H. Waite 
425 
Comparing the two hands we see that the differences in the middle, ring and 
little fingers are insignificant ; in the thumb and index, however, there is a marked 
difference in favour of the right hand. 
The order of the standard deviations in the right hand is : 
(1) Ring Finger, (2) Index, (3) Thumb, (4) Little Finger, (5) Middle Finger. 
In the left hand the order of the last two is reversed, but the difference is 
small. 
With the exception of the middle finger, where the difference between the two 
hands is only about equal to the probable error and is therefore insignificant, the 
standard deviation is in every case greater for the right hand than for the left ; 
the differences are all of the san\e order of magnitude and range from about 
•39 to -54. 
Coming now to the coefficients of variation — the order in the right hand is : 
(1) Index, (2) Ring Finger, (3) Middle Finger, (4) Little Finger, (5) Thumb. 
In the left hand the order of the ring and middle fingers is interchanged. 
Comparing the two hands we see that in three cases — the thumb, index, and 
middle finger — the differences are each less than the probable errors ; in the other 
two cases the variability is considerably greater in the right hand than in the left. 
I have carefully revised the calculations involved but have been unable to 
detect any error ; neither can I suggest a reason for the large differences. 
In "A First Study of the Variability and Correlation of the Hand" (see p. 421), 
the writers find that the variability of bone lengths is closely related to the 
relative utility of the fingers, the least variability being that of the most useful 
finger. There appears, however, to be no such simple relationship between the 
ridges of the loops and the relative utility of the fingers. 
I have compared the distribution of ridges in the loops of the thumbs by 
Professor Pearson's method (p. 422 a, footnote), which gives = 166'64 ; hence the 
odds are much greater than 1,000,000 to 1 against the occurrence of two such 
divergent samples if they were random samples taken from the same population. 
The distribution — absolute and percentage — of the five groups is now as 
follows (Table 5). 
In comparing the large and small loops it will be seen that in both hands there 
is an excess of large loops in the thumb, ring and little fingers, and an excess 
of small loops in the index and middle fingers. The order of these classes agrees 
in the two hands with one exception in each case. 
An approximate measure of the relationship existing between the various 
combinations of digits is given by the number of cases in which two particular 
digits on the same or on opposite hands have the same pattern. Table 6 a gives 
the percentages for the same hand and for digits of the same name on opposite 
