The Million Quasar (Milliquas) Catalogue Annotations from the historical versions' ReadMe's ---------------------------------------------------------------------------------- Version Log: 0.1 22Apr09 1067648 rows. Extracted from a working database using rules. 0.2 27Apr09 Fixed de-duplication, was dropping too many objects. Also improved selection of which names to use. 0.3 06May09 1068558 rows. Added a radio identifier and an X-ray identifier. 0.5 11May09 1086791 rows. Added 30816 objects from newly calculated associations to XMM, Chandra and SUMSS detections, plus some tweaks. Also added QSO probability % for non-confirmed QSOs. 0.6 10Jun09 1100199 rows. Added a net 826 objects from recent small surveys, various minor fixes, tweaks, and de-duplications. 0.7 19Jun09 1100700 rows. Using 2008 FIRST release, and improved de-duplication. 0.8 22Jun09 Added Chandra Source Catalog 1.0 (2009). 1.0 03Aug09 1101377 rows. Extensive de-duplication. XMM Slew Catalog data added. 1.1 09Aug09 Misc minor fixes. Removed NBCKDE objects of redshift="-1". 1.2 15Aug09 Restored NBCKDE objects which had been removed in v1.1. 1.3 17Aug09 1101820 rows. Misc fixes in the de-duplications. 1.4 20Aug09 More fixes in the de-duplications. 1.5 29Aug09 1101872 rows. Now using XMM Slew Catalog v1.3. Counts of classifications: 144985 QSOs, 23912 AGN, 1130 BL Lac types, 860295 candidates from published candidate catalogues, and 71550 radio/X-ray associated candidates unique to Milliquas. 2.0 14Jun10 1108117 rows. Includes 3590 new radio/X-ray associated objects, calculated from the new XMM3, XMM Slew 1.4, and XASSIST4 data using the QORG/ARXA algorithm. 2.1 17Aug10 1099490 rows. Now more careful to exclude narrow-line objects than v2.0 was (8627 NELGs removed), and also presents more original names of QSOs. Included objects are broad-line QSOs (core-dominated), AGN (disk-dominated), and Bl Lac types (lobe-dominated). 2.2 07Jan11 1101650 rows. Changes from the previous edition are: (1) All objects should be type I only. Type II (occluded) objects are not intended for inclusion, although some presented AGN may be type II. (2) QSOs are intended as objects for which the core optically dominates. AGN should be Seyfert-type objects where the disk outshines the core. Bl Lacs are featureless objects for which radio lobes are detected. (3) Veron QSO/AGN boundary is no longer used. Instead: (a) SDSS QSOs are taken from their DR7Q catalog (Schneider D. et al, 2010,AJ,139,2360); other SDSS emission objects are shown as AGN. (b) Non-SDSS objects are presented as AGN if their psf is non-stellar, and otherwise as QSOs if brighter than the magnitude-redshift boundary indicated by the SDSS schema, and as AGN if fainter. Accordingly, about 10K AGN from MQ v2.1 are now presented as QSOs. (4) The previous edition expunged about 8K NELGs. However, it seems that 796 AGN were also deleted in that exercise. These have been restored. (5) Data updated to that published by 31 December 2010 & small fixes done. 2.3 05Mar11 1221065 rows. Changes from the previous edition are: (1) SDSS-DR8 (Aihara H. et al, arXiv:1101.1559; http://sdss3.org/dr8) is incorporated. They present one large catalog encompassing all previous releases, instead of incremental releases as before. I have accordingly relied on DR8 to determine overall SDSS inclusion/exclusion, but DR7 astrometry is retained as better, as described in the DR8 paper. (2) A catalog of BOSS targets has been released concomitant with DR8, this is the SDSS-XDQSO photometric catalog (Bovy J. et al, arXiv:1011.6392). 350,664 BOSS targets plus 80,533 extra objects are included from XDQSO. (3) The threshold for inclusion of candidate objects is raised from 60% to 70% confidence that the object is a quasar. 2.4 05Jun11 1194462 rows. Candidate QSO likelihoods were recalculated using the larger training set provided by SDSS-DR8. 64578 candidates consequently fell below the 70% inclusion threshold and so were removed. 2.5 20Jul11 1194595 rows. Changes from the previous edition are: (1) Early (1970's-1980's) quasars are often poorly positioned due to the approximate astrometry of those days. Following investigation and some queries with original authors, about 200 early quasars have been moved to their correct optical sources, with some resultant deduplication as the original name is retained in all cases. This includes Cyril Hazard, Gail Reichert, Halton Arp, and Sramek & Weedman quasars, and others including the bright Soviet quasars TB 0948+722 and TB 0958+735. (2) About 200 recent NED quasars have been added. (3) Some radio/X-ray associations to USNO-B optical data have been added. 2.6 10Sep11 1182207 rows. Changes from the previous edition are: (1) The USNO-B catalog has been added and hard-deduplicated against APM and SDSS data to make a background pool of 1.04 billion optical objects. Also, about 12 million orphaned one-color APM objects have been removed. There are consequently improvements in this catalog's optical selections and astrometry, but the net totals have not changed much. (2) XMM Slew 1.5 catalog added. (3) Some further astrometric fixes and deduplications of early quasars. 2.7 05Nov11 1182703 rows. Changes from the previous edition are: (1) New quasar publications to 4 November 2011 added. (2) a few fixes on individual objects. 2.8 29Jan12 1182964 rows. Changes from the previous edition are: (1) Completed the fixes of non-astrometric relic data from the Veron-Cetty & Veron 13th edition, a paper for which is in preparation. About 150 further fixes were done since version 2.7 of this catalog. (2) Quasars compilation is current to the 27th January 2012. 2.9 02Apr12 1184209 rows. Changes from the previous edition are: (1) Data brought up to publications as at 31 March 2012. (2) About 50 astrometric fixes done, from my paper submitted to PASA. This concludes the astrometric cleanups performed over the recent editions. (3) The new FIRST radio catalog, 12Feb16 version, has been incorporated. 