542 



Fishery Bulletin 89(4), 1991 



Figure 1 



Pavlof Bay, Alaska, sampling area showing 

 55 m isobath (dotted/dashed line) and approx- 

 imate location of the 20 m isobath (dotted). 



I have separated the dominant size modes 

 representing year-classes from Pavlof Bay 

 length-frequency distributions using the 

 maximum-likelihood technique of Macdonald 

 and Pitcher (1979). Annually, calculated 

 mean lengths and numerical abundance of 

 year-classes were then used to estimate 

 growth and mortality using methods similar 

 to those of Frechette and LaBonte (1981) 

 and Anderson (1981). The present study 

 expands on previous work by using modal 

 analysis of length-frequencies from commer- 

 cial catches to estimate fishing and natural 

 mortality rates. Results from these analyses 

 were used to examine the yield-per-recruit 

 function in relation to mortality and growth 

 as well as year-class strength. 



Materials and methods 



Pavlof Bay was surveyed annually from 

 1972 to 1986 by trawling at randomly 

 selected stations from August to mid-Sep- 

 tember (Table 1). During these months 

 dense aggregations of P. borealis form in 

 relatively deep water prior to mating and 

 spawning. Earlier surveys had shown that 

 shrimp concentrate in depths greater than 

 70 m (Ronholt 1963). Consequently, all sur- 

 vey tows were restricted to depths greater 

 than 55 m (Fig. 1). Each year, except 1973, 

 10 to 13 randomly selected (with replace- 

 ment) locations were chosen from a grid of 

 the Bay's shrimp habitat divided into 22 

 stations (~13.7km 2 each). Sampling was 

 conducted with a 30-minute tow (~ 1.8 km) 

 during daylight using a 32 mm mesh (32 mm 

 mesh codend liner) high-opening shrimp 

 trawl with an 18.6 m headrope and footrope 

 (Wathne 1977). The mesh size and config- 

 uration of this trawl were similar to com- 

 mercial fishing gear which allowed for direct 

 comparison between survey and commercial 

 data. 



Carapace lengths (CL) were measured to 

 the nearest 0.5 mm (eye socket to midpos- 

 terior carapace edge) for approximately 300 

 specimens of P. borealis selected at random 



