ding to the nature of the correction to be made. It is well to
know, also, that it takes hold much more actively if used in the
direction of the grain, than transversely. You may, according to
circumstances, commence with a piece of coal having considerable tooth,
continue with another that is less aggressive, and wind up with a
somewhat soft piece. The heavier the charcoal the coarser its tooth, the
lightest being the softest. The plate must be washed, so as to keep the
charcoal always clean; as otherwise the dust produced, which forms a
paste, will wear down the bottom of the furrows, and the result, in the
proof, will be dull and reddish lines.

Charcoal is also used to remove the traces of the needle in those parts
of the plate in which changes were made while the drawing was still in
progress.

53. =The Scraper.=--The scraper is more efficacious than the burnisher
in the case of small places that have been deeply bitten. If the scraper
is sufficiently sharp, it leaves no trace whatever on the lowered
surface of the copper.

To sum up:--

_Charcoal_ and _scraper_ are used to remove part of the surface of the
copper. The furrows, having been reduced in depth, receive less ink in
printing; the lines gain in delicacy in the impressions.

The _burnisher_ simply displaces the copper; _charcoal_ and _scraper_
wear it away. It follows that they must be used with discernment.

54. =Hammering Out (Repoussage).=--These three means are employed when a
moderate lowering of the plate is required. When it becomes necessary to
go down to half the thickness of the plate or more, the result will be a
hollow, which will show as a spot in printing. In that case recourse is
had to the fourth means; that is to say, to hammer and anvil. Get a pair
of compasses with curved legs (_calipers_); let one of the legs rest on
the spot to be hammered out; the other leg will then indicate the place
on the back of the plate which must be struck with the hammer on the
anvil. In this way places which have been reduced with charcoal or
scraper may be brought up to the level of the plate; but if the lines
should be found to have been flattened, which would result in a dull
tone in the proofs, it will be best to have the part in question planed
out entirely, and to do it over.

55. =Finishing the Surface of the Plate.=--The charcoal occasionally
leaves traces on the plate, which show in the proof as rather too
strong a tint. You can get rid of t