2.10 06Aug12 1224423 rows. Changes from the previous edition are: (1) Quasar data brought up to publications as at 31 July 2012, including the SDSS-DR9 release. SDSS-DR9 object classifications used, their galaxies with subclass of AGN taken to be the AGN class. (2) Xassist XMM-Newton and Chandra X-ray data updated to 30 June 2012. (3) As this is a catalog for type I (broad-line) objects, 1870 Seyfert-2s, 532 LINERs, and 6 NELGs have been removed. Bl Lacs are retained as unconstrained emission objects. 3.0 09Sep12 1223884 rows. Changes from version 2.10 are: (1) Quasar data brought up to publications as at 31 August 2012. (2) Redshifts are now calculated for the XDQSO photometric quasar candidates. (3) 30 DEEP2 objects were misplaced in v2.10, and 12 omitted. Now fixed. (4) About 15 SDSS-DR9 objects were conflated with others in the de-duplication processing done here; now separated out. (5) QSO-AGN separator restored to that used thru v2.9, thus ~5000 QSOs have been re-classed as AGN. v2.10 used SDSS-DR9 object classes. (6) AGN historic names now sourced from the Principal Galaxy Catalogue. 3.1 03Nov12 1224966 rows. Quasar data brought up to publications to 20-Oct-2012, including the DR9 Quasar catalog (Paris I. et al, 2012,arXiv:1210.5166) which manually amends SDSS-DR9, including 1341 quasar removals from DR9. 3.2 10Feb13 1221803 rows. Changes from version 3.1 are: (1) Quasar data brought up to publications as at 9 February 2013. (2) MMT quasar positions fixed, and a few small misc fixes. 3.3 07Apr13 1246268 rows. Changes from version 3.2 are: (1) Quasar data brought up to publications as at 6 April 2013. (2) Individual quasar data from 2013,A&A,551,A29 "Quasar Luminosity Function from dedicated SDSS-III and MMT data" by N. Palanque-Delabrouille et al, have been kindly provided by the lead author. This adds 713 new quasars. (3) SWIRE data is now included, consisting of 117 confirmed quasars and >20K photometric quasars from Rowan-Robinson M. et al, 2013,MNRAS,428,1958. (4) My paper 2013-PASA-30-4 marked 9 QSOs as uncertainly located. Of those: (a) I now find Q 0112-27 at J011446.4-271408, via 1RXS J011446.4-271419. (b) XSF3:29 is found at J034116.3-435748, via 2RXP J034116.6-435754. (c) RXS J11167-1711 found at J111645.6-171155, via 1RXS J111644.6-171135. (d) 1H 0828-706 is removed as there are no credible candidates. (5) 26 SDSS-DR9 quasars removed as inspection shows they are only artifacts. (6) Iovino et al, 1996 A&AS,119,265 presented 1581 quasar spectra of which 144 were marked by them as questionable ('?'). Those 144 are often anomalously bright and none show radio/X-ray associations. Those 144 are now removed. (7) Eight blazars without redshift have been dropped as those optical objects have been classed as white dwarfs by the SDSS DR7 WD Cat, 2013,ApJS,204,5. (8) Miscellaneous: 4 moves, 2 de-dups, and 4 deletions found after extensive trawling. This catalog is looking very clean now. 3.4 11Jul13 1248691 rows. Changes from version 3.3 are: (1) Quasar data brought up to publications as at 11 July 2013. (2) My paper 2013-PASA-30-4 marked 9 QSOs as uncertainly located. Of those: (a) Q 0112-27 is finally sited at J011517.2-271223, courtesy Cyril Hazard. (b) Q 0752+617 is moved to J075646.6+613639, r=16.9, b=17.4, as a better fit with its co-published quasar (NOTNI 16, r=16.7, b=17.2). (3) 92 new quasars from Cyril Hazard (!). (4) Q 1409+732 was found to be sited on a plate flaw, and is now moved one arcmin due East to perfect optical fit at J141003.2+725939, r=19.3, b>22. (5) XAssist v4 XMM-Newton & Chandra X-ray data updated to 30-June-2013. 3.5 11Aug13 1302716 rows. Changes from version 3.4 are: (1) Quasar data brought up to publications as at 10 August 2013, including SDSS-DR10. Note that the DR10 is an automated catalog, so its object classifications are not as trustworthy as manual classifications. Thus, earlier manual classifications are retained over DR10 classifications. (2) XMM-Newton X-ray data updated to 3XMM-DR4 version (24-July-2013). (3) FIRST radio data updated to 13Jun05 version (05-June-2013). 3.6 23Oct13 1277383 rows. Changes from version 3.5 are: (1) Quasar data brought up to publications as at 22 October 2013. (2) Two additional columns of radio/X-ray detections are included now. This is so that any radio lobes will always be fully displayed, and extra detections from other source catalogs give a sense of fuller coverage. (3) 158 additional quasars are added from SDSS-DR10 which they flagged as bad astrometry but which are confirmed OK by my optical background database. (4) Photometric quasar candidates from surveys other than NBCKDE and XDQSO, e.g., SWIRE, have been dropped because much of their yield are type 2. This includes Palanque-Delabrouille et al, 2011,A&A,530,122 which has been well investigated spectroscopically and its residue thus less likely to be QSOs. (5) X-ray detections from the Champ2 catalog, Kim M. et al, 2007,ApJS,169,401 have been removed. It's a small catalog which, unlike others, does not present its raw astrometry needed for calculating optical solutions. (6) The Veron quasar catalog v13 (2010 A&A,518A,10) included some objects denoted as non-quasars by their discovery papers. I've identified leading examples of these to remove them from the Million Quasars catalogue: (a) 484 emission-line galaxies (ELGs) of z<0.46 from Schneider/Schmidt/Gunn 1994 AJ,107,1245, incorporated as QSOs by VCV13, are now removed. (b) Similarly, 55 ELGs of Schneider/Schmidt/Gunn 1999,AJ,117,40 are removed. (c) 58 galaxies from 2dF-GRS (Colless M. et al, astro-ph/0306581) removed. (d) 41 galaxies from La Franca F. et al, 1999,A&AS,140,351 are removed. (7) a few positional fixes, de-duplications, and removals of type-2s done. 3.7 26Nov13 1273361 rows. Changes from version 3.6 are: (1) Quasar data brought up to publications as at 26 November 2013, including the SDSS DR10Q release by Isabelle Paris et al, 2013 arXiv:1311.4870. (2) 318 SDSS "quasars" of auto-redshift>4 which were not subsequently confirmed by visual inspection are removed as bogus (communication Adam Myers). (3) The Veron quasar catalog v13 (2010 A&A,518A,10) flagged some objects as questionable, "?" in their spec field. These have unclear spectra or are described as "possible" by their discovery papers. 