FISHERY BULLETIN: VOL. 78, NO. 3 



Table 4. — Estimated parameters for models that predict live 

 northern anchovy body part size (L) from net-treated size (X,). 



Model' 



P3 



Standard length 

 Head length 

 Body depth 



0.289 



0.177 

 0.413 



0.434 



3858 



29.746 



-0,680 

 -0605 

 -0620 



'Equation (4), see text. 



ters for the standard length, head length, and body 

 depth models are given in Table 4. These equations 

 can be used to estimate live size of each body part 

 from measurements of northern anchovy larvae 

 after net capture. 



Preservation Shrinkage 

 (After Net Treatment) 



After larvae shrank during net treatment, addi- 

 tional shrinkage caused by Formalin was nearly a 

 constant proportion of length (Figure 7). The ratio 

 of preserved length to net-treated length (Figure 

 7) may decrease slightly (i.e., shrinkage may in- 

 crease) with increasing fish size, but this slight 

 decrease has no practical significance for length 

 calibration of larvae taken in routine plankton 

 samples. Because ossification begins at 14 mm SL, 

 I expect large fish would shrink less, not more, 

 than small fish. The overall mean ratio (preserved 

 size/size after each timed-net treatment) of 104 

 standard length measurements was 0.9668± 

 0.0020; percent shrinkage of the other body parts 

 was the same as standard length shrinkage. I rec- 

 ommend using 39c shrinkage for all body parts in 

 Formalin after net treatment. Preservation in al- 

 cohol after net treatment did not cause further 

 shrinkage (only standard length measured). 



To adjust the shrinkage models (Table 4) to pre- 

 dict live size from net-treated and Formalin- 

 preserved size (equivalent to field-collected lar- 

 vae), the preserved size is multiplied by 1.03. No 



6 8.0 10.0 12.0 14.0 16.0 



NET TREATED STANDARD LENGTH (mm) 



Figure 7. — Size specific shrinkage of northern anchovy larvae 

 preserved in Formalin after 10-min net treatment. Dots are 

 means of 10 larvae. Circles represent individual fish. 



adjustments are needed for alcohol-preserved 

 samples. 



The difference in shrinkage between 

 laboratory-preserved larvae and net-treated and 

 preserved larvae of the same initial live size de- 

 creased with age. For example, 3 mm larvae that 

 were net treated and preserved in Formalin 

 shrank IB'A more in standard length than 3 mm 

 larvae that were laboratory preserved in Forma- 

 lin, but shrinkage of 20 mm larvae was the same 

 for both treatments (Table 5). Shrinkage of 



Table 5. — Comparison of standard length for live (L), 

 laboratory-preserved and net-treated northern anchovy larvae 

 (X^). Numbers in parentheses are preserved length divided by 

 live length (ratio of 1.00 = no shrinkage). 



'Includes 30 s handling time; no shrinkage in alcohol and 8% shrinkage in 

 Formalin (see text). 



^Estimated preserved size calculated from Equation (4) for 10 mm net treat- 

 ment: size adjusted for 3% additional shrinkage in Formalin (see text and 

 Table 4). 



laboratory-preserved larvae in Formalin probably 

 decreases to something <8% (Table 5) as the skele- 

 ton develops and ossification occurs. Formalin pres- 

 ervation of 90 mm and larger salmon, Oncoryn- 

 chus spp., smolts caused 3-49^^ shrinkage in length 

 (Parker 1963). Laboratory shrinkage of northern 

 anchovy after transformation may be similar; 

 thus, shrinkage of net-collected and preserved 

 northern anchovy >35 mm (Table 5) should be 

 similar to shrinkage of laboratory-preserved fish 

 >35 mm. 



Eye Diameter 



Netting live larvae for 10 min caused the eye 

 diameter to shrink an average of 0.0443 ±0.0069 

 mm, and Formalin preservation after net treat- 

 ment caused an increase in eye diameter that av- 

 eraged 0.0177 ±0.0046 mm. The increase in eye 

 diameter after preservation was similar to the in- 

 crease after preservation noted for eye diameter of 

 laboratory-preserved larvae. The ^tests for paired 

 data (n=23) showed that in all cases the differ- 



690 



