824 



Fishery Bulletin 92(4), 1994 



£: 3300 



o 

 O 



10 



Point number 



Figure 6 



Comparison of parallel fine-scale (open circle) and coarse-scale (solid circle) 

 transects of the outer margin of an adult Nemadactylus macropterus sagitta. 

 Distance between the parallel scans was approximately 20 urn. Fine-scale 

 analysis was done with a 6-um diameter beam at 8-um intervals, center to 

 center, and with a beam-power density of 2.92 uW unr 2 ; coarse-scale analysis 

 was done with a 14-um beam at 16-um intervals and with a beam-power den- 

 sity of 2.44 uW urn -2 . Concentrations are reported in units of ppm by weight 

 for all elements other than Ca, which is reported as percent by weight. The 

 large disparity between the two analyses in Na and Ca concentrations is an 

 expected consequence of the different beam-power densities used in the two 

 runs (see Gunn et al., 1992). 



variation in the three specimens depicted range up 

 to 45.4 for Sr, 43.4 for CI, and 31.3 for S (as opposed 

 to, at the other extreme, 1.8 for Ca). Fourth, ontoge- 

 netic patterns in the variation are often consistent 

 across specimens. All N. macropterus that we have 

 analyzed, for example, show steep gradients in Sr 

 levels in the region immediately around the primor- 

 dium. Similarly consistent, though less pronounced, 

 patterns are evident in Na and Ca. 



The quality of these data were assessed by com- 

 paring life history scans from the left and right 

 otoliths from the same fish. The quality of the match 

 within each otolith pair differs markedly (Fig. 7). The 

 comparisons suggest two principal sources of error. 

 First, there is consistent evidence of the difficulty of 

 tracking identical growth trajectories even within a 

 pair of otoliths from the same individual. In all three 

 pairs, the match between left and right otoliths de- 

 teriorates as the otolith margin is approached. We 

 attribute this to the decline in the growth rate of the 

 otolith with age, a corresponding compression of on- 

 togenetic variability and, therefore, a larger effect of 

 errors in tracking through the growth axis on the 

 apparent ontogenetic pattern of composition. Slight 

 differences in the shape of the otoliths also give rise 

 to differences in the length of each section, and hence 



the spacing of scan points relative to the distance 

 along the growth axis. Most of the left and right dif- 

 ferences in specimens #304 and #312 appear to re- 

 sult from these tracking errors; that is, the same 

 ontogenetic patterns and mean concentrations are 

 generally evident but variously expanded or com- 

 pressed along the growth axis. Second, in four of the 

 six elements examined (Na, K, S, and CD, mean con- 

 centrations occasionally differ between left and right 

 otoliths over relatively large portions of the otoliths. 

 This second source of error is difficult to assess. The 

 mismatch is most evident for CI in #339 and S in 

 #312, where the scale of the mismatch greatly ex- 

 ceeds machine-induced measurement error. The pat- 

 tern of the mismatch varies widely and inconsistently 

 among the samples: for example, CI levels match well 

 in #304, match intermittently and poorly in #312, 

 and differ markedly near the margin of #339, whereas 

 S matches well in #304 and #339 but very poorly in 

 #312. Comparisons of parallel life history scans 

 across a single otolith (e.g. Fig. 6) suggest that dif- 

 ferences between otolith pairs of the magnitude ob- 

 served cannot easily be attributed to either measure- 

 ment error or slight differences between otoliths in 

 the position of the scan line relative to the main 

 growth axis. We conclude, therefore, that the differ- 