178 of these are removed, and 52 such Bl-Lacs with redshift>0.5 are re-classed to AGNs. (4) Some positional fixes and de-duplications of individual objects. 3.8 16Feb14 1252004 rows. Changes from version 3.7 are: (1) Quasar data brought up to publications as at 15 February 2014. (2) NBCKDE and BOSS quasar likelihoods have been calibrated against the DR10Q (Paris I. et al, 2013 arXiv:1311.4870) classifications. (3) The Swift X-ray Point Source catalog (Evans P.A. et al, 2014,ApJS,210,8) is added to the background data; 12604 associations are displayed here. (4) 110 SDSS-DR10 automated survey objects are removed as artefacts. (5) De-duplication of radio/X-ray detections improved. 3.9 30Jun14 1262723 rows. Changes from version 3.8 are: (1) Quasar data brought up to publications as at 30 June 2014. (2) XMM-Newton Slew Survey source catalog release 1.6 added to X-ray data. (3) 1SXPS X-ray detections were being dropped; fixed, now 21268 are included. (4) The Véron-Cetty & Véron Quasar Catalog (VCV) 13th edition (2010,A&A,518,10) is now fully mapped into optical; corrections from this are incorporated. 4.0 05Aug14 1021959 rows. Changes from version 3.9 are: (1) Quasar data brought up to publications as at 5 August 2014. (2) The threshold for inclusion of QSO candidates is raised to 80% confidence. This loses ~240,000 objects from v3.9, but improves confidence of use. (3) XAssist v4 XMM-Newton & Chandra X-ray data updated to 21-July-2014. (4) 98 rows were found to be type II or not quasars, and so removed. (5) This edition is numbered 4.0 as the catalog layout has changed. Milliquas now conforms better to the NED guide "Best Practices for Data Publication" by Marion Schmitz et al (v1.2, 2013) in these ways: (a) citations are now all to original discovery papers and no longer to VCV 13th edition or to NED; thus there are now 1981 discovery citations which are indexed in a separate file "milliquas-references.txt". (b) quasar and radio/X-ray names which used to run together are now spaced out, e.g., 2SLAQJ003449.56-002103.5 --> 2SLAQ J003449.56-002103.5 4.1 20Oct14 1015615 rows. Changes from version 4.0 are: (1) Quasar data brought up to publications as at 15 October 2014. (2) RA and DEC are presented in degrees in the first two columns. (3) Type 2 objects are now included for the convenience of the users, but completeness of these is only about 95%. The type I objects are believed to be 100% complete from the literature. (4) About 3000 SDSS "quasars" are removed as not quasars. These are auto-objects which were not taken up by the manual Paris files, and which were not given a subclass by the SDSS pipeline -- thus, the lowest quality objects. Another 2000 have been reclassified as AGN or type 2 in this catalog. (5) VCV mis-copied the redshifts from 2008 A&A,488,417 -- now fixed. (6) QSOs & AGN re-partitioned by psf class and modified luminosity equation. (7) misc tweaks & fixes done when revealed by publication-related analysis. 4.2 06Dec14 1018318 rows. Changes from version 4.1 are: (1) Quasar data brought up to publications as at 6 December 2014. (2) SDSS DR10/DR10Q and DR7/DR7Q data have been reprocessed, yielding a few more quasars and using the DR10Q redshifts more as the current standard. (3) Further adjustments to the luminosity equation separating QSOs from AGN; about 2000 AGN were reclassified as QSOs. (4) 18 new quasars from Dan Weedman based on his re-analysis of the Weedman 1985 spectra in combination with NBCKDE photometric redshifts for those objects. 4.3 11Jan15 1161859 rows. Changes from version 4.2 are: (1) Quasar data brought up to publications as at 10 January 2015, including the newly-released SDSS-DR12 and SDSS-DR12Q (www.sdss.org). (2) approx 800 quasars sourced from the AGES catalog, 2012-ApJS-200-8. (3) approx 200 extra quasars added in a full round-up from NED and SIMBAD. 4.4 06Feb15 1151011 rows. Changes from version 4.3 are: (1) Quasar data brought up to publications as at 25 January 2015. (2) Some reprocessing of the SDSS-DR12 and SDSS-DR12Q data. About 1500 type-1 objects reclassified to type-2. (3) Photometric candidate likelihoods and redshifts have been recalibrated using the SDSS-DR12Q quasars as the benchmark, causing ~10K candidates to fall below the 80% confidence threshold of inclusion, thus dropped. (4) The Half-Million Quasar catalog (HMQ: Flesch 2015,PASA,32,e010) was extracted from this version. 4.5 10May15 1153110 rows. Changes from version 4.4 are: (1) 780 new quasars from LAMOST-DR1 (Luo et al 2015,arXiv:1505.01570) are added. LAMOST-DR1 is a pipeline catalog with much provisional data, so I selected best-eligible objects only. LAMOST plans a manual (reliable) QSO catalog later this year. Also, NED/SIMBAD quasar data is included up to 9 May 2015. (2) Citations are greatly simplified, referencing the recently published Half Million Quasars catalog (HMQ: Flesch 2015,PASA,32,e010. arXiv:1502.06303) for smaller legacy citations while still citing large surveys like SDSS and 2QZ. These citations are shown as alphas instead of numbers, so the Milliquas file outlay is slightly modified. (3) X-ray data now includes 3XMM-DR5 (Rosen et al, 2015 arXiv:1504.07051). Also, high-confidence data from 2XMMi-DR3 is included, as acknowledged in section 8.2 of the Rosen paper (dropped 2XMMi-DR3 data recognized as valid). (4) Minor fixes on optical data result in a few more radio/X-ray associations. 