H. ANDERNACH
Observatoire de Lyon
9 Avenue Charles André
F - 69561 Saint-Genis-Laval Cedex, France
CDS, Obs. Astronomique de Strasbourg
11 Rue de
l'Université
F - 67000 Strasbourg, France
Abstract: From the literature of 1987 to 1993 a representative sample was prepared of 500 papers with tabular information for 50 supposedly extragalactic objects each. Regardless of their scientific value, but corrected for redundancy, these papers contain 842,000 entries. For the 374 papers with >100 entries each, the CDS FTP-archive of catalogues contains tables for 21%of them, increasing from 11%for 1987 to 29%for 1993. The coverage in terms of entries is much higher, and half of all papers with >1500 entries are in the archive. Valuable and/or large data sets are identified among the missing ones. Some of these can be found in other archives, and it is argued that tabular material published in future should be preserved more systematically in a collaborative effort between authors, journal editors and data centres.
Current computer networks allow almost instantaneous access to data obtained or compiled by other workers anywhere around the world (see [5] for a review). The Strasbourg Astronomical Data Centre (CDS) offers the largest collection world-wide of commonly used astronomical compilations and published tables in a publicly accessible FTP-archive, whichin May 1994 comprised 900 data sets [13]. The CDS, originally a stellar data centre, was renamed to ``Astronomical'' Data Centre in late 1991, and it now covers extragalactic data equally. Both Galactic and extragalactic objects may be extended, and published data may be classified into global (integrated), one-dimensional (spectral), two-dimensional (imaging) or even three-dimensional (imaging spectroscopy) data. The global data are technically the simplest to archive. They are usually the only to enter databases like SIMBAD, NED, or LEDA, and thus form their basic ``food''. However, such tables also allow other users to explore the data in different ways. Both these aspects motivated a quantitative study of the coverage achieved in the CDS catalogue archive (i.e. not that of SIMBAD) for extragalactic objects.
The test sample of references is presented in section 2, and section 3 assesses the frequency with which electronic information from these references are found at the CDS. In section 4 some other sources of electronic catalogues are described. Section 5 identifies examples of data sets missing at CDS and examines who are the customers of the CDS archive. In section 6 some conclusions are drawn and suggestions for improvements are made.
The aim was to assess the completeness with which major compilations, catalogues, atlases, and tables with published observational data are found in the electronic CDS archive. A minimum of 50 supposed extragalactic objects per paper was required for the current test sample, regardless of their scientific value. This number of objects was allowed to be lower if many lines (entries) of data were given per object. Papers dominated by one- or higher-dimensional data were only included if they also contained tabular information on global data. Thus, e.g. Tully's Nearby Galaxies Atlas was excluded, while the accompanying Nearby Galaxies Catalogue was included. Individual samples of 100 objects may be of limited scientific use, but their existence in electronic form would make the work of database managers and compilers of catalogues more efficient, e.g. to attach a reference to hundreds of objects for which this paper displays a spectrum. However, as the emphasis was on a representative sample of the literature the above conditions were not always taken as ironclad. Statistical statements will be based on a subsample of papers with over one hundred entries only.
Radio-source catalogues were also included, as the large majority of unresolved sources is extragalactic, even near the Galactic plane [9]. Papers dedicated to Galactic sources (planetary nebulae, HII regions or pulsars) were excluded. Optical counterparts of radio sources are frequently below the magnitude limits of current galaxy catalogues, thus only a small fraction of them (probably 15%) presently have an optical identification.
The period from 1987 to 1993 was chosen. This is long enough to study trends with time, and virtually all tabular data published since 1987 must have existed in electronic form at some stage. Instead of visually scanning the entire literature for such a period, I used the bibliography of atlases and catalogues as given in the IAU triennial reports [1][11][15] and in section 002 of Astronomy & Astrophysics Abstracts (AAA) up to vol.57. Many of the IAU and AAA references (with 50 objects or data difficult to prepare for CDS's electronic archive) were rejected. Some journals (like A&A, ApJ and MNRAS) were clearly underrepresented in the IAU lists as compared to others (like A&AS, ApJS and AJ). Eventually I browsed several journals (especially ApJS, A&AS, AJ and MNRAS) and also filled in the second half of 1993 not covered by IAU or AAA bibliographies. Finally, I cross-checked with my own unpublished compilation of references, as well as against a list of several hundred radio-continuum surveys published before 1991 [19].
