Williams et al : Scales of spatial variation in demography of a large coral-reef fisfi 



677 



10 



^ 0.5 



o 



5 



I 



<D 

 O) 

 (8 





0.0 



-0.5 



-1.0 



19 21 22 



32 20 20 ^l 66 ^ 



tr 



27 10 



'^ 5 



8 10 12 14 

 Sectioned age (years) 



16 



20 22 



Figure 2 



Mean difference in age estimates between whole and sectioned otoliths 

 plotted against estimates of age from counts of annuli in sectioned oto- 

 liths of L. miniatus. Horizontal line represents no difference between 

 whole and sectioned readings. Sample sizes are shown for each age class. 

 Error bars are 95% confidence intervals. 



Age structure 



The youngest fish sampled from any reef was two years 

 of age, suggesting that L. miniatus becomes vulnerable to 

 standard line fishing gear at this age. All age-frequency 

 distributions for individual reefs were unimodal and the 

 most common mode was 6 years of age. This age is thus 

 assumed to represent the age at which L. miniatus is fully 

 recruited to the sampling gear. The relative frequencies of 

 the seven age classes 4 to 10* were not significantly differ- 

 ent among reefs within each region (Townsville; x^= 23.59, 

 P=0.17: Mackay; x'^=27.97, P=0.06; Storm Cay: x'-=20.29, 

 P=0.32). As a result, age structures from individual reefs 

 were pooled for each region (Fig. 4) and multidimensional 

 contingency tables were used to test for regional differences 

 in age structures. The relative frequencies of the seven age 

 classes were significantly different among all three regions 

 (all regions: x-=193.31, P<0.0001, Townsville vs. Mackay: 

 X-=172.70, P<0.0001; Townsville vs. Storm Cay: x'-^=91.88, 

 P<0.0001; Mackay vs. Storm Cay: x^=22.27, P=0.001). The 

 most obvious difference among regions was the greater 

 relative abundance of older fish (>6 years) in the Towns- 

 ville region than in the Mackay and Storm Cay regions 

 (Fig. 4). However the oldest fish were from the Storm Cay 

 region, where a small number offish persisted in the older 

 age-classes up to 19 years of age. The relative abundances 

 of age classes 4 and 5 were greater in the Mackay region 

 than in the Townsville and Storm Cay regions (Fig. 4). 



Mortality 



Estimates of annual total mortality rates (Z) for individ- 

 ual reefs were generally similar among reefs within each 



600 

 500 

 400 

 300 



200 

 100 



A Townsville 



« * 

 : 



:.r' 



B Mackay 



5 10 15 20 



Age (years) 



Figure 3 



Least-squares linear regression of 

 otolith weight on age for L. miniatus 

 from three regions of the Great Barrier 

 Reef: (A) Townsville, (B) Mackay, and 

 (C) Storm Cay. 



region, with the exception of the Storm Cay region where 

 the estimated Z appeared much lower for reef 21-131 than 

 for other reefs in that region (Table 1). ANCOVA indicated 

 no significant difference in mortality among reefs in any 

 region (Townsville: F3 3o=0.80, P=0.50; Mackay: F3 2o=0.08, 

 P=0.97; Storm Cay: P.j ,5=1.14, P=0.37). Therefore mortal- 

 ity rates were estimated for each region from the pooled 

 age structures for all reefs within each region (Fig. 5). A 

 comparison among regions of the regression slopes from 

 the pooled age structures indicated significant differences 

 among regions {F2^g=T-ll, P=0.005). Tukey's multiple 

 comparison tests revealed that the estimated Z for the 

 Townsville (Z=0.42) and Mackay (Z=0.71) regions differed 

 significantly, whereas the estimate from the Storm Cay 

 region (Z=0.60l did not differ significantly from either 

 Townsville or Mackay. 



