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



675 



Figure 1 



Location of reefs sampled for L. miniatus from October 1995 to January 1996 in the 1) 

 Townsville, 2 ) Mackay, and 3 ) Storm Cay regions of the Great Barrier Reef, AustraUa. 

 Reefs maps indicate the relative position of the four reefs closed to fishing that were 

 sampled in regions 1, 2, and 3. 



open to line and spear fishing (referred to as "open reefs"). 

 Fishing had been prohibited from the closed reefs for at 

 least seven years prior to sampling. Each reef was sampled 

 for two days by the same four commercial line fishermen 

 using gear and sampling designs standardized across all 

 reefs (Davies et al.''). Fork length (FL) was measured to 

 the nearest millimetre immediately upon capture. Sagittal 

 otoliths were removed from frozen frames in the laboratory, 

 cleaned of any residual material, dried, and weighed to the 

 nearest 0.1 mg. 



A total of 1015 L. miniatus were collected from the four 

 closed reefs in each region between October 1995 and Janu- 

 ary 1996. Sample sizes from the open reefs were small and 

 mortality and growth estimates from these reefs were 

 unreliable. Therefore, these samples were used only to in- 

 crease the sample size of older fish for a comparison of the 

 two methods for reading otoliths (whole and sectioned). 



Comparison of otolith reading methods 



The annual periodicity of opaque increment formation in L. 

 miniatus otoliths has been validated (Brown and Sumpton 

 1998). A subsample of 355 L. miniatus otoliths from both 

 open and closed reefs was used to assess whether readings 

 of whole otoliths provided age estimates similar to those 

 from sectioned otoliths, but at substantially lower cost (in 

 time). Otolith weight was used to select a broad range of 

 age classes for this assessment on the assumption that 



otolith weight was a coarse indicator of age, thus avoiding 

 the need to preread otoliths to obtain a sample covering 

 all age classes. Each otolith in the subsample was read, 

 both whole and sectioned, on three separate occasions in 

 random order with no prior knowledge of collection date, 

 location, or fish size. For consistency, the right otolith was 

 chosen to estimate the age of all fish unless it was missing 

 or damaged, in which case the left one was used. Otoliths to 

 be read whole were placed in a small black dish of immer- 

 sion oil and examined under reflected light with a stereo 

 dissecting microscope. Counts of opaque increments were 

 made from the nucleus to the dorsoposterior edge on the 

 convex face of the otolith. For otoliths from older fish it was 

 necessary to rotate the otolith approximately 45° to clearly 

 observe increments on the otolith margin. 



Otoliths to be sectioned were embedded in epoxy resin 

 and cut transversely, adjacent to the anterior side of the 

 nucleus with a Buehler Isomet low-speed saw. The poste- 

 rior portion of the otolith was retained and mounted on a 

 glass microscope slide with Crystalbond adhesive. A second 

 transverse cut adjacent to the posterior side of the nucleus 

 resulted in a thin section, incorporating the otolith nucleus, 

 remaining on the slide. Otolith sections were then ground 

 on 800- and 1200-grade sandpaper to remove saw marks 

 and a single drop of immersion oil was placed on sections to 

 fill surface irregularities. Otolith sections were examined 

 under a stereo dissecting microscope with reflected light 

 and a black background. Counts of opaque increments were 



