Hair side, flesh side and Gregory's rule in the VMS

[quimqu]

As @nablator quoted me above - it is extremely difficult to accurately distinguish hair side from flesh side in the VMS parchment. Generally, you would expect to see Gregory's Rule if the color and texture of the two sides are clearly distinct. They aren't in the VMS, so I would not expect Gregory's Rule to be followed consistently.

Thanks Lisa, and thanks nablator for digging out the quote.

So, I might have been following a Halloween ghost... 

Nevertheless, in case it's of interest, this is how and what came out.

I cut each page into small squares of about 11 mm and kept only the ones that are blank parchment: at least 1 cm from any edge or hole, no ink or paint within 1 mm, and only the smoother half of them, to leave out folds, cockling and shadows. Then I compared each square of the recto with the square at the same spot on the verso (at least with the image coordinates). The leaf is the same piece of skin, so the same spot shows up mirrored left to right (as approximately as I could). On each square I used the same filter (lightness minus a slightly blurred copy, which keeps detail under about half a millimetre) and took its spread as the grain.

   

I could do this on 66 leaves, the ones where each side has its own photo, which gave 1356 pairs of squares. In 82% of them the same side came out grainier. The two leaves of a bifolio are photographed separately, but they still agree on which side is outside in 34 of 41 bifolios, or 18 of 20 if I only count the clear ones. The middle opening of each quire is a single piece of skin, and there both pages come out as the same side in 9 of 10 quires. And it matches my first, whole-page measurement in 46 of 50 bifolios.

   

So there does seem to be a small, consistent difference in fine texture between the two sides of a leaf, and it follows the skin from one leaf to its conjugate. It's small, though. In 12 of the 50 bifolios it isn't clear, and the foldouts can't be matched because each side comes from a different photo. A grainy zone on one side also tends to be a bit grainy on the other (correlation 0.29), so the part of the skin matters too, which is why comparing the same zones helps.

What I can't say is which side is hair. That still rests only on f112r. I also can't say that it's visible by eye. It's a statistical difference of the Beinecke images at less than a millimetre, which fits with what you say about the parchment. What I really wanted to know was whether hair and flesh can be told apart at all, because the sides might have mattered when the quires were first made, and that order could have been lost when the book was rebound later.

This is the table per bifolio, with the outer and inner pages of each one:

   

I have it clear that I've been measuring the photographs and not the skin... Either way, thanks again for answering!