296 



METHODS OF INTERPOLATION. 



The result is shown in cohmm (/).* This is a roughly adjusted series, 



approximating' closely to the form of the original series {d); too closely, 



however, for it retains at least one undulation which is abnormal, and 



would doubtless not have appeared if the number of observations on 



which the earlier portion of series {(1) is based had been very greatly 



increased. It is an acknowledged principle that after the age of 12 or 13, 



at which the probability of dying within a year is a minimum, the rate 



of mortality ought to go on increasing continuously up to the limit of 



old age. But in series (/) it increases up to the age 22, then diminishes 



up to 25, then increases agqin continuously. To remedy this fault, and 



also to perfect the graduation, some further process of adjustment will 



be required. 



Table II. 



*In all the terms of series (d), (c), and (/), the fifth figure might as weU have been 

 neglected. It has no real value, aud does not assist the graduation. 



