clearly marked scale rule. Suitable articles will
generally be found in an ordinary case of mathematical instruments.

A simpler and equally accurate method of taking measurements of
handwriting is by the aid of the transparent paper known as foreign
letter paper. It is usually of quarto size, very thin and transparent,
and is ruled horizontally and vertically, dividing the sheet into tiny
squares. It is laid over the writing to be examined, and the various
measurement marks are made with a finely pointed lead pencil. The lines
and squares are used for measurement as the parallels of latitude and
longitude are used on a chart. For example, a letter is said to be so
many lines high, so many lines wide. One of the tiny squares should be
carefully divided into two, or, if possible, four parts, so as to ensure
finer and more accurate measurement. A letter may then be measured in
parts of a line, being described, for example, as, height 6-3/4 lines,
breadth 2-1/2 lines. It is of course important that the same gauge of
ruled paper be used uniformly, otherwise the measurements will vary. If
the student has had practice in the use of the dividers and scale rule,
he may prefer to employ these, but the ruled paper and a finely pointed
lead pencil will be found sufficient for most purposes. A paper
specially prepared for surveyors, ruled in squares of one-eighth of an
inch may be obtained. For measuring the slopes of letters a transparent
protractor is necessary. The letters measured are all topped and tailed
small letters, and all capitals having a shank. Letters like _O_, _C_,
_Q_, _S_, and _X_ can only be measured approximately.

The method of applying the measurements of heights and angles of slope
is shown in the case illustrated by the table on page 15.

The subject of enquiry was a signature containing the letters _B_, _l_,
_k_, _b_.

The measurements of these letters in the forgery are given at the top of
the table, and show the height in lines and angle of slope in degrees.

The measurements of the corresponding letters in twelve genuine
signatures are shown in the table as Examples 1 to 12.

The total is averaged by dividing by twelve.

The presumption in favour of the suspected signature being a forgery is
strongly supported by the arithmetical result.

A difference of more than 2 per cent. in angle of slope, and 3 per cent.
in height may be safely relied upon as ground for suspicion, for it is
rarely that a man