(Kortlandt 192) and herring gulls (Marshall 1947; Paynter 1947). 
Time-specific analyses.--Our original question may be approached 
by examining the turnover data in another hypothetical population outlined 
in table 9. Under these exaggerated circumstances, the time-specific 
analysis of mortality data breaks down completely. Its use in ornithologi- 
cal literature appears to have been based on the following reasoning: 
(a) Survival rate equals the number alive at the start of one 1 
year divided by the number alive at the start of the previous year. s, = 2 
x 
(vo) The numerator and demoninator of this fraction can both a 
be multiplied by the mortality rate without changing the equation. S = 24x 
1aq 
(c) But the number alive at the start of each year times the 1°x 
mortality rate equals the number dying that year; do = 159, and dqsl4q, 
(d) Therefore, survival rate can be found by simply dividing , 
the deaths of one year by the mumber of deaths in the preceding year. s,=72 
Table 9,.--Dynamic Versus Time-specific Computations in a Theoretical 
Population. Based on the use of mortality data only 

Symbols i. dx q Ss a .__ we 
Population Dynamic Life Table Time-specific 

Number Number Mort- Sur- Life Table 
Age Alive at Found ality vival Mortality Survival 
Interval Start Dead Rate Rate Rate Rate 
Qu-} 1000 250 25% 75% 0 100% 
1-2 750 250 33% 67% 0 100% 
2-3 500 250 50% 50% 0 100% 
3-h 250 250 100% 0 0 100% 
The weakness in this line of thought of course lies in the fact 
that the mortality rate may differ from one year to another. Hence in 
step (b) above, instead of being able to use qy in both numerator and 
denominator, only qq and qj are available. In the consideration of 
constant mortality rates in table 7, the time-specific method stood. up 
well because both mortality rates Gq and qg equaled 70 per cent. In 
table 9 these mortality rates equaled 25 and 33 per cent respectively. 
The time-specific table broke down as a result. 
Buss (1946, pp. 76-77) has divided the number of birds shot in 
two successive years to show survival of game-farm ring-necked pheasants 
released in various states. His results almost certainly are not sur- 
vival rates and are subject to correction. Because ~9 equals survival 
36 
