Mathematica => PDF => Illustrator problem solved! (Sort of...)

[AES]

I posted recently about a situation in which a vector graphic was
created using Mathematica, including the Thickness[] and PointSize[]
options, and then transferred to a PDF format HD file using either
Export[] or the Save Selection As... menu command.

In either case when the resulting PDF file was then opened in
Illustrator, the plot lines that had been slightly thickened using the
Thickness[] command appeared enormously over-thickened in the
Illustrator display (these lines could in fact be selected in
Illustrator and thinned back down to something like the intended
thickness), although neither the axes nor the frame outlines nor the
points in the plot appeared to be similarly distorted.

I've since tried creating a PDF file from the selected graphic in the
Mathematica notebook in two additional ways, one of which seems to work.

The first way was to select only the _figure_ itself, in the cell in
which it appeared; invoke the Mathematica 6 Print Selection... menu
command; and direct the resulting printed output to a PDF file using the
Mac OS X Print to PDF option.

Opening this PDF file in Illustrator revealed a tiny 0.6733" X 0.4541"
figure located in the upper left corner of the standard 8.5" X 11"
artboard. (The original graphic had been created using the option
ImageSize->7*72.) Selecting this tiny image and enlarging it by 1000%
revealed the same grossly over thickened lines in the image.

The second way, which appears to be a more or less successful
workaround, was to select the _cell_ containing the desired graphic;
invoke the Mathematica 6 Print Selection...; and direct the printed
output to another PDF file by again using the Mac OS X Print to PDF
option.

The image in this case, which again appears located on an 8.5" X 11"
artboard, now appears entirely undistorted -- except that the Out[123]=
label also appears in the figure and has to be hand-removed.

I also note that the ratio of the widths given above, viz. 0.6733/5.6111
is 0.120000, which is suspiciously similar to the 120% reduction factor
which I believe Mathematica somehow applies to printed images??? Did someone
(at WRI or Adobe) slip a factor of 10 in their scaling somewhere?