162 



Fishery Bulletin 88(1). 1990 



Figure 5 



Nonlinear regression of sagittal radius (SR) on days 

 after first feeding (days ./J) of Atlantic menhaden lar- 

 vae from control (solid line, -f ) and treatment (1-3 day 

 starved, broken line, D) tanks. Model coefficients and 

 other descriptive statistics are also given. Some sym- 

 bols represent more than one observation. 



significant (/)<0.0001) and resitluals were distributed 

 at random over the entire size range examined, in- 

 dicating a good fit of tlie logistic model to the log- 

 transformed data. The relationship between sagittal 

 radius and days after first feeding was also fit with a 

 logistic function (Fig. 5). The regressions were highly 

 significant (<0.001) and residuals were distributed at 

 random, but the variance increased with age of the lar- 

 vae. Predictions of sagittal radius fi-om the regressions 

 indicated that increment width increased from 0.6 \xm. 

 at first feeding to 3.8 ^m at 30 days in larvae from the 

 control tanks, while increment width increased from 

 0.7 tol.6 \ixvi in larvae from the treatment tanks. 



Discussion 



The results of our laboratory experiments indicate that 

 microstructural growth patterns in sagittae of Atlan- 

 tic menhaden can be used to accurately estimate age 

 from first feeding, detect short-term variations in 

 growth rate caused by starvation, and estimate the 

 growth chronologj' of individual larvae. Age of in- 

 dividual larvae from first feeding can be estimated 

 within +3 days during the first month of life (based 

 on inverse regression: Days after first feeding = o -i- 

 h (increment count); Draper and Smith 1966, Rice 

 1987). Atlantic menhaden larvae appear to initiate in- 

 crement formation at hatching, although growth incre- 

 ments formed prior to first feeding are narrow (< 1 ycca), 

 poorly defined, and could not be consistently resolved 

 with light microscopy. Poorly-defined increments ob- 

 served during yolk feeding in the Atlantic menhaden 



have been observed during similiar periods in other 

 species (Lough et al. 1982, McGurk 1984, Bolz and 

 Lough 1983). The formation of these increments may 

 be related to an immature circadian rhythm (Campana 

 1984b) and/or the presence of increments too narrow 

 to detect using light microscopy (Campana et al. 1987). 

 Formation of the first prominent growth increment in 

 sagittae at first feeding may be related to the shift 

 to exogenous feeding and related circadian activity 

 patterns. 



Transitions within the egg and larval stages of Atlan- 

 tic menhaden were characterized by particular micro- 

 structural features. Hatching was characterized by a 

 wide discontinuous zone, and the transition to exog- 

 enous feeding coincided with a prominent growth in- 

 crement. Similiar patterns have been observed in other 

 species (Brothers and McFarland 1981. Campana 1983, 

 Lagardere and Chaumillon 1988). Particular micro- 

 structural features formed during the early life stages 

 may be the result of changes in physiological metab- 

 olism, stress, and/or growth cycles that are generally 

 associated with these transitions. 



Studies that have examined the rate of increment for- 

 mation in fish larvae reared under conditions pro- 

 moting rapid growth have shown that growth incre- 

 ments form daily in most cases ( Jones 1986). Atlantic 

 menhaden larvae fed ad libitum formed increments at 

 a rate of unity, consistent with the hypothesis that in- 

 crements are formed daily in sagittae of well-nourished 

 fish larvae. Estimates of statistical power obtained in 

 our study indicate low variability about the relation- 

 ship between increment count and days after first 

 feeding and provide additional support that growth 