The final sample of 500 references is available on request from the author. Standard reference codes are given in the Appendix. The sample is not complete in the sense that the A&A and ApJ main journals were not scanned in full. Obvious redundancy among the papers was removed, e.g. for the quasar catalogues by Hewitt & Burbidge and Veron-Cetty & Veron only the most recent version was used. Tables from some papers may have been integrated in other compilations (e.g. redshift compilations like ZCAT or SRC, available as CDS catalogues VII/164 and VII/142), but this may not be an argument against having the data as published in the CDS catalogue collection: other users might want to use the data in a way different from the compilation catalogue, or use other parts of the data which did not even enter the compilation catalogue. I made sure that all relevant entries in the CDS archive were present in the test sample, i.e. any further addition of references would only lower the completeness levels of the CDS archive derived below.
The sample is believed to be at least 80%complete and rich enough to allow some quantitative analysis. Table 1 gives a statistical overview. The origin of the references is given for each year in the first seven rows and four columns. The next two rows give the sum over the period 1987-93 for all references, and the last rows only for those papers with N entries. The IAU and AAA references provided almost half of all references gathered in this exercise. Less than 5%of the references are monographs or appeared in non-periodical publications.
The public file cdsarc.u-strasbg.fr:/pub/cats/cats.all, with the contents of the CDS archive was searched for all 500 papers. As of May 21, 1994 the file listed 680 catalogues in sections I-VIII and 220 items in the ``Journal-''(J-)directories [13]. Note however, that almost half of the items listed in this file for sections I-VII are not yet stored in the FTP archive because of lack of electronic documentation. These are mostly older and sometimes superseded data sets which will be incorporated upon request. They were thus considered as ``available'' in the present study.
Table 1: Test references and presence in the CDS electronic catalogue archive
The search result for allreferences and for the subset of references with N100 is shown in the last two columns of Table 1, respectively. The fraction of papers for which data are archived at CDS increases with year of publication, for the sample with N100 from 11%for 1987 to 29%for 1993, with an average of 21%over the seven years. Larger data sets are evidently better represented (cf. also Figure 2 below). The 84 extragalactic catalogues in the CDS archive include 24 which were collected and forwarded to CDS by the present author.
For a more quantitative assessment I also recorded the number of objects contained in each of the 500 papers. Sometimes this number could only be estimated. For some papers with few objects, but many data rows (entries), the total number of entries was used as the relevant measure of size. Generally the references deal with extragalactic objects only, but for the two largest data sets, namely the IRAS Faint Source Catalog (including the Associations) and the Catalog of IR Observations, a correction for their Galactic ``contamination'' was made. According to an estimate by M.Schmitz (priv.comm.) a total of 120,000 and 70,000 entries was assumed for the latter catalogues, respectively. For the radio source catalogues this correction would have been much smaller (10%) and was ignored. Again, the statistics derived is meant to be representative rather than exact. The total extragalactic content of the 500 references was found to be 842,000 entries (not necessarily different objects). The average is 1700 entries per paper, the median 200 (see also Fig.2).
The top seven rows of Table 2 show how all500 test references are distributed among the major journals for individual years (but note that A&A and ApJ were not systematically scanned here). The next two rows give the total number of papers N and the total number of entries N they contain. While N is highest for A&AS, followed by ApJS, AJ and MNRAS, N is highest for ApJS, followed by A&AS, MNRAS and AJ. The next two rows give the number of references N covered by the CDS archive, and the number of entries N contained in these data sets. Both N(CDS) and N(CDS) are higher for ApJS than for A&AS. The results of this table practically do not change if the references are restricted to those with N100, as can be seen from Figure 2 below.
Table 2: Partition of test references by journal and size
The Editors of Astronomy & Astrophysics (A&A) and its
Supplement Series (A&AS) agreed that ``extensive tabular data'',
accepted for publication in A&A or A&AS after 1992, and selected
either by the Editors or the authors, would be stored at CDS (see
A&AS 280,E1). To assess the realization of this agreement, I extended
the literature search for A&ASup to vol.105 (1994, see
last 12 entries of Appendix). As for A&AS alone, for the
period when the agreement was implemented, the archive contained data from
11 of 19 references, with 42%(3547/8485) of the entries.
Missing references which in my opinion merit to be stored at CDS,
are preceded with a ``-''-sign in the Appendix.
Figure 1 shows the size distribution of the 512 publications (i.e. including
the 12 from A&AS 1994). The cumulative number of papers with
>N entries is plotted vs. N.
For N100 the curve is almost a power-law
of index near , a manifestation of Zipf's law in
bibliometrics [12].
The turnover of the curve for N100 is most likely due
to incomplete sampling of smaller data sets in the literature survey.
Figure 1 (see text) Figure 2 (see text)
Figure 2 displays the completeness of the CDS archive as function of
the number of entries per paper.
