This distribution contains the scanned article "Knots for
Climbers" by C. E. I. Wright and J. E. Magowan, Alpine Journal,
Volume 40, Number 236, May 1928, and Volume 40, Number 237,
November 1928.  This was published by the Alpine Club of London.
This is a major article that seems to be forgotten in current
climbing literature, though the name "butterfly knot", which the
authors coined (see page 137) is well established in both the
climbing and knot literature.  The only reference to the article
in knot books is in the chapter on life support knots by Charles
Waner in "History and Science of Knots", edited by J. C. Turner
and P. van de Griend.

There are two distributions:
   (1) Just this file and a two PDF files of the article.
   (2) This file, the PDF files, raw scans, and intermediate
files.  The files with the scans bring the total size to over
3 Gbytes, which is too big too fit on a CD-ROM.

The text pages were scanned at 1200 dpi as grayscale images with
an 8-bit color depth or brightness range.  The grayscale scans
were slightly rotated so that the text lines were approximently
horizontal.  The grayscale images were converted to black and 
white (1-bit color depth), reducing the size of the files by a
factor of 8.  Sone stray spots were cleaned up.  The text and
margins were adjusted to a standard size.  The files were then
compressed using CCITT Group 4 compression.

The initial scans were done at 1200 dpi to get maximum quality.
It is always possible to take the scans and reduce the
resolution and quality, but not not possible to increase
quality.  Grayscale scans were used so that abrupt changes would
not occur along sharp edges when a rotation is done.  After
making the enclosed PDF, I looked at some of the factors that
went into its quality.  I found that my scanner did not give the
quality that I expected from it its "1200 dpi optical
resolution" specification.  The transition from black to white
at the sharp edges of letters was about 6 pixels wide at 1200
dpi.  Effectively, I was getting only a 200 or 300 dpi scan.  By
choosing the gray level at which to convert from 8-bit to 1-bit
color depth (gray to black-and-white) I could change the width
of the strokes that make up the letters.  The sharp corners on
the letters in the original were always rounded, giving the text
in the PDF file a softer, less crisp look.

The photo plates were scaned at 600 dpi because they could not be
compressed as much as the text images.  Various compinations of
color depth and descreening were tried.  It was found that 8-bit
grayscale with no descreening was satisfactory.  The white 
background around the photos was scanned to a light shade of
gray.  These areas were cleared except around the lettering.  
The white background prints darker on paper than it appears on
screen.  The photos of the ropes are dark in the original.  
Various adjustments to brightness and contrast were tried, but 
none of the combinations improved over the raw scans.  

The file types and resolution were chosen for maximum quality.
This produces large files.  The text pages are still readable
and print clearly when reduced to 300 dpi.  The photos still
show the knots as 150 dip JPG files.  I decided to make a
smaller PDF file for ease of downloading.
 
Producing the W&M.PDF file was a learning experience.  My 
scanner has an unusual user interface.  I used a chain of three
prpgrams to convert and prepare the image files.  I made a
mistake and had to do much of the work a second time.

The files I am distributing are:

W&M.PDF -- the end result of my efforts and the item of most
           interest to others.  This has 1200 dpi text and
           600 dpi pictures.

W&M300.PDF -- A smaller PDF file size version of the article.
              This has 300 dpi text and 300 dpi pictures.

PG*.BMP -- 8-bit grayscale scans of the text pages.
           These have been rotated so text is nearly horizontal.
           These are not compressed.  Compression would cut the
           size in half, but the files would still not fit on a
           CD-ROM.

PA*.TIF -- black and white images of the text pages.  The text
           has fuzzy edges because dithering was used.  This 
           produced a dotted pattern in the gray regions at the
           edges of letters.  The margins on many of these
           images have not been trimmed.   CCITT Group 4 
           compression was appled because it gave the most
           compression of any method.

PB*.TIF -- black and white images of text pages.  No dithering
           was used so the text has sharp edges.  At high
           magnification the text looks thin, but it looks ok
           at less magnification or when printed.  The images
           sizes and margins were adjusted.  CCITT Group 4
           compression was applied.

PLT*.TIF -- scans of the photo plates.  These are slightly 
            rotated.  Blank ares were cleared.  Compression
            was applied.  The brightness and contrast were
            not adjusted from the original scans.  The darkness
            of the ropes and background are ok on screen, but
            they print too dark.  Perhaps someone can adjust
            the brightness and contrast to be better.  An
            interesting effect is to go negative, giving dark
            ropes on a light background.  With the programs I
            have, the whole image must go negative, resulting
            in white lettering on a huge black background.

ALLPGS.ZIP -- has the original scans of the pages with no 
              rotation or trimming.

   Bob Thrun
   February 2007