19 
parisons are to be drawn. The reason for saj'ing in the above that the 
odds or chances are 2*57 to 1*0 is justified by the fact that it has been 
established mathematically that when a difference[divided by the probabb 
error of that difference is: 
1 • 0, the odds are as 1 • 0 to 1 ■ 0 l 
1- 6 2-57 to 10 
2- 0 4-64 to 10 
2-2 6-25 to 10 
2-5 9-90 to 1-0 
2-6 11-58 to 1-0 
2-7 13-58 to 10 
2-8 15-96 to 1-0 
2- 9 18-82 to 1-0 
3- 0 22-24 to 1-0 
4- 0 142-30 to 1-0 
5- 0 1341 -.00 to 1-0 
6- 0 19300-00 to 1-0 
7- 0 427000-00 to 1-0 
As it would clearly not be reasonable to place reliance on such slender 
odds as 2*57 to 1*0, it is, therefore, assumed that the mean difference of 
2*4 cm. in stature is not significant in the present instance, and on this 
account in the fourth column of table YI, under the heading of stature, a 
space has been left blank. It is, in fact, usual when dealing with such data 
as these to pay little attention to odds of less than 22 to 1 (which is the 
equivalent of a Diff./P.E. diff. of 3*0), for such odds give reasonable assur- 
ance that a difference is a genuine one. If a Diff./P.E. diff. be 4, 5, 6, or 
more, then surely when dealing with problems such as these, we have what 
is tantamount to proof that such a difference would be found to exist were 
we to measure entire populations and not be limited to relatively small 
samples of them. In table VI, a blank space has been left in the fourth 
column whenever a Diff./P.E. diff. has been found to be less than 2*2 (i.e., 
the odds are less than 6*25 to 1*0). 
The Three Bands of Chipewyan Indians 
The three bands of Chipewyan Indians, those from Fond-du-lac, 
Chipewyan, Fitzgerald, and Fort Smith, may first be compared with each 
other; and while this is being done it will be kept in mind that the maximum 
number of individuals in these three bands available for any measurement 
(or any index) is small, being respectively, thirty-three, fifteen, and twenty- 
four. In sections A, B, C, of table VI, each of the bands is compared with 
the other two. The first and second columns of each section record the 
mean measurements (and indices) of the two bands under consideration; 
the third column records the differences between these means; and the fourth 
column records the number of times the difference between the two means 
is greater than the probable error of its difference; no entries are made, 
however, unless differences are more than 2*2 times greater than their 
probable errors. 
'This is an excerpt from table 40, “Medical Biometry and Statistics,” by Ramond Pearl. 
