DETAILED EXPLANATION OF COLUMNS IN CATALOGUE.


(1) Object name, in IAU format based on truncated RA and dec.

(2) Repeat = the rank of repeated spectra by S/N. The highest S/N 
    spectra is given the value 1.

(3) Right ascension (2000) in decimal degrees.

(4) Declination (2000) in decimal degrees.

(5) lrgsam = LRG sample. From mid-2003 onwards `lrgsam'
     was usually 8 or 9, with Sample 8 being the primary LRG sample,
     Sample 9 the secondary sample, and lrgsam=0 denoting extra targets
     used as `fillers'.  For Semester 2003A `lrgsam' ranged from 3-5,
     with 3 being the highest priority, and with pry=6 used to denote
     a special faint subsample in field d10.

(6) Template spectrum used for best cross-correlation redshift.  T1 is
    a mean SDSS LRG spectrum, T2 and T3 are other galaxies used for
    the 2dFGRS, T4-T8 are stars with spectral types ranging from A
    to M and T9 is a composite e(a) spectrum.

(7) xcor = cross-correlation (rx) parameter, a measure of the height
     of the peak relative to the noise in the cross-correlation.  In
     general, redshifts are reliable if xcorr > 4 and very unreliable
     if xcorr < 3.

(8) za = absorption line redshift from best cross-correlation.

(9) qa = quality of auto redshift assessed by Zcode, after checking
     for emission lines.  If there are convincing lines consistent
     with za (usually only [O II] 3727A), qa is set to (qx + 1).

(10) z_fin = redshift finally adopted (= za unless manually modified).


(11) qop = redshift quality adopted (= qa unless manually modified by
     observer).  Range 1-5, where Q=1 means no plausible redshift
     could be derived, Q=2 means a possible redshift but with
     reliability 50% or lower, Q=3 means z_fin is very likely to be
     correct (>95% of cases) and a redshift with Q=4 or 5 is virtually
     certain with >99% probability.

(12) fld = 2dF field number.  Format LNN, where L is a letter a-e for
     northern (Galactic) fields and L=s for the southern strip.

(13) `date' is a 6-character identifier for the 2dF data set from
     which the spectrum was taken, i.e. the central part of the 2dfdr
     NDF filename which has the format LNNg_YYMMDD_2.sdf.  When the
     data come from a single night it is in fact the UT date of
     observation.  For most fields the data were taken over several
     nights, in which case DD is either `fi' or `00'.  A few spectra
     were taken from special subsets of the data, to avoid
     instrumental problems; these are identified by other characters
     in the DD columns.

(14) S/N = average signal-to-noise ratio of spectrum.  This is derived
     by the Zcode, using the signal and variance (error) arrays
     generated by 2dfdr.  It is thus a true S/N ratio, taking account
     of the sky and other sources of noise, and gives the best
     indication of the quality of each spectrum.  It can however be
     negative and still give a reliable redshift, since occasionally
     fibre calibration errors result in over-subtraction of sky or
     scattered light.

(15-69) SDSS identifiers, magnitudes and magnitude errors.