The solid line shows, for a given minimum N, the fraction of
papersfor which electronic tables are archived at CDS.
All references with >10,000 entries are available at CDS, and half of
those with N1500 entries.
The dashed curve shows, for a given minimum number of entries
N per publication, the fraction of entriesstored at CDS. The fact that the largest
catalogues (i.e. those typically available at CDS) dominate by far the number
of data entries, implies that the coverage of the CDS archive never
drops below 74%even when all papers down to 50 objects are considered.
As a byproduct (but not as an assessment of SIMBAD's extragalactic data content!) the bibliographical tool simref was queried for those test references published in journals covered by the SIMBAD team (see SIMBAD User's Guide & Reference Manual III, p.66). Thirty references were not resolved by simref, and these are marked with an asterisk in the Appendix. According to [10]it is maybe not surprising that half of the missing references are of Russian or Chinese origin (both original andEnglish translation were queried). The other 15 papers, however, appeared in major journals, and seven of them deal with radio sources only. All thirty references contain original data.
While the high level of SIMBAD's literature coverage is reassuring, a recognition of a paper by simref does not imply that the datafrom these papers are also offered by SIMBAD. It usually means that the reference is attached to at least those objects having a cross-identification in SIMBAD. The on-line notes appended to some of the SIMBAD bibliographic references, especially the ones with manyobjects (like e.g. ``to be scanned'' or ``to be requested in electronic form''), document the ongoing efforts of the SIMBAD team to gradually recover missing electronic information, and to establish cross-identifications for these objects in SIMBAD.
Also, there is a frequent misunderstanding that all entries in the CDS catalogues would automatically be accessible through SIMBAD. This interesting feature is now under development within the project ALADIN [17], which will eventually allow an interactive overlay of entries from both SIMBAD and the CDS catalogues on top of the digital images of the optical sky surveys. Both SIMBAD and ALADIN would clearly benefit from a more complete coverage of published tabular data in the electronic archive of CDS.
Several on-line systems like e.g. DIRA2, ESIS, HEASARC, STARCAT, or STARLINK also offer astronomical catalogues, among which one can occasionally find valuable items not archived at CDS. Below I describe three other possible sources of catalogues.
Early in 1994 the American Astronomical Society (AAS) issued the first CD-ROM of a series with bulky data published in the AAS journals ApJ, ApJS, and AJ (see ApJ 402,1). This CD-ROM covers the year 1993, for which my test sample lists 38 references published in ApJ, ApJS and AJ. However, the CD-ROM contains data for only six (16%) of them. One of the reasons for this small coverage could be the fact that data contribution by the authors is voluntary. Also, one CD-ROM per year seems too infrequent. It not only burdens the authors with data requests (cf.sect.5), it may even inhibit full comprehension of papers for which data are given in electronic form only (e.g. 1994AJ.107.1629T, not in the present test sample) until the CD-ROM appears. Database managers would like to use the data within weeks from publication. The amount of data on the AAS CD-ROM is small enough to be integrated in the CDS archive.
Within the framework of the Working Group ``Radioastronomical Databases'' of IAU Commission 40 [2] the present author has now collected 120 electronic radio-source catalogues, virtually all of them were not previously archived at data centres. Currently 55 of these tables (comprising 511,000 records) are searchable in the Einstein On-line Service EOLS [4] and NASA's Astrophysics Data System ADS. Separate ASCII tables are not provided by EOLS or ADS. However, their integration into the CDS archive only require a translation of the EOLS documentation files from ADS to CDS standards [14].
Table 3 compares the total number of existing electronic radio-source catalogues
with those integrated in either CDS or EOLS/ADS. Note that the statistics in this table is not restricted to (although it is dominated by) the period from 1987 to 1993. As already seen in Figure 2 above, the archiving statistics is clearly better when the number of entries rather than the number of references is counted.
Table 3: Electronic radio-source catalogues in EOLS and in CDS
According to one of the most productive surveyors of the radio sky [7] ``most extragalactic radio astronomers live in a distant part of the universe, isolated from the rest of astronomy.'' His remedy to the problem, namely to make a deeper large-scale radio survey, in order to detect objects which the ``normal'' astronomer can study, is now being realized, and will contribute another 2 million radio sources before the year 2000 [6][8].