4.6 30Nov15 1392544 rows. Changes from version 4.5 are: (1) Additional QSO candidates have been sourced from the new NBCKDE v3 catalog (Richards G.T. et al, 2015-ApJS-219-39) and a variability & refraction sourced catalog (Peters C.M. et al, 2015-ApJ-811-95). These have added about 300,000 new candidates to Milliquas. Note that the NBCKDE v3 catalog uses different selection criteria from its predecessor, and so augments it rather than superseding it. (2) Threshold for inclusion of QSO candidates is raised from 80% QSO-likelihood to 90%, so should be reliable even for spot checks. (3) NED and SIMBAD listed quasars are included to 30-Nov-2015. Note that SDSS-DR12Q haven't yet published their paper which is probably slowing the uptake of their data by dynamic databases, but that data is included here. 4.7 02Mar16 1393164 rows. Changes from version 4.6 are: (1) 527 new QSOs added from recent publications, including 33 new QSOs with z>5 of which 4 have z>6. (2) The UV-Bright Quasar Survey (UVQS, T.R. Monroe et al, arXiv:1602.06255) is included and shows two mag-14 and 43 mag-15 QSOs plus many brighter AGN. 4.8 22Jun16 1422497 rows. Changes from version 4.7 are: (1) The Chandra ACIS source catalog (Wang S. et al,2016,ApJS,224,40) has been added and new X-ray associations calculated. (2) The 2nd RASS source catalog 2RXS (Boller Th. et al, 2016, A&A,588,103) replaces 1RXS for the most part, with significant differences in places. (3) Radio/X-ray processing has been tightened and data recalibrated to improve likelihood reliability, as part of a new radio/X-ray catalog in preparation. (4) 23,727 radio/X-ray associated objects, recently dropped, are restored to this catalog on the criterion that they are each 80%+ likely to be true quasars. This was done because this information is not found elsewhere. 4.9 30Jul16 1422219 rows. This version is the same as v4.8 but with these changes: (1) The 3XMM-DR6 X-ray source catalog (www.cosmos.esa.int/web/xmm-newton/xsa) has been added and new X-ray associations calculated. (2) Radio/X-ray data have been reprocessed in line with that of the new Million Optical Radio/X-ray (MORX) associations catalogue, in preparation. 5.0 19Jun17 1507287 rows. Version 5.0 presents J2000 coordinates of 1/10th arcsecond precision, and magnitudes to 1/100th mag precision, either from the original surveys or from the All-Sky Portable (ASP) optical catalog (arXiv:1705.05434). This new format prompts this release just prior to SDSS-DR14 publication. Additional changes are: (1) 5414 net QSOs/AGN added from AAOz, OzDES, LAMOST DR2, and a reload of AGES. (2) The 3XMM-DR7 and XMMSL2-2.0 Slew X-ray source catalogues have been added and new X-ray associations calculated. (3) A few duplicates have been removed, notably SDSS J131715.46+601533.1 being found to be duplicate to SBS 1315+605. 5.1 30Jun17 1541639 rows. Changes from version 5.0 are: (1) Full coverage of the Subaru-XMM Deep Field from multiple authors. (2) Inclusion of LAMOST DR3 and other miscellaneous papers to bring complete coverage to 30 June 2017. 5.2 05Aug17 1998464 rows. Changes from version 5.1 are: (1) SDSS-DR14 and SDSS-DR14Q are added, using the processing rules from the Half Million Quasars catalog (HMQ: Flesch 2015,PASA,32,e010). (2) WISE quasar candidates have been added from Secrest et al, 2015,ApJS,221,12; these are ~435K candidates over the whole sky for which 2-color optical objects were found within a 2-arcsec radius. They have been processed into pQSOs from calibration against SDSS multi-class supersets, and photometric redshifts using the the four-color based method from the HMQ Appendix 2. The four colors used were B-R, R-W1, W1-W2 & W2-W3. (3) NELGs (narrow emission-line galaxies, class='N') are added as the type-II class corresponding to the type-I AGN galaxies. Unresolved type-II objects (NLAGN, class='K') are retained as corresponding to the type-I quasars. The NLAGN/NELG divider is the same luminosity/psf function which separates QSOs from AGNs. Thus, ~30K objects previously classed in Milliquas as type-II NLAGN now appear as type-II NELGs. NELGs include unquantified contamination by LINERs, so it serves as a catch-all category presented for completeness, rather than as a strict type-II class. (4) Small publications through 5 August 2017 were added. (5) Positional fixes (of about 2 arcsec) were applied to ~150 objects. 5.3 12Sep18 1982374 rows. Changes from version 5.2 are: (1) Radio/X-ray association likelihoods are now calculated to a granularity of 1/10th of an arcecond for Chandra, XMM-Newton, & Swift X-ray source catalogs and the FIRST radio source catalog. All associations have been recalculated onto the APM/USNO-B optical background of Flesch 2017,PASA,34,25 (Figure 3). (2) Radio/X-ray association likelihoods to faint SDSS sources not in the optical background of (1) were previously halved, based on the approximately doubled sky density of SDSS. This deduction is now classed by PSF & red-blue color, consequentially most SDSS likelihoods were reduced more, often by 60%-90%. (3) Associations to faint objects neither in (1) nor (2) are now calculated as (2) instead of (1) as before. Thus their likelihoods have also reduced. The overall effect of (2)&(3) was that ~20,000 candidates fell below the 80% likelihood threshold for inclusion in Milliquas, and so were removed. (4) Visually inspected quasars from the LAMOST DR3Q (Dong et al 2018,AJ,155,189) and auto-quasars from the LAMOST DR4 (http://dr4.lamost.org) were quality checked and a total of 10,624 quasars/AGN added to Milliquas. Quasars from individual papers since August 2017 have not been added into this edition, but will be added into the next edition whenever that is. 5.4 17Oct18 1982686 rows. Changes from version 5.3 are: (1) Quasars from individual papers from August 2017 forward to 15 October 2018 have been added, plus some from a couple of older papers. (2) Quasar doublets are more carefully processed to show 2 quasars instead of 1. Discoverers don't always publish the positions of both. (3) A few gravitiationally lensed images are added as type='L', additional to the already-catalogued brightest one. These are added only where my optical data shows them, to account for any radio/X-ray associations to them. (4) The 3XMM-DR8 X-ray source catalog (www.cosmos.esa.int/web/xmm-newton/xsa) has been added and new X-ray associations calculated. (5) A bug in v5.3 dropped radio/X-ray associations to some high-z quasars; now fixed. 