ON THE NOETII-WESTEKN TRIBES OF CANADA. 



437 



of observations great variations become more probable, and smaller ones 

 consequently less probable. Or the same fact may be expressed in this 

 ■way : — the limits of variation are probably the wider, the greater the 

 series of observations. Therefore the curve computed from a long series 

 is by no means the same, not even theoretically, as that computed from 

 a shorter series. 