Two public on-line databases exist for extragalactic objects. The NASA/IPAC Extragalactic Database (NED) [20] covers all extragalactic objects, while the Lyon-Meudon Extragalactic Database (LEDA) [18], covers nearby (z0.2) galaxies only and offers up to 66 different parameters per galaxy. These databases provide very valuable cross-identifications between individual catalogues, but they do not provide data tables as published, nor do they generally make use of all data columns, nor can the input data sets be reconstructed by database interrogation. However, both need these tables to feed their databases, and frequently these are page-scanned and must then be proof-read. Much time could be saved if database teams could obtain tables routinely from a central archive. Vice versa, major tables which NED or LEDA received from the authors or converted from paper into electronic form could be provided to that archive. The well-documented and easily searchable FTP-facility for electronic catalogues at CDS would be an obvious repository.
It is often argued that modern (i.e. larger and more sensitive) data sets would supersede previous data anyway. This is invalid if these older data were taken with different equipment, in another waveband or with different angular resolution, in addition to the different epoch, which would allow to check for variability. A complete recovery of data sets ``missing'' in the CDS electronic archive is equally impossible and unnecessary, but selection criteria for those to be recovered should be defined. Among the eighty references with >1000 entries I found, e.g. catalogues of HI Observations of 10,000 galaxies (AAA.49.002.026H); 6445 Southern Peculiar Galaxies (AAA.43.002.088A); 4000 Nearby Galaxies (AAA.45.002.054T); 2300 Southern Ringed Galaxies (AAA.53.002.023B); Visual and IR photometry of 2000 galaxies (AAA.50.002.113D); lists of Low Surface Brightness Dwarfs (AAA.49.002.069K) and of a Drift Scan Survey (1993AJ.105.393K) with 1500 galaxies each, not to mention over a dozen radio-source catalogues. Two authors whose data I requested recently stated that they had been asked for the same data sets (1993AJ.106.1273 and 1980ApJS.42.565) over 50 resp. over 100 times (!). This proves the benefit of such a storage at CDS. Clearly, a regular archiving of all (even medium-sized) data sets would be essential in saving manpower needed to keep extragalactic databases complete and up-to-date, and to help compilers of reference catalogues, like e.g. this one:
During my work ... I repeatedly had to refer to the observational data whic h involved looking through a large number of individual papers each time. ...much time would be saved if observational data were kept on file for individual sources. This Handbook is the result of the integration of a part of such a file into a volume. I hope the use of this volume will now save time for others[16].
This work of 1978 collected for the first time the most relevant data for almost 200 strong radio sources from 0 to 12 of RA(1950). However, the other half of the sky, and another announced volume on weaker sources, was never finished for the obvious reason of an overwhelming effort required to compile them.
With present-day computers and networking capabilities we are clearly in a much better starting position. However, along the way of developing such tools, and even now, much information which once existed in ``machine-readable'' form (starting with punched cards!) was not saved on appropriate media after publication. Some of this material is now being recovered (at CDS and elsewhere) with optical character recognition (OCR) devices, but at the expense of large amounts of manpower for proofreading. Several large compilations were published in extremely small and sometimes broken fonts (many of them on microfiche), impossible to recover even with modern OCR methods. It is surprising that even for papers published during the past 9 months (in A&AS by authors in central Europe) it was necessary to page-scan the tables at CDS because the authors either did not provide the data or had lost them. A supply of the data directly to the data centre would be a minimal effort and make such exercises unnecessary.
To get an idea on who are the ``customers'' of CDS data, I inspected the files that record the downloading of data sets from the FTP archive for sections VII-VIII of the archive (i.e. for non-stellar and radio data). After exclusion of file accesses for local backups, I found that the great majority of FTP-copies were made by staff of other data centres, database managers, a few compilers of reference catalogues, but very few by other individual astronomers. This differs from the other sections of the archive: among the 27 most copied catalogues since 1991, only two are extragalactic (Ochsenbein, priv.comm.).
The catalogue archive at CDS concentrates the world's largest set of astronomical catalogues in one place, and this paper provides the first quantitative assessment of its extragalactic content. While a clear trend for better coverage of more recent publications and larger data sets is evident , data from only 21%of the 374 papers with 100 extragalactic entries and published between 1987 and 1993 are in the archive. This does not imply that all other material should be restored, but rather gives in impression on how the astronomical community has preserved material that was originally quite costly to prepare. For the period since implementation of the A&A-CDS agreement (see sect.3.2), the CDS archive was found to contain 11 of 19 (extragalactic) papers with 42%of the data entries. This is less than expected from (at least my own) judging of the scientific value and/or size of the data. Several data sets could be integrated from other electronic archives, but 140 of the 500 references were not found in any of these. The existence of different sets of astronomical catalogues in different archives prompted me to suggest in 1993 a Master Index of Astronomical Catalogues [3] in which the CDS could play an important role.
A possible reason for the low number of FTP-copies of extragalactic catalogues might be that CDS is not yet sufficiently known as an archive of such data. Also, most astronomers are probably looking for a database which would automatically provide them with entries from many individual catalogues, and are much less ready to select, copy, and browse these catalogues by themselves.