5.5 19Nov18 1982872 rows. Changes from version 5.4 are: (1) Quasars from publications to 14 November 2018 have been added. (2) SDSS data have been reprocessed to ensure DR14 data is used where available. (3) Close quasar pairs & lens images processing is complete. (4) 631 "unknown class" (type=U) objects are added, they are SDSS pipeline QSOs without subclass, thus are not usually accepted into Milliquas but these 631 also have probable radio/X-ray associations of p>80% which thus qualify them for inclusion. The SDSS pipeline redshift is shown but is provisional only. (5) The J2000 decimal round-up is increased to avert truncation. 5.6 17Dec18 1984120 rows. Changes from version 5.5 are: (1) Quasars from publications to 15 December 2018 have been added. SDSS-DR15 is out but has no new quasar data, however the 3rd release of SDSS-DR14Q (v4.4) is uploaded, adding ~300 quasars which had no astrometry in their earlier releases. All SDSS data reprocessed to get rid of artifacts which crept in. (2) LAMOST DR5Q (Yao et al 2018,arXiv:1811.01570) is accepted but not published yet by ApJS. In preparation for that, the LAMOST DR4 auto-quasars are here demoted to quasar-candidates because the true quasars will appear in LAMOST DR5Q, with the leftover auto-quasars thus selected against. LAMOST pipeline quasars are now processed using the same rules as with SDSS pipeline. (3) Removed 223 objects from Iovino/Clowes/Shaver,1996,A&AS,119,265 which had no quasar profile nor radio/X-ray/WISE associations, leaving 917 in Milliquas. This resolves the cautionary note at HMQ(2015,PASA,32,10) section 2.B.4 end. 5.7 07Jan19 1983749 rows. Changes from version 5.6 are: (1) Quasars added from publications to 7 January 2019, including LAMOST QSO DR5. (2) Logic tightened up for using SDSS SUBCLASS and ZWARNING fields to decide classifications for SDSS objects, including pipeline-only candidates. (3) Milliquas has been classifying all QSOs/AGN of z<0.1 as AGN. This boundary is now moved in to z=0.05 because of many core-dominated QSOs between z=0.05 and 0.1, e.g., IRAS 01267-2157 (SDSS J012910.99-214156.8) at z=0.93. HMQ (2015,PASA,32,10), sec 5, discusses these luminosity-based boundaries. (4) 22 objects from Boyle et al, 1990,MNRAS,243,18, classified by them as "possible" QSOs with "uncertain" redshifts, are removed from Milliquas. (5) Positional fixes done on ~90 legacy objects, mostly moves of <5 arcsec. 6.0 15Mar19 1983729 rows. Changes from version 5.7 are: (1) Quasars added from publications to 15 March 2019. (2) For ease of use, the "HMQ" citation (necessitating consultation of the HMQ catalog 2015-PASA-32-10 to identify the discovery article) is replaced by the HMQ-given 4-digit citation number. 30512 entries were changed. Now the HMQ references file can be directly consulted for any numeric citation. This makes Milliquas into a stand-alone catalog, so version # is reset to 6.0 . (3) positional fixes done on 9 legacy objects. (4) Where SDSS finds a different redshift for a LAMOST-discovered quasar, if the difference is >0.1z, the name & discovery credit are transferred to SDSS. This was done for 6 objects. (5) 42 AGN are dropped because their discovery paper Mauch & Sadler 2007, MNRAS,375,931 flagged those classifications as "not certain", also those 42 have neither stellar psfs nor secure (80%+) radio/X-ray associations. (6) The unseen Soviet BAL quasar "SBS 1401+566" is dropped. Its discovery paper 1986 Afanasiev/Erastova/Lipovetsky/Stepanian/Shapovalova, although cited as "in press" by Markarian/Stepanian/Erastova 1987-IAUS-121-25, was evidently never published. This object was only an artifact of the literature. (7) The quasar "Q 1409+732" is moved to the co-ordinates stated by its discovery paper Anderson & Margon 1987,Natur,327,125, where a suitable object resides. Confusion came from that paper's discrepant finding chart onto a plate flaw. 6.1 12May19 1981004 rows. Changes from version 6.0 are: (1) Quasars added from publications to 12 May 2019. (2) Gaia-DR2 astrometry is now used & flagged for 63% of Milliquas (MQ) objects. Gaia sources were matched 1-to-1 with MQ objects on the criterion that the Gaia source is that nearest to the MQ object and that the MQ object is that nearest optical (out of a 1.163G optical DB) to the Gaia source. Care was taken to avoid false matchings. 99% of all matchings are within an arcsec offset, but to find valid farther matches I binned all matches by object class and offset distance in 0.1 arcsec bins, with hundreds of targetted spot checks done to refine offset limits and to check objects with anomalous Gaia BP-RP colour suggestive of a false match. SDSS quasar candidates could not be matched beyond 1 arcsec offset because they are often optically faint and/or in close groups for which Gaia had a different object only. AGN cores match to 1.5 arcsec beyond which Gaia often shows nearby stars only. QSOs/Bl-Lacs match well out to within 4 arcsec; for these I spot-checked all matches beyond 2 arcsec and all with off-colours beyond 1 arcsec offset and removed 14 false matches. However, for Gaia sources without BP & RP colours, QSOs match reliably only within 2 arcsec and Bl-Lacs within 1.5 arcsec. X-ray/radio-only candidates (unique to MQ) match reliably out to within 2.5 arcsec offset. In all cases it was paramount to avoid false matchings, thus very many true matchings were lost beyond the offset cutoffs. (3) Reclassifications were done for Milliquas objects deemed by the SDSS-DR14 pipeline to be plain galaxies -- these were heavily spot-checked to confirm. 