830 



Fishery Bulletin 92(4), 1994 



then increase towards the otolith margin in samples 

 from all three sites. Second, the pattern of ontoge- 

 netic variation in concentrations differs among ele- 

 ments. And third, similar mean differences are evi- 

 dent among sites irrespective of where the analysis 

 was done in the otolith. NSW/Victorian specimens, 

 for example, at all stages of their life histories to an 

 apparent age of at least 2-3 years tend to have CI 

 levels higher than those of fish collected elsewhere. 

 As a result, discriminant analyses based on mean 

 concentrations at points 36 and 80 result in site de- 

 lineations virtually identical to those derived from 

 concentrations measured near the primordium. 



Evaluation of signatures specific to nursery 

 areas and the links between nursery areas 

 and adult groups 



The links between nursery areas and spatial compo- 

 nents of the adult populations can be assessed in two 



complementary ways: 1) by determining the source 

 affinities (e.g. spawning site) of juveniles collected 

 in each nursery area and 2) by developing a specific 



P«0001 



P«0.001 



P<0.05 



NS 



NS 



*> **> $u ^ ty&, Q.}/, 



Source location 



Figure 13 

 Distribution among sites (see Fig. 1) of values of 

 each of the five discriminant functions defined 

 by the initial (six-site) discriminant function 

 analysis for adult Nemadactylus macropterus, 

 based on mean concentrations for points 2-6, in- 

 clusive. Horizontal bars indicate sites that pool 

 together based on post-hoc analysis (Steffe's F- 

 test). In the case of Function 3, the overall 

 ANOVA is just significant at P<0.05, but pair- 

 wise comparisons among sites indicate no single 

 or set of sites that differs consistently and sig- 

 nificantly from the others. NS=not significant. 



