Frazer.] '^O^ [jlay IS, 



across the intersection. When the paper is inclined, however, but few 

 of the rays of iiglit which reach the eye by reflexion from the intersection 

 traverse and lose rays by absorption from both ink films ; but the greater 

 number penetrate only the upper ink and do not suffer absorption by 

 the lower. 



The second is a method of judging whether or not two lines have been 

 made with the same ink and consists in passing over each in succession 

 prisms of red, yellow or blue glass (or two of these), and noting the num- 

 ber of millimeters through which it is necessary to move each prism 

 from the position where its thin edge is in contact with the mark to be 

 judged to that where the color is extinguished and the line is black. The 

 prism is pushed horizontally over the ink mark continually adding to the 

 thickness of the colored glass over the latter. When the line ai)pears 

 quite black the distance in mm. over which the prism has been pushed 

 is read off and compared with the number of mm. which the other line 

 requires to attain the same result. If the inks have the same colors these 

 results must agree. 



Third method. In 18S6 I read before the Society a paper on the use of 

 composite photography for the purpose of establishing the type of an 

 individual's writing and especially the signature. Since then the mechani- 

 cal difficulties in the way have been greatly lessened and the method has 

 given most valuable results in cases before various courts. But there 

 are many occasions where it cannot be employed for one reason or an- 

 other, and in such cases I have devised a system of measurement and 

 tabulation which accomplishes by figures what composite photography 

 established automatically by form. The older method may be called the 

 graphic average of the handwriting and the later the numerical average. 

 The advantage of the former is that it takes into account at once all the 

 elements of character, while the latter can deal only with comparatively 

 few, but in spite of this the results attained have been very interesting. 



By the system here alluded to a given number of heights, breadths and 

 angles of letters, and spaces between them and between words, are 

 selected and measured in a large number of undisputed signatures. The 

 same elements are then measured in the signature in dispute. The aver- 

 ages of all the elements in the genuine series is then compared with the 

 latter, and their agreement or disagreement will generally lead to a cor- 

 rect judgment as to the genuineness of the disputed signature. 



This method has given successful results in a direction which extends 

 the original idea to a study of "guided hands," and it has been possible 

 to extract from the columns of measurements, proofs of the existence of 

 characteristics of each of the separate handwritings. 



