PALM AND SOLE IMPRESSIONS. 387 



definite though arbitrary limits, themselves the result of much experi- 

 ence ill measurements, and designed to divide any average set of 

 measurements into three approximately equal divisions, rather than to 

 divide equally the range of millimeters between the extremes of a given 

 measurement. Thus, to quote an example furnished, "the numerical 

 limits of the 'medium' head-length as used at the Prefecture of Police 

 in Paris include an interval of but G millimeters (185-190), while 

 those included under 'large' extend from 191 mm. to the greatest 

 dimensions possible, an extent of more than three centimeters."* 



Now if we were to conceive of each one of these eleven measure- 

 menfs as varying independently of one another and as being divided 

 into three subdivisions, the number of possible subdivisions under 

 which individual anthropometric records could be filed would reach 

 the large number of 3 to the 11th power, or 177,147, but in practical 

 use M. Bertillon employs for purposes of identification only a few of 

 these measurements, which he gives in the work just quoted, together 

 with an hypothetical application, as follows: 



He supposes the case of 90,000 sets of measurements, a number 

 approximately corresponding to that of the adult male prisoners re- 

 corded in the Paris prisons up to 1893. Of these the first classifica- 

 tion is made by means of the length of head, and as the subdivisions, 

 small, medium and large, are fixed with reference to equality of 

 division, approximately 30,000 of these records will be placed in each. 



Each of these subdivisions is now divided again into three parts, 

 in accordance with the breadth of head, a division which leaves ap- 

 proximately 10,000 in each of the nine compartments, i. e., 10,000 

 individuals whose head length and head breadth fall into the same 

 categories. The third division, which reduces the number of records 

 in each of 27 compartments to about 3,300, is based upon the length of 

 the left middle finger, and the fourth, resulting in 1,100 in each of 

 the 81 compartments, is based upon the length of the left foot. The 

 length of the cubitus then follows, which increases the number of com- 

 partments to 243 (=3^) and the number of individual cases in a 

 compartment to less than 400. By the addition of the standing height, 

 the number of compartments is increased to 729 (=3^) and that of 

 the cases in each compartment to approximately 130, and these num- 

 bers become respectively 2187 (=3'^) and 42 by the use of measure- 

 ments taken from the left little finger. These are finally reduced to 

 small sets of a dozen records each by such criteria as the color of the 

 eye and the length of the right ear, after which this small number may 

 be carefully compared for individual measurements. 



With some modifications the above system is in official use in most 

 of the civilized countries of the world, including England, Eussia, 



* Bertillon, A. ' Instructions Signal^tiqiies,' 1893. Introduction. 



