H. Waitb 
423 
number of radial loops except in the forefingers. A comparison of the distribution 
in the two hands shows considerable differences ; e.g., in the left thumb the num- 
ber of arches is about double the number in the right ; again, the whorls in each 
finger of the right hand are greatly in excess of those on the left, while the left 
hand has, in every case, an excess of ulnar loops. 
If we arrange the numbers of each class in order of magnitude, we see that the 
order for the arches is identical for the two hands and also for the ulnar loops. 
In each of the other classes there is one exception to the " identical " order. 
I have tested these distributions for each type by the method referred to in the 
footnote of p. 422 a, with the following results : — In the arches the odds are more 
than 500 to 1 against the occurrence of two such divergent samples which are 
random samples taken from the same population ; in the ulnar loops the odds are 
more than 200,000 to 1 ; in the radial loops about 5 to 2 ; in the whorls more than 
1,000,000 to 1, and in the composites more than 1300 to 1. 
We may thus fairly conclude that with the exception of the radial loops the 
frequency distribution of the classes between the fingers is different in the two 
hands and the radial loops are so few, except in the forefinger, as to be almost 
negligible. 
6. Subdivision of Loops. The great preponderance in the number of loops 
and the insignificance of the number of radial loops, except in the forefinger, make 
another subdivision of this class necessary. The method adopted is as follows : — 
All loops, in common with whorls and composites, contain certain well-defined 
points; these are (1) the "delta," or "outer tei'minus," and (2) the " point of the 
core," or " inner terminus." [See Henry, pp. 22 — 24.] The number of ridges 
intervening between the delta of a loop and the point of the core may be anything 
from one up to about thirty ; in only 38 cases out of the 13,095 loops does the 
number of ridges exceed 25 ; two of these are over 30, one being 32 and the other 
35. The complete distribution of ridges is given in Table 4 a. 
In dividing the loops into two sub-classes according to the number of ridges 
the nearest approach to equality is obtained by taking (a) those containing from 
1 to 12 ridges, and (b) those containing 13 or more ridges. For brevity I have 
called these classes (a) Small Loops, and (6) Large Loops ; the terms " Small " and 
"Large" have no reference to the relative sizes of the patterns. The numbers in 
the two groups, thus arranged, are 7033 and 6062 respectively. 
Table 4 6 gives (1) the number of loops for each finger, (2) the means, (3) the 
standard deviations, and (4) the coefficients of variation in the numbers of ridges. 
Examining the Table below consider first the means. The order which is 
identical in the two hands runs : 
(1) Thumb, (2) Ring Finger, (3) Little Finger, (4) Middle Finger, (5) Index. 
It will be noticed that this order of the means is quite different from that 
of the relative areas of the patterns. 