Clearly, the CDS will have to cope with ever more data sets.
Hundreds of users worldwide now rely on public databases for their
bibliographical object studies.
To accomplish a reasonable completeness, database managers must
rely on the input of electronic tables ``as published, since an
enormous amount of time will be saved from optically scanning the
tables and proofreading'' [20].
More coherent collaboration is needed between authors, editors,
and data centres to arrive at a satisfactory level of completeness.
Above all, the authors should be more willing to supply their data
directly to the data centers. Editors of journals can help ``stimulate''
this attitude, and editors other than those of A&A and A&AS should join the commitment to supply tables contained in publications
to the data centres.
In a suitable collaboration between data centres, these
should carefully monitor the literature for items missing in their archives
and actively request these data.
Eventually the user should be encouraged to
provide quality control even years after the information was archived.
A considerable boost of motivation for
authors to provide their data to the CDS archive could come from the
integration of more data (rather than only references attached
to objects) into SIMBAD, NED or LEDA, even more so if data become
transparent graphically, as is indeed planned with the powerful
ALADIN tool.
Acknowledgements
I am grateful to M.Crézé for suggesting this study, and for the
hospitality I received at CDS where part of this work was done.
F.Ochsenbein patiently explained and installed several tools facilitating
this study. C.Petit page-scanned ref.[11] and S.Okamura provided
ref.[15] in electronic form. The ADS abstract service
saved me several visits to the library. E.Pecontal helped with the
design of the figures, and valuable comments improving the manuscript were
received from C.Stern Grant, M.Crézé, M.Pakull, G.Paturel,
M.Schmitz, M.Kurtz, F.Simien, R.E.M.Griffin, and an anonymous referee.
Appendix: ~500 test references 1987-1993, plus 12 papers
from A&AS (1994)
See SIMBAD User's Guide III, p.144 for the 19-digit standard reference coding.
Items are sorted by year of publication, then journal.
For AAA the publication
year is that of the reference, and not that of the AAA
volume. Other codes are: AApTr = Astronomical and Astrophysical
Transactions; UCapT = Astron.Dept.Univ.Cape Town Publ..
The code is preceded by a ``+'' if stored at CDS, a * if
not recognized by the simref tool, and a ``-'' if published in
A&AS since mid-1993 with data desirable for the CDS archive.
The code is followed by a ``-'' if the paper contains
<100 entries.
1987A&AS...67..237R 1988A&AS...73..265A 1989A&AS...79..283W 1989PASP..101..360S 1987A&AS...67..261R- 1988A&AS...73..471S- 1989A&AS...80..215A 1989PBeiO..12....8Z 1987A&AS...67..341B 1988A&AS...73..515B +1989A&AS...80..299P 1990A&A...231..327R- 1987A&AS...68..427R +1988A&AS...74...83S 1989A&AS...81..253D 1990A&AS...82...41K 1987A&AS...69...23T- 1988A&AS...74..315R 1989A&AS...81..291S- 1990A&AS...82..113J *1987A&AS...69...91S 1988A&AS...74..475P 1989AAA.49.002.026H 1990A&AS...82..279R- 1987A&AS...69..487W- 1988A&AS...75...67L 1989AAA.49.002.027B +1990A&AS...82..391B 1987A&AS...70...77F 1988A&AS...75..317S 1989AAA.49.002.128M 