FISHERY BULLETIN: VOL. 86, NO. 2 



discernible without any further preparation. 

 Sagittae from fish >25 mm SL were mounted in 

 epoxy resin and were ground, above and below, 

 with carborundum paper (600 grit). The resulting 

 thin section was secured to a microscope slide 

 with epoxy resin and etched with 6% EDTA (pH 

 7.0). Both the grinding and etching procedures 

 were monitored periodically by viewing the 

 sagitta under a dissecting microscope. 



The sagittae were then viewed under a Zeiss 

 compound microscope with transmitted light. The 

 number of growth increments were counted from 

 the image projected by a drawing tube onto a 

 Zeiss MOP Digital Image Analyzer System. 

 Under transmitted light each growth increment 

 was composed of a light and dark ring (Fig. IB), 

 which corresponded to the heavily calcified incre- 

 mental zone and the organic-rich discontinuous 

 zone of Watabe et al. (1982). Depending on the 

 size of the otolith, magnifications used ranged 

 from 400 X to 1,000 x . Three counts were made on 

 one of the 2 sagittae from each larva or juvenile, 

 and those otoliths with a repeatable increment 

 count of >90% were used in the growi:h analysis. 

 The other sagitta was counted once for compari- 

 son, as were the 2 lapilli. The number of incre- 

 ments on the 2 asterisci also were enumerated. It 

 was found in the previous study (Bolz and Lough 

 1983) that the asterisci were not detectable at 

 hatching, in contrast to the sagittae and lapilli, 

 but appeared later in the larval period. This was 

 reflected in the asterisci having on average 27 

 fewer growth increments than the sagittae. In 

 those instances where the sagittae and lapilli 

 were particularly difficult to read, the number of 

 asteriscal increments plus 27 was consulted as an 

 additional check. Maximum and minimum di- 

 ameters and planar surface area of the entire 

 otolith were measured routinely on all sagittae, 

 lapilli, and asterisci. 



The differential shrinkage of Atlantic cod and 

 haddock larvae and juveniles with respect to 

 standard length was corrected using Theilacker's 

 algorithm ( 1980), which is specified and discussed 

 in Bolz and Lough (1983). All lengths referred to 

 in the results and discussion portions of this 

 paper are reported as corrected lengths. 



RESULTS 



Haddock Larval and Juvenile Growth 



From analysis of the 189 larval and juvenile 

 haddock, ranging from 3.5 to 123.4 mm SL, we 



found that growth was best described by a 

 Gompertz-type curve. Previous uses of the Gom- 

 pertz growth curve and methodology for fitting 

 the curve are described in Pennington (1979), 

 Lough et al. (1982), and Messieh et al. (1987). The 

 variance was stabilized by using the natural log 

 form of the growth equation, and parameters 

 were derived by nonlinear estimation techniques 

 resulting in the relationship: 



ln(L) = 1.1987 + 4.8438(1 -e 



0.0088R\ 



(1) 



where L = standard length in mm, and 



R = number of days (increments) from 

 hatch. 



A plot of the Gompertz curve fitted to the natural 

 log of standard length vs. age in days is shown in 

 Figure 2. 



The predicted hatch-length fi-om the curve of 

 3.32 mm falls within the range of previous studies 

 (Colton and Marak 1969; Fahay 1983). An aver- 

 age growth rate of 0.24 mm/day (Table 2) for the 

 first 30 days is also reasonable (Laurence 1978; 

 Laurence et al. 1981) and agrees with the earlier 

 study of Bolz and Lough (1983). As a generalized 

 model the Gk)mpertz equation described haddock 

 growth through the first six months (175 days), at 

 which point it intersected (Fig. 3) the von Berta- 

 lanffy growth curve generated fi'om an analysis of 

 adult haddock by Clark et al. (1982): 



Table 2. — Mean standard length at age, 95% confidence limits, 

 and growth rate (mm/day and %/day) of larval and juvenile had- 

 dock from hatch through 200 days estimated from the Gompertz 

 growth model fit. 



226 



