
the other hand, the normal curve gives estimates that are more in accord- 
ance with other observations, in particular with the know kill of 4-year- 
olds. 
The procedure and tables of Hald (1952) are used in estimating 
the parameters of the normal curve. One advantage of this procedure is 
that the escapement can be estimated without evaluating all of the 
parameters. 
The following four series have been analyzed in this manner 
to yield estimates of escapement: 

1950 i d/ 1950 2/ 1951 i12/ 1951" 
Round of 3-year- returns of 3-year- returns 
centered at--_. olds of A-tags olds = of B-tags 
June 20 a/ 7TA5 | 5/ 47 8/ 519 &/ 10 
25 1, 265 64 956 33 
30 3,083 132 799 69 
July 5 4,312 213 2,496 111 
10 4,938 251 4,593 161 
15 5,737 316 3,567 212 
20 5,597 324 7,651 331 
25 B= 10) | Saree Fe)” A A MR A! ce Sp A oP 
Total 31,746 1,671 27,960 1,263 
Estimated per- 
cent escapement 
(after July 27) 21 24 47 43 

1/ From table 30. 
From tables 26 and 27 combined. 
From table 31. 
By similar analysis to the 1950 A-tag returns. 
Including preseason. 
SS te Bes 
Two days only (postseason kill not included). 
-61l- 
