxy group SL2S\,J02140-0535 ($z_{\rm spec}=0.44$), which we have already studied in the past. We use new VLT/FORS2 spectroscopy of multiple images and group members, as well as shallow X-ray data from \xmm. We confirm that the observed lensing features in SL2S\,J02140-0535 belong to different background sources. One of this sources is located at $z_{\rm spec}$ = 1.017 $\pm$ 0.001, whereas the other source is located at $z_{\rm spec}$ = 1.628 $\pm$ 0.001. With the analysis of our new and our previously reported spectroscopic data, we find 24 secure members for SL2S\,J02140-0535. Both data sets are well reproduced by a single NFW mass profile: the dark matter halo coincides with the peak of the light distribution, with scale radius, concentration, and mass equal to $r_s$ =$82^{+44}_{-17}$ kpc , $c_{200}$ = $10.0^{+1.7}_{-2.5}$, and $M_{200}$ = $1.0^{+0.5}_{-0.2}$ $\times$ 10$^{14}$M$_{\odot}$ respectively. These parameters are better constrained when we fit simultaneously SL and dynamical information. The mass contours of our best model agrees with the direction defined by the luminosity contours and the X-ray emission of SL2S\,J02140-0535. The simultaneous fit lowers the error in the mass estimate by 0.34 dex, when compared to the SL model, and in 0.15 dex when compared to the dynamical method.The combination of SL and dynamics tools yields a more accurate probe of the mass profile of SL2S\,J02140-0535 up to $r_{200}$. However, there is tension between the best elliptical SL model and the best spherical dynamical model.

We report the discovery of a large timescale candidate microlensing event of a bulge stellar source based on near-infrared observations with the VISTA Variables in the Via Lactea Survey (VVV). The new microlensing event is projected only 3.5 arcmin away from the center of the globular cluster NGC 6553. The source appears to be a bulge giant star with magnitude Ks = 13.52, based on the position in the color-magnitude diagram. The foreground lens may be located in the globular cluster, which has well-known parameters such as distance and proper motions. If the lens is a cluster member, we can directly estimate its mass simply following Paczynski et al. (1996) which is a modified version of the more general case due to Refsdal. In that case, the lens would be a massive stellar remnant, with M = 1.5-3.5 Msun. If the blending fraction of the microlensing event appears to be small, and this lens would represent a good isolated black hole (BH) candidate, that would be the oldest BH known. Alternative explanations (with a larger blending fraction) also point to a massive stellar remnant if the lens is located in the Galactic disk and does not belong to the globular cluster.

We use spectra of the double lensed quasars HE0047-1756 and SDSS1155+6346 to study their unresolved structure through the impact of microlensing. There is no significant evidence of microlensing in the emission line profiles except for the Ly$\alpha$ line of SDSS1155+6346, which shows strong differences in the shapes for images A and B. However, the continuum of the B image spectrum in SDSS1155+6346 is strongly contaminated by the lens galaxy and these differences should be considered with caution. Using the flux ratios of the emission lines for image pairs as a baseline to remove macro-magnification and extinction, we have detected strong chromatic microlensing in the continuum measured by CASTLES (http://www.cfa.harvard.edu/castles/), in both lens systems, with amplitudes $0.09 (\lambda 16000) \lesssim |\Delta m |\lesssim 0.8 (\lambda 5439)$ for HE0047-1756, and $0.2 (\lambda 16000) \lesssim |\Delta m |\lesssim 0.8 (\lambda 5439)$ for SDSS1155+6346. Using magnification maps to simulate microlensing, and modeling the accretion disk as a Gaussian source (I $\propto$ exp(-R$^{2}$/2r$^2_s$)) of size r$_{s}$ $\propto$ $ \lambda^{p} $ we find, r$_{s}$ = 2.5$_{-1.4}^{+3.0}$ $\sqrt{M/0.3M_{\odot}}$ light days and p = 2.3 $\pm$ 0.8, at the rest frame for $\lambda$ = 2045, for HE0047-1756 (log prior), and r$_{s}$ = 5.5$_{-3.3}^{+8.2}$ $\sqrt{M/0.3M_{\odot}}$ light days and p = 1.5 $\pm$ 0.6 at the rest frame of $\lambda$ = 1398, for SDSS1155+6346 (log prior). Contrary to other studied lens systems, the chromaticity detected in HE0047-1756 and SDSS1155+6346 is large enough to fulfill the thin disk prediction. The inferred sizes, however, are very large compared to the predictions of this model, especially in the case of SDSS1155+6346.

We measure the differential microlensing of the UV Fe II and Fe III emission line blends between 14 quasar image pairs in 13 gravitational lenses. We find that the UV iron emission is strongly microlensed in 4 cases with amplitudes comparable to that of the continuum. Statistically modeling the magnifications we infer a typical size of r ~ 4*sqrt(M/Msun) light-days for the Fe line emitting regions which is comparable to the size of the region generating the UV continuum (3 ~ 7 light-days). This may indicate that a significant part of the UV Fe II and Fe III emission originates in the quasar accretion disk.