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