FISHERY BULLETIN: VOL. 87, NO. 3, 1989 



tion or response time for each size, diet, and cell 

 height category is needed. The duration is the 

 number of days that larvae belonging to the size, 

 diet, and cell height category remain within that 

 category. Daily rates in the sea (estimates of 

 numbers of larvae per day) are determined by 

 dividing the number of field-collected larvae 

 classified to each category by the category dura- 

 tion. See the following section on "Applying the 

 Calibration Criteria" for a more detailed ex- 

 planation of this procedure. 



Duration for Fed Category, <4.00 mm SL 



Fed larvae began to eat at 3.8 mm (preserved 

 length), yet there was little or no growth for 2 

 days (Fig. 2). Even though, on the average, the 

 fed anchovy grew at 0.41 mm/d, when the delay 

 in growth was considered, the resulting duration 

 for the <4.00 mm fed group was 2.5 days. Thus 

 it took 2.5 days for the fed fish to move out of 

 this size, diet, and cell height category. This 

 initial lag in growth was reported earlier for 

 northern anchovy (Theilacker 1987). 



teria due to indistinct nucleoh, diffused nuclei, 

 and intercellular spaces. However, measure- 

 ment of midgut cell height was still possible after 

 25 minutes. 



The height of the midgut cells, whether 

 measured in formahn- or Bouin's-fixed individ- 

 uals, showed little change in height over the 25 

 min processing period (Fig. 6a-d; slope b is not 

 significant from 0; P = 0.494 and 0.596 for the 

 formalin group and 0.077 and 0.039 for the 

 Bouin's gi'oup, T^ble 5). Although the height of 

 midgut cells is stable within each diet and 1 mm 

 interval size category (Table 3), we weighted the 

 midgut cell height by fish size for this analysis 

 because it included both fed and starved larvae 

 ranging between 2.8 and 5.6 mm. When the mid- 

 gut height of fish belonging to all diets was plot- 

 ted over the 2.8-5.6 mm size range, midgut cell 

 height increased linearly with size (Fig. 4). And, 

 because the fish shrink in the collecting net and 

 the amount of shrinkage is related to the time 

 elapsed in the net, we adjusted fish lengths in 

 this analysis to equal "capture" size us,ing the 

 model developed by Theilacker (1980). 



Durations for Intermediate Categories, 

 <4.00 mm SL 



The durations were 5 days for diet groups 

 where feeding was delayed 1 and 2 days (Sl/F 

 and S2/F). The gi'oup that was starved 3 days 

 before feeding (S3/F) remained within the 

 midgut cell height interval for 8 days (Table 1). 

 The duration for larvae that were starved for 1 

 and 2 days (S1&2) was 2 days. 



Duration for Starved Category, 

 <4.00 mm SL 



The duration for the group starved 3 and 4 

 days (S3&4) was 2 days. Northern anchovy 

 larvae died after starving 3-4 days. No larvae 

 belonging to the starved category were larger 

 than 4.0 mm. 



Field Feasibility Study 



Because larval fish tissues decompose rapidly 

 due to autolysis (Theilacker 1978; O'Connell 

 1980), we tested the effects that the prolonged 

 processing periods encountered at sea have on 

 the integrity of the midgut cells. In this study, 

 after 5 minutes in the net, the condition of both 

 formalin- and Bouin's-fixed larval tissues could 

 not be gi-aded using traditional histological cri- 



Applying the Calibration Criteria 



In practice, when applying this analysis to the 

 field to estimate starvation-induced mortality 

 rates (Owen et al. 1989), it was deemed neces- 

 sary to use only three diet categories (fed, inter- 

 mediate, and starved) instead of the four distinct 

 categories determined by ANOVA (Table 4), 

 and discussed earlier. Because the durations 

 were 5 days for both diets where feeding was 

 delayed 1 and 2 days (Sl/F and S2/F), there was 

 no need for the Sl/F larvae to be a separate, 

 fourth category. Thus we included Sl/F in the 

 intermediate category and selected 5 days as an 

 average duration for the <4.00 mm larvae in this 

 category. 



The cell-height intervals for the three cate- 

 gories were the midpoints between the means of 

 the group cell heights determined for the labora- 

 tory-reared larvae. We regarded all larvae with 

 a midgut height measurement >17.5 |j,m as be- 

 longing to the fed category; the break between 

 the intermediate and starving categories was 

 11.25 |xm. This classification scheme correctly 

 identified 95% of the fed larvae, 77% of the 

 starved larvae, and 74% of the intermediate 

 larvae <4.00 mm SL. 



To calculate the starvation rates of northern 

 anchovy larvae in the sea, we measured the 

 length of each field-collected larva (corrected for 



464 



