FISHERY BULLETIN: VOL. 79, NO. 1 



RANGE 



OCT. 24 • • 



N = 4 



N=29 



RANGE 



NOV 29 



N = 6 



15 

 10 

 5 



SEPT. 



TOTAL 



OCT. 



NOV. 



5 12 19 26 3 10 17 24; 



SEPT. 



OCT 



14 21 28 5 



NOV. 



Figure 4. — Age distributions of Atlantic herring larvae caught 

 in the towed nets as determined by daily growth increments 

 in the otoliths. The ages are represented as the times of first 

 daily growth increment formation. The distributions for the 

 individual sample dates and the total for all the sample dates are 

 given. The dates on the abscissa correspond to the time of 

 formation of the first otolith daily growth increment. These 

 dates are determined by subtracting the number of days, 

 represented by the number of otolith increments, from the date 

 of capture. Two modes are apparent in the total histogram. 



age distribution of larvae from all the towed sam- 

 ples. This bimodal age-frequency distribution was 

 analyzed graphically by the method of Harding 

 (1949) (Figure 5). The total age-frequencies were 

 plotted on probability paper as cumulative per- 

 centages, and were found to fall on a sigmoidal 

 curve having a point of inflection on the 32% 

 vertical. The fitted line EF in Figure 5 is the resul- 

 tant of the two straight lines AB and CD which 

 were found by assuming the data to be bimodally 

 distributed (see Harding 1949 for a more com- 

 plete discussion). Line AB represents those larvae 

 (about 32% of the total) in which the first otolith 

 groMd;h increment was deposited before 30 Octo- 

 ber. This date is approximately the point of inflec- 

 tion of the plotted data. Line CD represents the 

 second, younger group of larvae (about 68% of the 

 total). The two straight lines, AB and CD, indicate 

 that the two groups of Atlantic herring larvae 

 which entered the estuary were normally distrib- 

 uted with respect to age. The mean dates on which 

 the first otolith growth increments were laid down 

 for each age-group were determined by the inter- 

 section of lines AB and CD with the vertical at 

 50%. The standard deviations of these dates were 

 estimated from the points where the two lines 

 intersect the verticals at 15.87% and 84.13%, the 

 standard deviation being half this distance on the 

 Y-axis. The mean dates on which the first daily 

 grovd;h increments were laid down were approxi- 



^ 3 

 26 



O '9 



g 12 



5 



Is 



> 21 



i 1^ 



7 



24 



H 



O 17 

 O 

 10 



26 

 5 



GROUP 2 MEAN DATE -^ NOV 23 

 SO = 19 WEEKS. 



GROUP I MEAN DATE =^ OCT 10 

 SD = i8 WEEKS 



001 01 I 2 5 10 1 20 30 405060 70 80| 90 95 98 99 998 



1587 84 13 



CUMULATIVE PERCENTAGE (N =300) 



Figure 5. — Probability plot of the total bimodal age distribu- 

 tion of Atlantic herring larvae depicted in Figure 4. The dots are 

 the dates on which the first otolith daily growth increment was 

 formed in each larva and are plotted as cumulative percentages. 

 The circles are the cumulative percentages for each of the 2 age- 

 groups using the inflection point of the sigmoidal curve as a 

 dividing point for the two groups. See text for explanation. 



126 



