The manner of addressing papers adopted in the WITNESS office is to print the names and addresses, with the date when the subscriptions expire, directly on the papers themselves, in red ink. This method has several disadvantages, but these are counterbalanced by the fact that when once the name is printed it can never come off, as is the case when addresses are printed on little slips of colored paper, and then pasted on. In either method the subscribers' names are first set up in columns, under their respective post-offices, these offices being arranged alphabetically for facility of reference. It will be noticed that the post-office is only printed once, and then in large heavy type, the subscribers' names following it in the column. Five of these columns, containing on an average two hundred names, are placed in a "chase" and locked up. There are altogether in the office some three hundred and fifty of these chases constantly in use. They have to be continually revised, at which from two to ten men are constantly engaged. When the mailing time comes the chase which is to be used is inked and placed in the mailing machine, which is shown in the engraving. The machine is worked by the operator's foot. A paper is put under the hammer, as shown, and the treadle being pressed the name in the chase beneath is plainly stamped on the paper. Only the first paper of each parcel has the name of the post-office as well as that of the subscriber. When all the papers going to one post-office have been stamped, they are tied in one parcel and that with the name of the post-office being uppermost, the general address of the whole is known. When the parcel arrives there it is opened, and the postmaster makes the further distribution. [Illustration: ADDRESSING MACHINE.] Those who read this account will understand how it is that sometimes papers go astray. It would be wonderful if, out of nearly a hundred thousand names always in type at the WITNESS office, while changes are constantly being made in the lists, there were not some mistakes, and it is creditable to the system adopted by newspaper publishers that the number is comparatively so small.