As the broadest letter measures only 12 mm., and the stamps are about 23mm. from center to center of perforations, there are usually parts of two or three letters on each stamp. I have found a large number of single letters, pairs, portions of three letters, and in one instance, a pair and parts of two letters. Of many combinations I have found several examples. I have also found quite a number of stamps showing parts of two rows of letters, one above the other. Taking these in sequence we reconstruct the watermark [Illustration: E. & G. BOTHWELL CLUTHA MILLS] The reader will please bear in mind, that when the stamps are viewed from the back, the letters read from right to left (at least when the sheets were placed normally in the press) as is usual with the Crown and CC, CA and other watermarks. The letters are plain double lined capitals, except the third in the first line, C, which is more fancy, having a decided hook at the end of the lower curve and the upper curve ending in a point, instead of being cut off squarely, as in the case of the other letters. The E and C are followed by periods 2-1/2 mm. square. The initial capitals E, C and B are 13 mm. high, the other letters 12-1/2 mm. The upper row is about 140 mm. long, the lower about 122 mm., and the distance between the rows 11-1/2 mm. The watermark will thus fall on twelve stamps in each sheet of one hundred. But it cannot be argued from this that the stamps with watermark are only eight times as rare as those without, as we must take into consideration the proportionately large number of sheets on ordinary unwatermarked paper. The sheets were apparently placed on the press without much care, as the letters are frequently found reversed and inverted. I have not however found any placed vertically, nor have I found any other letters than the above.... As to the position of the watermark in the sheets, I believe it to be central. Its height, 37 mm., is great for a marginal watermark, and the fact that none of the letters have been found vertically, as is so frequently the case with marginal watermarks, is also in favor of a central location. We might also expect to find stamps on watermarked paper showing, as is not uncommon, the imprint of the contractors above or below, if the watermark were marginal. I, at least, have found none. Mr. Luff considers that the watermarked stamps "are on an unofficial paper used temporarily," which is without doubt the case, at least as far as the temporary nature goes. He says further: "Compared with the large number without watermark, they are sufficiently scarce to indicate a provisional use of the paper and at the same time there are enough of them to show that a considerable number of sheets were printed."