The Masson staining of collagen ? an explanation of an apparent paradox