-158- 



populations, where they underlie actuarial studies (Pianka, 1978). 



Deevey (1947) pointed out three ways in which life table may be 



developed: 1) by directly observing the age at death (d x ) for a large 



and reasonably random sample of the population; 2) by following the 



survival (l x ) of a large cohort (born more are less simultaneously) at 



fairly close intervals throughout its existence; 3) by obtaining the 



age structure from a random sample of the population, and inferring dx 



from shrinkage between successive age classes. Morris and Miller 



(1954), stated that only the second sort of information is 



statistically respectable. These authors constructed a life table for 



the spruce budworm, Choristoneura fumiferana (Clem.), using a 



modification of Deevey's second method. They used column headings 



conventionally employed in life tables (Southwood, 1978) to include 



x Age interval or life stage; 



l x The number surviving at the beginning of the age 

 interval stated in column x; 



d x The number dying within the age interval stated in 

 column x; and 



q x The number dying in the age interval divided by the 

 number of survivors at the beginning of the 

 interval (the rate of mortality). 



In addition, they inserted a new column, "d x f," between l x and d x in 



which to record the factors responsible for death (d x ). This placed 



the emphasis on the causes of mortality rather than the order of dying 



(Morris and Miller, 1954). 



The life table constructed in Table 3-7 for J_u. diabol ica is 



patterned after that of Morris and Miller (1954) but differs in two 



respects: 1) It is based on a cohort of laboratory-reared individuals 



rather than on a naturally occurring population. Although it may have 



