326 



Fishery Bulletin 94(2). 1996 



sideration. Mortality estimates derived with the 

 catch-curve regression method are least sensitive to 

 minor violations of this assumption (Kicker, 19751 

 and this is particularly evident inL. quinquelineatus. 

 Despite this, it is evident from Figure 7 that mortal- 

 ity rates of L. quinquelineatus are variable among 

 reefs and are similar to those for populations of L. 

 adetii in that they are characterized by low rates of 



total mortality and high rates of survivorship. How- 

 ever, mortality estimates cannot be derived with 

 great confidence from populations where significant 

 recruitment variability is retained in the age struc- 

 ture (e.g. Ferreira and Russ, 1995), and although 

 Ricker (1975) has suggested that irregularities in 

 catch curves caused by variable recruitment can be 

 reduced by combining samples over successive years, 



, 3 999 - I T> < 



r- I) 729 



(ages 8-24 \i-;irs only) 



16 2H 



Kelso 



v - 3 979 



, 0.795 

 (ages l >- 15 years only) 



Aye (years) 



Figure 6 



Catch curves for L. adetii among reefs, based on 



observed ayes 



