to the type, became hard and useless, and was taken off in the form of a thick black crust.  Another inconvenience also arose—­the quantity of ink spread on the block not being regulated by measure, and the number and direction of the transits of the inking-balls over each other depending on the will of the operator, and being consequently irregular, it was impossible to place on the type a uniform layer of ink, of the quantity exactly sufficient for the impression.  The introduction of cylindrical rollers of an elastic substance, formed by the mixture of glue and treacle, superseded the inking-balls, and produced considerable saving in the consumption of ink:  but the most perfect economy was only to be produced by mechanism.  When printing-presses, moved by the power of steam, were introduced, the action of these rollers was found to be well adapted to their performance; and a reservoir of ink was formed, from which a roller regularly abstracted a small quantity at each impression.  From three to five other rollers spread this portion uniformly over a slab (by most ingenious contrivances varied in almost each kind of press), and another travelling roller, having fed itself on the slab, passed and repassed over the type just before it gave the impression to the paper.

In order to shew that this plan of inking puts the proper quantity of ink upon the type, we must prove, first—­that the quantity is not too little:  this would soon have been discovered from the complaints of the public and the booksellers; and, secondly that it is not too great.  This latter point was satisfactorily established by an experiment.  A few hours after one side of a sheet of paper has been printed upon, the ink is sufficiently dry to allow it to receive the impression upon the other; and, as considerable pressure is made use of, the tympan on which the side first printed is laid, is guarded from soiling it by a sheet of paper called the set-off sheet.  This paper receives, in succession, every sheet of the work to be printed, acquiring from them more or less of the ink, according to their dryness, or the quantity upon them.  It was necessary in the former process, after about one hundred impressions, to change this set-off sheet, which then became too much soiled for further use.  In the new method of printing by machinery, no such sheet is used, but a blanket is employed as its substitute; this does not require changing above once in five thousand impressions, and instances have occurred of its remaining sufficiently clean for twenty thousand.  Here, then, is a proof that the quantity of superfluous ink put upon the paper in machine-printing is so small, that, if multiplied by five thousand, and in some instances even by twenty thousand, it is only sufficient to render useless a single piece of clean cloth.(1*) The following were the results of an accurate experiment upon the effect of the process just described, made at one of the largest printing establishments in the metropolis.(2*) Two hundred reams of paper were printed off, the old method of inking with balls being employed; two hundred reams of the same paper, and for the same book, were then printed off in the presses which inked their own type.  The consumption of ink by the machine was to that by the balls as four to nine, or rather less than one-half.