Leffler and Shaw Age, growth, and mortality of larval Chloroscombrus chrysurus 



713 



a random subsample of 50 fish were measured, as well 

 as the shortest and longest. Ethanol-related shrinkage 

 was assumed to be uniform for each fish collected and 

 preserved (3-min tow, alcohol preservation; see Radtke 

 1989). 



Validation, age, and growth 



Sagittal otoliths were removed from each Atlantic 

 bumper larvae using a dissection microscope equipped 

 with polarized light. The sagitta from nine postlarval 

 and juvenile Atlantic bumper (8. 3-25. 0mm SL) that 

 were immersed in the calcein-seawater solution were 

 prepared and viewed using the method described by 

 Wilson et al. (1987). Growth increments, following the 

 fluorescent mark, were counted at 400 x and verified 

 at 1000 X . The number of growth increments counted 

 from the calcein mark to the otolith edge were com- 

 pared with the number of days fish were held in cap- 

 tivity after marking. 



Age estimation of larval Atlantic bumper was per- 

 formed using sagitta that were air-dried and mounted 

 in S/P Accu-mount 60 on a glass microscope slide. Most 

 larval otoliths were thin enough that only viewing 

 under a compound microscope was necessary to make 

 total increment counts and otolith radius measure- 

 ments. A few larger otoliths were ground with 600 

 WetorDry grit sandpaper and polished using 0.3/:/ 

 Alumina 2 Alpha Micropolish until growth rings were 

 countable. The counting and measurement procedure 

 was enhanced by using a digital imaging system which 

 produced images on a video monitor at 400 x or 1000 x . 

 Independent increment counts were made twice by the 

 same person without knowledge of fish length or 

 previous otolith count. Only otoliths for which replicate 

 counts were identical were used in the analysis. Eleven 

 of the 170 otoliths prepared were discarded. 



Separate linear growth equations of standard length 

 on increment number were developed for fishes col- 

 lected on the five cruises. These five equations were 

 compared using analysis of covariance (ANCOVA, 

 a 0.05; SAS Inst. 1985). Exponential and other non- 

 linear models (e.g., Laird-Gompertz) used to describe 

 larval growth were also tested (Campana and Neilson 

 1985). A General Linear Model ANOVA, followed by 

 a multiple comparisons test (Duncan, a 0.05; SAS Inst. 

 1985), were used to detect differences in surface-water 

 temperature between years, months, and cruises. 



Zooplanl<ton biomass 



Zooplankton displacement volumes (mL/m''^) were 

 determined (Yentsch and Hebard 1957) for each net 

 tow. A mean zooplankton standing stock value was 

 then calculated for each cruise and net mesh type. A 



simple regression of zooplankton standing-stock values 

 (202 vs. 333^im mesh nets) was developed for both 

 September 1987 cruises. ANCOVA (a 0.05) was used 

 to test for differences between the two cruises. The 

 data from the two cruises were combined into one zoo- 

 plankton standing-stock regression to standardize the 

 values from the two mesh sizes. 



Dry weight-length relationship 



Larval and juvenile Atlantic bumper (N 120, 8.0-32.0 

 mmSL) collected by dipnetting for jellyfish, were 

 measured to the nearest O.lmmSL, oven dried for 

 6h at 62°C, and then weighed to the nearest 1.0 mg. 

 A log-log dry weight-length relationship was estab- 

 lished and described by the equation W = aL'^, where 

 W = logio dry weight (mg), and L = logio standard 

 length (mm). A 95% confidence interval placed around 

 the estimated slope (b) was used to test for differences 

 in the estimated length power term (b) and the classical 

 b estimate of 3.0 for adult fish (LeCren 1951) and 4.0 

 for larval fish (Power 1989). 



Mortality 



Atlantic bumper densities for each ImmSL category 

 were converted into mortality estimates following the 

 length-frequency method described by Essig and Cole 

 (1986). Sampling with respect to the windowshade 

 drogues allowed us to monitor larval densities from the 

 same mass of water for an entire collecting period. Only 

 nighttime collected larvae >2.0mm or <5.0mm were 

 utilized in our mortality estimates to minimize biases 

 from net avoidance by larger larvae or extrusion 

 through the mesh openings by the smallest larvae. The 

 descending limb of each age-frequency distribution cor- 

 responding to a length range of 2.0-5.0 mm SL was 

 described by the equation Dt=Doexp'"'^'>, where M = 

 the instantaneous daily mortality coefficient, Dt = 

 larval fish density at time t, Do = larval fish density in 

 the first fully recruited group (i.e., time = 0), and t = 

 time in days (Peebles and Tolley 1988). Mortality 

 estimates were tested for statistical differences be- 

 tween cruises and years using ANCOVA (a 0.05). 



Results 



Validation, age, and growth 



Daily increment formation on Atlantic bumper sagitta 

 was validated using calcein. Each otolith from the nine 

 fish treated had distinct growth increments between 

 the green fluorescent calcein mark and the edge of the 

 otolith (Fig. 2). On each sagitta examined, the number 

 of increments counted after the calcein mark was 



