Kitada and Tezuka Siirvey designs for estimating recreational fishery catch 



243 



R,= 



C(s) 



By using a method similar to that given in Appendix 3, we 

 get 



Cov(R,.,R.,)- 



N' 



1 



M,M,. 



-cov(c;-'",c;f') 



M,A/^. 



-CoviCl",M,.) 



J^'^ Cov(M„C;r') 



Cov{M,,,Mi,) 



Cov(Q'",Af^,.) and Cov(M,,,Cj.?'). These are the covariances 

 between a total estimate and a sample mean. By a method 

 similar to that in Appendix 1, we have 



Cov(i-,y ) = EiNX - NiJ^)(Y - fi^.) 



= NCov(X,y ) = Cov(X.y). 



The covariance is estimated by 



Cov(i'T)= ^ " y{X~X)(Y-Y). 

 n(n - 1) f^ 



For our case, the two covariances are as follows; 



n (« - 1) ■"' 



Here Cov(C'i'",C'i'')and CovlM^M,. I are already given by 

 Equation 18 and Equation 24. Hence we can estimate 



Cov(M, , CI? ' ) = !y " X ' C,*- - Q- ) (^,* - M, 

 nin- ll •^ 



