NOTE Pooler et al : Assessment of sampling methods to estimate egg density of Limutus polyphemus 



701 



S 0.6 



a 0.4 



0) 0.2 



0.0 



500 



1000 



1500 



Density of eggs per core 



Figure 2 



The relationship between density and coefficient of variation (CV) for (A) shallow 

 sediment ( 0-5 cm ) and ( B ) deep sediment ( 0-20 cm ). Curves in each figure depend 

 on sample size: circle is n = 20, triangle is a; = 40, square is n = 60, diamond is 

 n = 80, and x is n = 100. Vertical lines represent median egg densities that we 

 observed in 1999. 



09 



= 0,6 



0.4 



3 4 5 



Rate of decline over five years 



0.3 04 0.5 



Rate of decline over ten years 



Figure 3 



Probability of detecting a decline (i.e. statistical power) for various magnitudes 

 of decline and for annual surveys over five (A) and 10 IB) years. Statistical power 

 was calculated for a one-tailed f-test with a type-I error rate of 0.2, and a constant 

 annual rate of change. 



it is important to realize that if sampling is restricted to 

 a short segment of beach, then the scope of inference is 

 strictly limited to that segment. If a rehable estimate of 

 egg density along a beach is required, then it will be nec- 

 essary to take samples along the entire beach. Because of 

 the logistics of sampling sediment it would be difficult to 

 sample throughout a long stretch of beach in one stage of 



sampling. However, a two-stage sampling design could be 

 considered in which beach segments are selected at the 1*' 

 stage and sediment cores within segments are selected at 

 the 2"'' stage. 



Consistent with our findings on sampling within a beach, 

 bay-wide egg density can be more precisely estimated for 

 eggs 0-20 cm deep than for eggs 0-5 cm deep. A stratified 