1990A&AS...83..183M- 1987A&AS...70...95d 1988A&AS...76...21Q 1989AAA.51.002.070B 1990A&AS...83..393Q 1987A&AS...70..115S- +1988A&AS...76...65B 1989AJ.....97...69B- 1990A&AS...83..399G *1987A&AS...70..189S 1988A&AS...76..339M 1989AJ.....97..315D 1990A&AS...83..539R *1987A&AS...70..191S 1988AAA.45.002.054T 1989AJ.....97..633G 1990A&AS...83..569B 1987A&AS...70..465d- 1988AAA.49.002.069K 1989AJ.....97..708v- 1990A&AS...84...47T 1987A&AS...70..517H 1988AAA.50.002.113d 1989AJ.....97..957N 1990A&AS...84..455P 1987A&AS...71...25O 1988AJ.....95..284D 1989AJ.....97.1319C 1990A&AS...85..805F *1987A&AS...71..221O 1988AJ.....95.1340D 1989AJ.....97.1556v 1990A&AS...85.1049S- *1987A&AS...71..493K 1988AJ.....95.1659M 1989AJ.....97.1576M 1990A&AS...86..109G +1987AAA.43.002.088C 1988AJ.....95.1678N 1989AJ.....97.1721C 1990A&AS...86..167J 1987AAA.43.002.137A 1988AJ.....96...30C 1989AJ.....98...54J 1990A&AS...86..473F 1987AAA.45.002.053L +1988AJ.....96..816C 1989AJ.....98...64Z +1990AAA.50.002.111L 1987AAA.46.002.153S 1988AJ.....96.1655G 1989AJ.....98..351M +1990AAA.52.002.009K 1987AJ.....93....1B- 1988AJ.....96.1775O +1989AJ.....98..367F 1990AAA.52.002.059B 1987AJ.....93..788D 1988AJ.....96.1791F 1989AJ.....98..419V *1990AISAO..32...31A 1987AJ.....94..111L- *1988Afz....29..247S 1989AJ.....98..766S 1990AISAO..32...73A 1987AJ.....94..587B- *1988Afz....29..548L- 1989AJ.....98..931V 1990AJ.....99..463D 1987AJ.....94.1116C 1988Ap&SS.141..303B 1989AJ.....98.1148M- 1990AJ.....99.1071C 1987AJ.....94.1423F 1988ApJ...325..610H 1989AJ.....98.1175M 1990AJ.....99.1381V- 1987AcA....37..163R 1988ApJ...328..114P- 1989AJ.....98.1195K 1990AJ.....99.1435K 1987ApJ...314..129C 1988ApJ...328..530E- 1989AJ.....98.1959F 1990AJ.....99.1722K 1987ApJ...317..102S 1988ApJ...329..174T- 1989AN....310....7B 1990AJ.....99.1740H 1987ApJ...320..238S 1988ApJS...66....1U 1989AZh....66..897P- +1990AJ....100....1F 1987ApJ...321...94D- 1988ApJS...66..261K- *1989Afz....31...63S +1990AJ....100...47C 1987ApJS...63..247H 1988ApJS...66..297P 1989ApJ...339...12H 1990AJ....100.1028S 1987ApJS...63..265W 1988ApJS...67....1T 1989ApJ...343..659U- 1990AJ....100.1405W- 1987ApJS...63..311H +1988ApJS...68...91R- 1989ApJ...345...59C 1990AN....311....5B 1987ApJS...63..515L 1988ApJS...68..151H 1989ApJ...347..127F- *1990AZh....67....1A +1987ApJS...63..543S 1988ApJS...68..715K 1989ApJS...69...65H *1990Afz....32...29A 1987ApJS...63..555S 1988MNRAS.230....1B 1989ApJS...69..365F- *1990Afz....32..441S- 1987ApJS...63..771G- 1988MNRAS.230..639M- +1989ApJS...69..763F *1990Afz....33...89S- 1987ApJS...63..803D- 1988MNRAS.231..479S 1989ApJS...69..809M- 1990Afz....33..213A 1987ApJS...63..809S 1988MNRAS.231..977L +1989ApJS...70....1A *1990Afz....33..345A 1987ApJS...64..411S- 1988MNRAS.231.1065M- 1989ApJS...70..271B- 1990Afz....33..351S- 1987ApJS...64..417S- 1988MNRAS.232..111M +1989ApJS...70..329K 1990ApJ...354..124I 1987ApJS...64..581D 1988MNRAS.232..381M- 1989ApJS...70..447S 1990ApJ...357..388L 1987ApJS...64..601B +1988MNRAS.234..919H 1989ApJS...70..479S 1990ApJ...359....4v 1987ApJS...65..485C 1988MNRAS.234.1051G- 1989ApJS...70..687H 1990ApJ...365...66H 1987ApJS...65..543C- 1988MNRAS.235.1227C- 1989ApJS...70..699Y 1990ApJS...72..231M- +1987BICDS..32...81K 1988MNRAS.235.1313S 1989ApJS...70..723M- 1990ApJS...72..245S 