16 QSOs, 172 AGN and 2343 type-II objects were found to be just galaxies and so are dropped, and 433 host-dominated QSOs were reclassified to AGN. 6.2 22May19 1981261 rows. Changes from version 6.1 are: (1) About 50 fixes done on Gaia-DR2 opticals selected in crowded places. (2) Quasars added from publications to 22 May 2019, including the PS-ELQS. (3) 3LAC reprocessed to fix a few selections and improve completeness. (4) The z=2.136 QSO "SDSS J101012.77+560520.0", overlooked by SDSS, is added. 6.3 16Jun19 1986800 rows. Changes from version 6.2 are: (1) Quasars added from publications to 15 June 2019, including 4LAC. (2) positional fixes of a few arcsec were done for ~50 legacy objects. (3) That astrometry which is from the optical DB of 2017,PASA,34,25 (and not from Gaia-DR2) was tweaked by <0.01 arcsec to exactly match that DB. (4) 3581 galaxies with double radio lobes are added as type=G because the lobes show that they have active nuclei of some kind, even if well hidden. (5) Blazar candidates with neither redshift nor radio/X-ray association, about 30 objects, are dropped. Most were stated low confidence in legacy papers. 6.4 11Dec19 1968377 rows. Changes from version 6.3 are: (1) Quasars added from publications to 11 December 2019. (2) The SDSS-DR16 pipeline release is added. I have processed it using the standard HMQ rules which yield QSO classifications that accord well to the follow-up manual SDSS classifications. The DR16Q manual catalog, due out in July 2020, will give a definitive classification for the new DR16 objects. A new rule for Milliquas classification is that STARFORMING & STARBURST subclasses, or GALAXY classification by the DR12Q superset, are now taken to indicate optical host domination, e.g., classed as AGN instead of QSO. (3) More candidates with double radio lobes are added, active nuclei being indicated. A photometric threshold guards against starforming interlopers. (4) The pQSO's of SDSS candidates which are eBOSS targets are now lowered due to anti-selection, i.e., possibly DR16Q-classified as a non-quasar. Candidates which consequently fell below pQSO=80% (~57K) were dropped. (5) The Chandra Source Catalog v2 is added, but only as a supplement to the CSC v1.1 because v2 provides only stacked data for which the optical solution used by Milliquas cannot be calculated. 6.5 14Jun20 1980903 rows. Changes from version 6.4 are: (1) Quasars added from publications to 14 June 2020. (2) Some extra processing of the new SDSS-DR16 data yields 534 more type-I objects, and fixes the astrometry of 381 objects which had been offset by up to 3 arcseconds. Also about 800 QSOs are flagged as host-dominated and so are reclassified to AGN. (3) The 4XMM-DR9 XMM-Newton catalog is added and X-ray associations calculated, replacing 3XMM-DR8. (4) The 2SXPS Swift XRT catalog is added and X-ray associations calculated, replacing 1SXPS. (5) Gaia astrometry is removed from 18 quasars and 66 candidates which were in doublets (on the sky) for which Gaia had only the other object. (6) SDSS-DR16 redshifts more fully replace earlier SDSS pipeline data. (7) Gaia-sourced parallaxes & proper motions were analyzed to see if they could identify false quasars. But this test flagged some true quasars as well, regardless of cut-offs used, so individual quasars couldn't be disqualified in this way. However, legacy quasar publications could be group tested such that those with a high flag rate (>20% compared with 2% overall, signalling the presence of low-quality data) could have their flagged objects dropped, provided, as a confirmation, that those objects also had no radio/X-ray associations. This was done. The counts of dropped objects were: -- 28 out of 40 (70%) from Zhan & Chen, 1987 & 1989 ChA&A publications -- 400 out of 916 (44%) from Iovino/Clowes/Shaver, 1996,A&AS,119,265 -- 50 out of 194 (26%) from Savage/Trew/Chen/Weston, 1984,MNRAS,207,393 -- 111 out of 517 (21%) from Drinkwater 1987. Thus, in total, 589 "quasars" were dropped as having parallax/proper motion. (8) 136 SDSS-DR14 objects which were dropped by SDSS-DR16, were removed. (9) About 1500 objects had their photometry supplemented from Pan-STARRS data. 6.6 30Aug20 2019858 rows. Changes from version 6.5 are: (1) Quasars added from publications to 30 August 2020. (2) The SDSS-DR16Q data (Lyke+ 2020) are incorporated. Combined with the DR16 pipeline data these have contributed ~200K new quasars beyond DR14 plus many valuable reclassifications of the SDSS-IV data. In processing this vast data pool, I identified 37 extra new quasars not classified by DR16/DR16Q, these are presented in Milliquas with cite="MQ" and zcite of "DR16QN"/"DR16". Data selection of DR16 pipeline data was by HMQ (Flesch 2015) rules, including choice of spectra. Data flagged with ZWARNING=256 (bad astrometry) were validated by testing and are included. Details of usage of the SDSS-DR16/DR16Q data will be included in a Milliquas paper in preparation (Flesch 2021). 7.0 30Sep20 1445913 rows. Changes from version 6.6 are: (1) Quasars added from publications to 30 September 2020. (2) SDSS-DR16 represents the conclusion of the eBOSS & BOSS quasar surveys, and that conclusion somewhat orphans the leftover candidates from the SDSS-based photometric catalogs. Said candidates without any radio/X-ray/WISE associations are unlikely to be targetted in this decade, and so are herewith dropped from Milliquas, 574538 of them. Said candidates which do have WISE associations, numbering 35086, will now display AllWISE names. These significant changes prompt this release to be numbered v7.0. (3) Clean-ups have been done on the uptake of SDSS-DR16Q/DR16 by Milliquas v6.6, notably 4506 DR16Q quasars were presented as candidates, now shown as QSOs. Also, 404 low-quality pipeline objects were presented as QSOs, now dropped. (4) WISEA photometric redshifts have been recalculated using the DR16-augmented training set, and Pan_STARRS-based photometric redshifts are now calculated for radio/X-ray associated candidates on the PS1 footprint (all-sky N of decl S30) using the four-colour method of 2015,PASA,32,10, Appendix 2. (5) 222 candidates in a solid degree centred on J2033+412 in Cygnus are dropped. They are obvious stars albeit near to background Galactic X-ray sources. 