4 0 FINGER PRINTS CHAP.
shows on what parts of the card the impressions fall. Thus each of the four fingers is impressed twice, once above with a simple dab, and once below with a rolled impression, but each thumb is only impressed once ; the thumbs being more troublesome to print from than fingers. Besides, the cards would have to be made even larger than they are, if two impressions of each thumb had to be included. It takes from two and a half to three minutes to obtain the eighteen impressions that are made on each card.
The pocket apparatus is similar to one originally made and used by Sir William J. Herschel (see Plate 3, Fig. 4, in which the roller and its bearings are drawn of the same size as those I use). A small cylinder of hard wood, or of brass tube, say 14 inch long, and 2 or ~ inch in diameter, has a pin firmly driven into each end to serve as an axle. A piece of tightly-fitting india-rubber tubing is drawn over the cylinder. The cylinder, thus coated with a soft smooth compressible material, turns on its axle in two brackets, each secured by screws, as shown in Plate 2, Fig. 4, to a board (say 6 x 2- x I inch) that serves as handle. This makes a very fair and durable roller ; it can be used in the heat and damp of the tropics, and is none the worse for a wetting, but it is by no means so good for delicate work as a cylinder covered with roller composition. These are not at all difficult to make ; I have cast them for myself. The mould is. a piece of brass tube, polished inside. A thick disc, with a central hole for the lower pin of the cylinder, fits smoothly into the lower end of the mould, and a