1987MNRAS.224..895W- 1988PASP..100..452T 1989ApJS...71..433P- 1990ApJS...72..291S +1987MNRAS.227..563L 1988RMxAA..16..123C 1989ApJS...71..701H- 1990ApJS...72..433H *1987MNRAS.227..607W 1989Zelenchuk-7cm.A 1989AuJPh..42..633S +1990ApJS...72..471D 1987MNRAS.227..705W- +1989IRAS-GalQSOAssL 1989MNRAS.236..171D- +1990ApJS...72..567G 1987MNRAS.229..589P 1989A&AS...77...31V 1989MNRAS.236..207M 1990ApJS...72..621L 1987PASJ...39..709H 1989A&AS...77...75F 1989MNRAS.236..425U 1990ApJS...72..761B 1988A&A...189....7I- 1989A&AS...77..161C 1989MNRAS.236..737E- 1990ApJS...73..359C 1988A&AS...72..215P 1989A&AS...77..237R 1989MNRAS.238.1171D 1990ApJS...73..603T 1988A&AS...72..243R- 1989A&AS...78..269F 1989MNRAS.239..459D 1990ApJS...74....1Z 1988A&AS...72..415P 1989A&AS...78..277G 1989MNRAS.240..591S 1990ApJS...74..129G 1988A&AS...73..103O +1989A&AS...79...79S 1989MNRAS.240..657S- 1990ApJS...74..181Z 1988A&AS...73..125O- +1989A&AS...79..105S +1989MNRAS.240..785M 1990ApJS...74..325B
1990ApJS...74..347K- 1991ApJS...75..273O 1992ApJ...386..143C 1993AJ....105.1637H 1990ApJS...74..675B +1991ApJS...75..297H 1992ApJ...386..408H- 1993AJ....105.2107M- *1990ApJS...74..869B +1991ApJS...75..645K- 1992ApJ...387..591J 1993AJ....106..473C *1990ChA&A..15..131H 1991ApJS...75..751R 1992ApJ...400..410B 1993AJ....106..831H +1990IRASF.C...0000M 1991ApJS...75..801G +1992ApJS...78....1D +1993AJ....106.1273Z 1990MNRAS.243....1B *1991ApJS...75..935d 1992ApJS...78..365H- 1993AJ....106.1743V- 1990MNRAS.243..209M +1991ApJS...75.1011G 1992ApJS...79..157F 1993AJ....106.2197T- +1990MNRAS.243..390K +1991ApJS...76...23W +1992ApJS...79..255K- +1993AN....314...97K 1990MNRAS.243..504A- 1991ApJS...76..471S- +1992ApJS...79..331W 1993AN....314..225S 1990MNRAS.244..233R +1991ApJS...76..813S 1992ApJS...80....1S 1993AN....314..317S +1990MNRAS.244..408C 1991ApJS...76.1043P 1992ApJS...80..109B- 1993Ap&SS.200..279H 1990MNRAS.245..289I 1991ApJS...76.1055D- 1992ApJS...80..137J 1993ApJ...404...81B 1990MNRAS.246..110M 1991ApJS...76.1067d- 1992ApJS...80..211H 1993ApJ...404..521A- +1990MNRAS.246..169P 1991ApJS...77..203J 1992ApJS...80..257E 1993ApJ...405..498W- +1990MNRAS.246..256H 1991ApJS...77..331H 1992ApJS...80..479T 1993ApJ...407..470Z- 1990MNRAS.247..182B +1991ApJS...77..363S 1992ApJS...80..501O 1993ApJ...409..110S +1990PASJ...42..603S 1991ChA&A..15...19X- 1992ApJS...80..531F 1993ApJ...411..501L- 1990RMxAA..20...47C +1991MNRAS.248..112W 1992ApJS...81....5S 1993ApJ...412..541D 1991A&A...241...35B 1991MNRAS.248..398W 1992ApJS...81...35H 1993ApJ...413..453N +1991A&A...241..551W *1991MNRAS.248..483S 1992ApJS...81...83H 1993ApJ...416...36P 1991A&AS...87....1P 1991MNRAS.248..804L- 1992ApJS...81..413M 1993ApJS...84..109L 1991A&AS...87..309H 1991MNRAS.249..164O- 1992ApJS...82....1B 1993ApJS...85....1N 1991A&AS...87..389B 1991MNRAS.249..606C- 1992ApJS...82..471S 1993ApJS...85..111C- 1991A&AS...87..425R +1991MNRAS.251...46H 1992ApJS...83....1E 1993ApJS...85..249H 1991A&AS...88..559N- *1991MNRAS.251..253B +1992ApJS...83...29S +1993ApJS...86....5K 1991A&AS...89..389M- +1991MNRAS.251..330W 1992ApJS...83...65O 1993ApJS...86..365P- 1991A&AS...89..599H 1991MNRAS.252..593Z- 1992MNRAS.254...30G 1993ApJS...87...63L 1991A&AS...90...55v- 1991MNRAS.253..222W 1992MNRAS.254..655P 1993ApJS...87..135O 1991A&AS...90..121C 