7.1 14Feb21 1548897 rows. Changes from version 7.0 are: (1) Quasars added from publications to 14 February 2021. (2) For quasar candidates, the pQSO (likelihood that the candidate is a quasar) threshold for inclusion into Milliquas is lowered to 60%. This is to present more radio/X-ray associated candidates which are not available elsewhere. (3) Audits on the inclusion of DR16Q into Milliquas have led to tweaks of the selection of DR16Q objects; this topic will be detailed in the planned publication of Milliquas (Flesch 2021). Other SDSS data were also audited. Also, all galaxy data are removed in order to focus exclusively on quasars. (4) 108 visually-inspected high-z (range 2 - 4.6) SDSS-DR16 quasars added, using QuasarNET redshifts from the DR16Q Superset, with spectra re-checked by me. (5) LAMOST-DR6 is added -- being pipeline-only data, its quasar classifications are accepted into Milliquas only when supported by radio/X-ray/WISEA pQSOs. (6) 4XMM-DR10 X-ray sources added, replacing DR9. 7.2 28May21 1573824 rows. Changes from version 7.1 are: (1) Quasars added from publications to 30 April 2021. (2) About 230 quasars added from auditing of SDSS-DR16 pipeline catalog. (3) 92 SDSS-DR16Q quasars were somehow dropped. Now restored. (4) The VLASS Quick Look radio catalog is included, which adds 34189 new radio core associations and 6793 probable double radio lobe associations. (5) Ongoing audits of SDSS & LAMOST quasars have led to a few additions & drops. Some WISEA-supported SDSS/LAMOST pipeline quasars are now accepted which were otherwise marginal. Also some likely galaxies removed. (6) This release published on the ArXiv as arXiv:2105.12985 (Flesch 2021). 7.3 31Oct21 1556881 rows. Changes from version 7.2 are: (1) Quasars added from publications to 31 October 2021. (2) 17 classified quasars were dropped as moving stars (Gaia-DR2 flagged and confirmed on legacy DSS charts), 4 as galaxies, and 1 line poacher. (3) 42,637 quasar candidates for which Gaia-DR2 shows proper motion or parallax, have been dropped for efficiency. (4) 1365 quasars and ~100K candidates without Gaia-DR2 J2000 were reconciled to Pan-STARRS or SDSS J2000. Average positional shift was ~0.1 arcsec. (5) About 400 very faint (r>22) legacy quasars have been positionally tweaked by comparing to DES and Pan-STARRS images, usually ~1 arcsec, some more. (6) Selection and processing tweaks done on candidates, adding a net 93. (7) AGES (Kochanek+ 2012) was reloaded using the CDWFS X-ray survey (Masini+ 2020,ApJS,251,2) to support object selection. AGES did not publish their spectra; 368 of their objects are now cited to DR16Q which shows spectra. (8) RACS (Hale+ 2021, arXiv:2109.00956) radio associations are added. (9) RASS (Rosat All-Sky Survey) X-ray data is dropped, as it has become clear over time that its resolution is too coarse to confidently identify sources in isolation. (10) OzDES1 (Tie+ 2017,AJ,153,107) is now fully replaced by OzDES2. (11) 4XMM-DR11 X-ray sources added, replacing DR10. 7.4 12Dec21 1115590 rows. Changes from version 7.3 are: (1) Quasars added from publications to 12 December 2021. (2) SDSS-DR17 is processed, with 1003 new quasars and 266 new MaNGA AGN added. (3) The eFEDS AGN catalogue has been processed, with 532 new quasars added. Only the best-quality data were used (5029#) of which 1139 were new objects. I used redshift/photometry cut-offs on those to avoid galaxy contamination. (4) WISEA-only candidates have been removed to avoid repeating candidates from other catalogues; 448,009 candidates were removed. The only candidates now presented are those with radio/X-ray associations calculated here. (5) More Pan-STARRS astrometry added; was intended for MQ v7.3 but for a bug. (6) 1219 RACS associations showed a bad RACS name (wrong hemisphere); fixed now. 7.5 30Apr22 1122492 rows. Changes from version 7.4 are: 1. Quasars added from publications to 30 April 2022. 2. LAMOST DR7 pipeline adds 646 QSOs supported by radio/X-ray/WISEA association. 3. A search revealed six earlier (2011-2021) papers missed due to not being on ArXiv, paywalled data, or otherwise overlooked. Now added (156# new) and indexed in the references as AT20G, ICRFz, GB6hiz, Hdots1, Hdots2, and 4FGL3. 4. candidates selection tweaked for efficiency, net total is ~800 less. 5. The citations 2MAGN (Zaw+,2019) and 6dAGN (Chen+,2022) classify faint nuclear emission. Their type-I & Kewley-criterion type-II objects are accepted here. 6. HETDEX AGN of 2+ spectral lines are added into Milliquas after deduplication. 7.6 15Sep22 1613389 rows. Changes from version 7.5 are: 1. Quasars added from publications to 31 August 2022. 2. All objects/candidates calculated as having 80-100% likelihood of radio/X-ray association (RXpct) are now included, including classified galaxies. This has increased the Milliquas row count by ~50%. The threshold increase to 80% is to increase reliability. Both RXpct and Qpct (pQSO) are now shown in their own columns to clarify the distinction for users. Some radio-ray associated classified stars are also included which are possibly quasars. 3. Gaia-DR3 QSO candidates (indexed as "GAIA3" below) with redshifts are matched to radio/X-ray detections, and those with calculated associations of 80%+ likelihood are added into Milliquas as candidates. Their "QSOC" redshifts (arXiv:2206.06710) are here rounded to 0.1z to show that they are only approximate; because of the lack of spectral redshift, they cannot be fully classified as QSOs in Milliquas. 