This file contains a description of the electronic data tables from
Bershady (1995, AJ, 109, 87; hereafter B95). See also Bershady et
al. (1994, AJ, 108, 870; hereafter B94). PLEASE REFER TO THE
ORIGINAL PAPERS FOR IMPORTANT TABLE NOTES. For further information,
contact:

     Matthew A. Bershady 
     Pennsylvania State University           voice: (814) 865-0150
     Dept. of Astronomy & Astrophysics       email: mab@astro.psu.edu
     525 Davey Lab                           fax:   (814) 863-3399
     University Park, PA 16802               

-------------------------------------------------------------------------
* Tables 3a-e from B95: Rest-frame absolute magnitudes and spectral types
-------------------------------------------------------------------------

	B95_table_3a.txt	bg,bk
	B95_table_3b.txt	bm
	B95_table_3c.txt	ak,am
	B95_table_3d.txt	fm
	B95_table_3e.txt	gm

Description:

Absolute magnitudes in all bands for spectrally classified galaxies in
our sample are presented here in Tables 3a-e, matching Tables 6a-e
from paper I (B94). Rest-frame colors can be derived from these
data. The last three columns of the tables contain the stellar
fraction (f(c,F)) from the ``free'' two-star model described in
Section 2 of B95, as well as the two ``ID'' numbers of the hot and
cool stars, referenced to Gunn & Stryker (1983).

Absolute magnitudes were derived using the k-corrections described in
B95, and assuming a standard cosmology with the Hubble constant H0=50,
the deacceleration parameter q0=0.1, and the cosmological constant
lambda=0. All absolute magnitudes are scaled to the K total apparent
magnitudes from B94. As in Tables 6a-e (B94) photometric errors are
quoted in Table 3a-e for the positive root of the error
function. However, recall that to derive an effective k-correction for
each band we have interpolated via the models between the two observed
bands that bracket the redshifted band of interest. Effectively, then,
we form a weighted average of two observed bands, and we derive the
photometric errors for each rest-frame magnitude accordingly,
including as well the error introduced by scaling matched-aperture
magnitudes to total magnitudes. [footnote: This error estimate does
not include systematic or additional random errors introduced by the
interpolation process, since we have not determined their
magnitude. Such an estimate could be made either via some Monte Carlo
procedure, or simulations. We expect the magnitude of the errors
introduced via the interpolation process to be small because the
models are reasonably accurate (i.e.  physical) representations of the
true SEDs, and they fit the observed broad-band data well. We have
neglected these errors here because our results do not depend on more
precise values for the rest-frame photometric errors.] Note that the
photometric errors for colors formed from data in Tables 3a-e are not
necessarily the quadrature sum of magnitude errors (i.e. the formula
given in the notes to Table 6 of B94) because the rest-frame magnitude
errors may be correlated via the bracketing procedure.  Also note that
Tables 3a-e still contain galaxies with saturated photometry, etc.,
and the reader should refer back to Tables 6a-e of paper I (B94) for
these flags.
