signatures, and the elements of the composite were also measured. The agreement of these with those of the disputed signature was extraordinary, and tends to show the value of the method of study by composite photography.

Ratios of Parts of a Letter. It is not only in the ratios of parts of words or sentences that opportunity is afforded to identify the writer by his unconscious tendency to preserve relations which are hidden from the eye, but even in a single letter. As an example of this, a study was made of the relative lengths of the tails of a number of ys from the bottom of the loop to the junction with the first part of that letter, and of the lengths from the bottom of the loop to the crossing of the down stroke by the up stroke of the tail. The annexed table

will show the results. The first six lines give the page, line, word, and letter of the specimen from a known writing which was measured. The last five lines give similar data for the specimens taken from an anonymous letter. In the latter case the additions, averages, etc., are added and recorded upward, so that the results may be brought into comparison with the similar results of the first table which are brought down. The agreement within, practically, one per cent. is strong evidence of the identity of the writer of both letters.

This difference of ratio is very small, and far within the error of observation.

This test would tend to show that the two letters were by one hand.

Character of such Writing.-Cases frequently occur where it is admitted that the hand of the person who wrote a signature was guided by the hand of some one else.

The writing produced by two hands conjointly is usually erratic, and, at first sight, hard to connect with the handwriting of any one person. In appearance it changes abruptly from very high or very wide to very low or very narrow letters.

Theory. This is to be explained by the non-agreement in phase of the impulses due to each of the two

writers. If both are endeavoring at the same moment to write a given stroke in the same direction, the length of that stroke will be measured by the sum of the impulses given by the two writers. If they act in opposition to one another, one seeking to make a down stroke while the other is trying to make an up stroke, the result will be a line equal to the difference between the stronger and the weaker force.

As these coincidences and oppositions occur at irregular but not infrequent intervals, like the interference and amplification phases of light- and sound-waves, the result traced on the paper might be expected in advance to be, as it is, a distorted writing where maxima and minima of effect are connected together by longer or shorter lines of ordinary writing.

The tabular statement by numerical average of one case will be shown and more specifically explained elsewhere. Manner of Guiding.-The usual and perhaps only legitimate reason for guiding a hand executing a legal instrument is the feebleness or illness of its owner.

When such assistance is required it is usually given by passing the arm around the body of the invalid or otherwise incapacitated writer, and supporting the writing hand while the necessary characters are being made.

Both participants in this action are looking at the writing, and both are thinking of the next letter which must be written, and of the motion of the pen necessary to produce it. Unless the executing hand were absolutely lifeless or entirely devoid of power, it would be

impossible for it not to influence the guiding and presumably stronger hand; for the least force exerted cannot fail to deflect a hand, however strong, in an unnatural and cramped position. Nor can the hand of the guider fail to add its contribution to the joint effort, however much the brain which controls may strive to render it entirely passive. Both minds are busy with the same act, and insensibly both hands will write the same letter with the results just described.

Analysis of Guided Writing.-Can the characteristics of each hand be separated from those of the other and the relative amount of the two contributions to the joint signature be stated?

From the comparatively small number of experiments made by the author in this direction it would be hazardous to answer it by an unqualified affirmative, but it may be said that some of the characteristics of each hand can usually be made apparent by the system of measurement, and the indications seem to point to the probability of an increase in the number of characteristics elicited in proportion to that of the observations made. If the significance of every part of every stroke could be properly interpreted, a complete separation of characteristics might be effected, but this would require an indefinitely large number of observations and a quite unattainable skill in explaining them.

Example.-Table II. contains the data of the study of a disputed signature.

Columns one to thirteen inclusive refer to measure