4. The classification of "q" (quasar candidate) is added to denote candidates which are 99% likely to be true QSOs, as per the method of Flesch 2015,PASA, 32,10 Section 8. 5. 4XMM-DR12 X-ray sources added, replacing DR11. 6. MIGHTEE spectral AGN (indexed as "MIGHTE" below) are added, but with cuts to avoid galaxy contamination: 2/5 MIGHTEE "True" tests required, and z>0.7. 7. I readied Gaia-DR3 astrometry to replace DR2 in Milliquas, but testing found that DR2 has better completeness of desired objects in crowded sky. Examples are J121450.06-055228.1 and SDSS J222942.84+002954.0, both of which are close red-blue doublets for which VLASS gives a radio detection onto the blue object; Gaia-DR2 presents those blue objects but Gaia-DR3 gives only the red. Maybe a Gaia on-spacecraft software change has prioritized red/yellow sources over blue, but whatever the cause, Gaia-DR2 seems better suited. 7.7 15Oct22 1630662 rows. Changes from version 7.6 are: 1. Quasars added from publications to 15 October 2022. 2. Some quasar candidates of pQSO 60%-79% were missing; now added. 3. Positional fixes done on ~6000 candidates which had "epoch 2000" astrometry from USNO-B. Where USNO-B adjudged the object to be moving, the epoch 2000 projection could be discrepant by many arcsec. Pan-STARRS astrometry has been obtained for these objects as available, i.e., those of decl>-30. Those remaining have been flagged as questionably located. 7.8 04Dec22 1427250 rows. Changes from version 7.7 are: 1. Quasars added from publications to 4 December 2022. 2. LoTSS (2022,A&A,659,A1) radio associations are included, adding about 272K candidates to Milliquas. 3. VLASS has been reprocessed to include only Gaussian detections, and more of them. This adds a net 14,734 more quasar candidates into Milliquas. 4. Association likelihoods are now calculated in 0.1-arcsec offset bins for all radio/X-ray input surveys; thus likelihoods are tweaked from before. 5. Galaxies and low-likelihood QSO candidates are now removed as Milliquas heads towards planned journal publication in 2023. 7.9 05Feb23 1461834 rows. Changes from version 7.8 are: 1. Quasars added from publications to 5 February 2023, including SDSS-DR18 and LAMOST QSO DR6-DR9. 2. Gaia-EDR3 proper motion data were used to select objects to audit, especially those of unclear classification from legacy surveys. The net outcome was the addition of ~500 quasars to Milliquas, mostly from LAMOST QSO DR1-DR5. Also notable were 12 (out of 91) Hazard (cite=0800) objects dropped due to proper motion detected, and with no redshift confirmation from SDSS nor radio/X-ray. 3. DES DR2 astrometry and/or photometry were supplied to ~15K objects. 4. QSO candidates acceptance thresholds simplified. 7.10 15Apr23 1447613 rows. Changes from version 7.9 are: 1. Quasars added from publications to 15 April 2023. 2. Gaia-DR3 quasar candidates' redshifts are henceforth rounded to the 0.01z instead of the 0.1z because their performance is being seen (by the astronomical community) to be that good. 3. The method of Cristiani+ 2023 (cite='QUBzs') approaches full spectroscopy and is used to classify selected Gaia-DR3 candidates as quasars. That subset of their output which show radio/X-ray association are accepted as quasars here. 4. ~15,000 candidates with only single-band photometry and only one radio/X-ray detection were dropped because ~half of those are not optically seen at all. 7.11 30Jun23 1023326 rows. Changes from version 7.10 are: 1. Quasars added from publications to 30 June 2023. 2. The DESI-EDR is now included, adding 60,488 new quasars. Quality checks resulted in the removal of about 2000 poor-spectrum objects and duplicates. This selection anticipates that of the upcoming DESI-EDR Quasar VAC. 3. The Milliquas project is drawing to a close after 15 years of publications. 30-June-2023 is the date of "closing the books". Thus, this version (7.11) is much like the final Milliquas (v8) to come, which adheres to the selection rules of the Half-Million Quasars catalogue (HMQ: 2015,PASA,32,e010) which selects classified quasars and candidates of 99% pQSO. Accordingly, v7.11 also restricts to this selection, and so reduces to just over a million objects. Candidates of lesser pQSO are still available from the MORX catalogue. The final MORX (v2) will also be to 30-June-2023, and will be published alongside the final Milliquas (v8). 4. Citations are now all provided in a single file, Milliquas-references.txt. v8 02Aug23 1021800 rows. Changes from version 7.11 were: (1) 468 quasars added from the DESI Ly-alpha VAC. (2) 47 DESI objects reclassified to QSOs and 68 to AGN in a process described in the Milliquas v8 paper. This concluded work still in progress at the time of the v7.11 release. (3) 19 quasars were dropped via inspections of their spectra and/or images. (4) Two columns were removed, the radio/X-ray detection likelihood and the QSO likelihood, since all objects now are of pQSO >= 99%. (5) 4XMM-DR13 X-ray sources were added, replacing DR12. (6) LSXPS X-ray data were updated through 01-July-2023. (7) ~2100 candidates were dropped as low-confidence (<97%) of being QSOs. Postscript 2026: 13096 NBCKv3 candidates of <97% QSO-confidence are invalidly present in the Milliquas v8 data. These slipped through a gap in the v7.9-simplified program logic dealing with acceptance thresholds: GAIA3 candidates are very well performed and so were accepted with a lower acceptance threshold, but this lower threshold was inadvertently applied to the NBCKv3 candidates as well, causing the invalid retentions. This does not affect the verified catalogued type=Q quasars. The MORX v2 catalog, published alongside Milliquas v8, has all the radio/X-ray associated candidates (and classified objects) listed with their calculated confidences of being QSOs or other classifications. -- EWF 9/9/2026 --