1991MNRAS.253..584L- 1992MNRAS.256..349Z- +1993ApJS...87..451H +1991A&AS...90..327t 1991MNRAS.253..686C +1992MNRAS.258....1L 1993ApJS...88..383L 1991A&AS...90..355C- +1991Obs...111...75L +1992MNRAS.259..233S +1993ApJS...89....1R 1991A&AS...90..375L +1991PASAu...9..170O 1992PASAu..10..140W *1993ApJS...89...35G- 1991A&AS...91....1P 1991PAZh...17..787M- +1992PASP..104..678S 1993ApJS...89...57Y 1991A&AS...91...61v- +1991UCapT..11.....F 1992PAZh...18..396A *1993ApL....29....1H 1991A&AS...91..285T 1992A&A...260..355A 1992RMxAA..24..147M- +1993ESOSR..13.....V +1991A&AS...91..337S 1992A&A...261...57B +1993A&A...278....6X *1993MNRAS.256..404L- +1991A&AS...91..371P 1992A&A...264..203G +1993A&A...278..379B- +1993MNRAS.260...77P 1991A&AS...91..513R 1992A&AS...92...63N +1993A&AS...97..435Q 1993MNRAS.261...39T 1991AAA.53.002.023B 1992A&AS...92..749T 1993A&AS...98..165K- 1993MNRAS.261..445F +1991AAA.53.002.034d 1992A&AS...93..173B 1993A&AS...98..193V- 1993MNRAS.262..475C 1991AJ....101...18G 1992A&AS...93..211F 1993A&AS...98..229P- +1993MNRAS.262.1057H 1991AJ....101..127W +1992A&AS...93..255G 1993A&AS...98..275B +1993MNRAS.263...25H 1991AJ....101..148W +1992A&AS...93..399W- 1993A&AS...98..297B 1993MNRAS.263...98B- 1991AJ....101..362C 1992A&AS...94..121T 1993A&AS...99...71K- 1993MNRAS.263..425M- 1991AJ....101.1121H 1992A&AS...94..299S 1993A&AS...99..407C- 1993MNRAS.263..707L- 1991AJ....101.1183C 1992A&AS...94..327T 1993A&AS...99..545Z 1993MNRAS.263..999T- 1991AJ....101.1207G- +1992A&AS...95....1G 1993A&AS..100...33G- 1993MNRAS.263.1023M +1991AJ....101.1561P 1992A&AS...95..129P- +1993A&AS..100...47G 1993MNRAS.264...71v- 1991AJ....101.1983B 1992A&AS...95..249L 1993A&AS..100..263O 1993MNRAS.264..298M- 1991AJ....102..461C 1992A&AS...95..489J- 1993A&AS..100..431G- 1993MNRAS.264..665V 1991AJ....102.1258F- 1992A&AS...96..389d -1993A&AS..101..207v 1993PASAu..10..310D 1991AJ....102.1581B 1992A&AS...96..435G- -1993A&AS..101..447B 1993PASJ...45..153O- 1991AJ....102.1627M 1992A&AS...96..441S 1993A&AS..101..475Q 1993PASP..105..387S 1991AJ....102.1663C 1992A&AS...96..583P +1993A&AS..102...57B +1993PNAOJ...3..169T 1991AJ....102.1680W- *1992AApTr...2..265K -1993A&AS..102..251M 1993PPMtO..12...65H *1991AZh....68..681R 1992AJ....103...11F -1993A&AS..102..451M 1993RMxAA..25...51M *1991Afz....34....5S- 1992AJ....103.1057V +1993A&AS..102..611S +1994A&AS..103..245M *1991Afz....34...13A 1992AJ....103.1062L +1993AAA.57.002.026G +1994A&AS..103..349S *1991Afz....34..205S- 1992AJ....103.1107S *1993AApTr...3..197I- -1994A&AS..103..391M +1991ApJ...368...54D 1992AJ....103.1438R 1993AJ....105...35C- +1994A&AS..103..573P 1991ApJ...369...79L 1992AJ....103.1501V 1993AJ....105...53B- +1994A&AS..104....1F 1991ApJ...375...15O- 1992AJ....103.1746N- 1993AJ....105..393K +1994A&AS..104..179G- *1991ApJ...376....8F- 1992AJ....104....1H- 1993AJ....105..427R -1994A&AS..104..259A 1991ApJ...378...77G 1992AJ....104..891F- +1993AJ....105..788F +1994A&AS..104..271G- +1991ApJ...380...30M +1992AJ....104.1706C 1993AJ....105..853I -1994A&AS..104..529T 1991ApJ...383..524Z- +1992AN....313..189S 1993AJ....105.1060P- +1994A&AS..105...67S- +1991ApJS...75....1B 1992AZh....69..225A 1993AJ....105.1251W +1994A&AS..105..211S +1991ApJS...75..241D +1992ApJ...384..404P 1993AJ....105.1271G +1994A&AS..105..385K