Polarization and limb darkening of plane-parallel atmospheres
of solar composition for 586 wavelengths between 2800A and 7464A
spaced between approximately 5A to 11A apart.

These results were extracted from Kurucz model atmospheres with
solar abundances. The file name is based on the temperature and
log gravity: T11000lg3.5_2800-7464A is a model with Teff=11,000K 
and log(g)=3.5. 

The Kurucz models (http://kurucz.harvard.edu/grids.html) were read
in by Synspec (http://nova.astro.umd.edu/Synspec49/synspec.html).
The version used was synspec51 which was found at one time on
Ivan Hubeny's Arizona web page but does not seem available now. 
I modified the synspec source code to output the monochromatic 
absorption and scattering at each wavelength as a function of
depth, from which the emergent polarized radiation is obtained.
NOTE: Synspec is normally run with a much finer wavelength grid
(<~0.1A) to resolve the absorption lines. Because this would 
require enormous files to cover the entire visible spectrum, 
the results given here were obtained *by turning off the line
opacity* (aside from the broad hydrogen lines: IMODE=2). Thus the
opacity is that of the continuum and H lines only; the addition 
of atomic and molecular line opacity would depress the polarization.
For stars in this temperature range, this is not a large effect, at
least longward of the Balmer jump. To illustrate the effect of
absorption lines, we ran a model for 12,000K and log g=3.5 with
much higher wavelength resolution, ~40,000 wavelength points 
across the spectrum (spacing ~0.2 A). Fig. Hi-Lo_Hdelta.pdf shows
the region including H-delta, and it is seen that the lines other
than the H lines do not depress the polarization much. However, in
the UV shortward of 3000A, shown in Fig. Hi-Lo_UV.pdf, there are
enough lines to depress the polarization by 20%-40%, as seen by the
points connected by the thick blue line, which are the resolved 
spectrum averaged over groups of 100 wavelength points. The effect
of lines is discussed further elsewhere on this webpage.

There are 1785 rows in each file. The first row gives the 22 values
of  \mu (= cos \theta) at which I(\mu) and Q(\mu) are tabulated.
The polarization is highest for \mu near zero,i.e., for radiation
emerging nearly parallel to the surface of the stellar atmosphere,
varies rapidly as \mu goes to zero, so I have spaced the values
of \mu more closely there.

The third row gives the values of the 586 continuum wavelengths.
The fifth row gives the emergent (astrophysical) flux F_nu  at
these 586 wavelengths. (These rows are seperated by a blank row.)

The 7th row gives begins a group of 586 rows, one for each wavelength,
each of which has I(\mu) at the 22 values of \mu.

The 594th row begins another group of 586 rows, each of which has 22
values of the Stokes parameter Q(\mu). This is followed, beginning 
at row 1181, 586 rows containing the (signed) fractional polarization, 
P(\mu)=Q(/mu)/I(/mu).

Finally, begining at row 1768, we give values of I,Q and P convolved
with the U, B and V filter transmission functions. Thus rows 1768,
1770 and 1772 contain I_U, I_B and I_V, followed by Q's at 1774, 1776 
and 1778 and then P's at 1780, 1782 and 1784.

Note that the Stokes vector Q is measured with respect to the
vertical, so radiation near the limb (\mu small) is generally
negative, as the light is polarized parallel to the surface of
the atmosphere. But for hot stars in the visible, where the
gradient d B_nu/d tau_nu is small, the polarization away from
the limb may change sign and become positive (polarization parallel
to the z-axis).